1 2013-10-17 Joseph Myers <joseph@codesourcery.com>
3 * tree.c (build_common_tree_nodes): Fix typos.
5 2013-09-25 Andrew MacLeod <amacleod@redhat.com>
7 * doc/generic.texi (CP_TYPE_ATOMIC_P): Add documentation.
9 2013-09-25 Andrew MacLeod <amacleod@redhat.com>
12 (build_atomic_base): New. Build atomic base type node.
13 (build_common_tree_nodes): Build atomic{QHSDT}I_type_node, allowing
14 for override with target hook.
16 2013-09-25 Andrew MacLeod <amacleod@redhat.com>
18 * hooks.c (hook_uint_mode_0): Return 0 unit hook.
19 * hooks.h (hook_uint_mode_0): Prototype.
20 * target.def (atomic_align_for_mode): define hook.
21 * doc/tm.texi (TARGET_ATOMIC_ALIGN_FOR_MODE): Define.
22 * doc/tm.texi.in (TARGET_ATOMIC_ALIGN_FOR_MODE): Add.
24 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
27 * doc/extend.texi: Move Cilk Plus Builtins node before Other
30 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
31 Joseph Myers <joseph@codesourcery.com>
33 * ginclude/stdatomic.h: New file.
34 * Makefile.in (USER_H): Add stdatomic.h.
36 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
38 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
41 2013-11-08 Jeff Law <law@redhat.com>
43 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
44 * tree-ssa-threadupdate.c (delete_thread_path): New function.
45 (ssa_redirect_edges, thread_block_1): Use it.
46 (thread_through_loop_header, mark_threaded_blocks): Likewise.
47 (thread_through_all_blocks, register_jump_thread): Likewise.
48 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
50 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
52 * config/arm/aarch-common.c
53 (search_term): New typedef.
54 (shift_rtx_costs): New array.
55 (arm_rtx_shift_left_p): New.
56 (arm_find_sub_rtx_with_search_term): Likewise.
57 (arm_find_sub_rtx_with_code): Likewise.
58 (arm_early_load_addr_dep): Add sanity checking.
59 (arm_no_early_alu_shift_dep): Likewise.
60 (arm_no_early_alu_shift_value_dep): Likewise.
61 (arm_no_early_mul_dep): Likewise.
62 (arm_no_early_store_addr_dep): Likewise.
64 2013-11-08 Richard Biener <rguenther@suse.de>
66 PR tree-optimization/59047
67 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses
70 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
72 * common.opt (fcheck-pointer-bounds): Move to ...
73 * c-family/c.opt: ... here.
74 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
75 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
76 * langhooks.h (lang_hooks): Remove chkp_supported field.
77 * toplev.c (process_options): Remove chkp_supported check.
79 2013-11-08 Richard Biener <rguenther@suse.de>
81 PR tree-optimization/59038
82 PR tree-optimization/58955
83 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
84 previous change. Handle known dependences correctly.
86 2013-11-08 Tom de Vries <tom@codesourcery.com>
88 * config/rs6000/t-xilinx: Remove duplicate contents.
90 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
91 Joseph Myers <joseph@codesourcery.com>
93 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
94 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
95 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
96 (struct tree_base): Add atomic_flag field.
97 * tree.h (TYPE_ATOMIC): New accessor macro.
98 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
99 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
100 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
101 (atomicDI_type_node, atomicTI_type_node): New macros for type
103 * tree.c (set_type_quals): Set TYPE_ATOMIC.
104 (find_atomic_core_type): New function.
105 (build_qualified_type): Adjust alignment for qualified types.
106 (build_atomic_base): New function
107 (build_common_tree_nodes): Build atomicQI_type_node,
108 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
110 * print-tree.c (print_node): Print atomic qualifier.
111 * tree-pretty-print.c (dump_generic_node): Print atomic type
113 * target.def (atomic_assign_expand_fenv): New hook.
114 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
115 * doc/tm.texi: Regenerate.
116 * targhooks.c (default_atomic_assign_expand_fenv): New function.
117 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
118 * sync-builtins.def (__atomic_feraiseexcept): New built-in
120 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
122 * config/i386/i386.c (enum ix86_builtins): Add
123 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
125 (bdesc_special_args): Add __builtin_ia32_fnstenv,
126 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and
127 __builtin_ia32_fnclex.
128 (ix86_expand_builtin): Handle the new built-in functions.
129 (ix86_atomic_assign_expand_fenv): New function.
130 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
131 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
132 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
133 (fnstenv, fldenv, fnstsw, fnclex): New insns.
135 2013-11-07 Steve Ellcey <sellcey@mips.com>
137 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
138 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
139 (MULTILIB_DIRNAMES): Add fp64 directory.
140 (MULTILIB_EXCEPTIONS): Add new exclusions.
142 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
144 * gimplify.c (gimple_regimplify_operands): Do not set
146 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
147 (rewrite_close_phi_out_of_ssa): Same.
148 (rewrite_phi_out_of_ssa): Same.
149 (rewrite_degenerate_phi): Same.
150 (handle_scalar_deps_crossing_scop_limits): Same.
151 * tree-if-conv.c (predicate_scalar_phi): Same.
152 * tree-parloops.c (create_loads_for_reductions): Same.
153 (create_final_loads_for_reduction): Same.
154 (create_loads_and_stores_for_name): Same.
155 (transform_to_exit_first_loop): Same.
156 (create_parallel_loop): Same.
158 (move_computations_dom_walker::before_dom_children): Same.
159 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
160 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
162 * tree-ssa-propagate.c (substitute_and_fold): Same.
163 * tree-vect-loop.c (vect_finalize_reduction): Same.
164 * tree-vect-stmts.c (vectorizable_call): Same.
166 2013-11-07 Mike Stump <mikestump@comcast.net>
168 * config/pdp11/pdp11.c: Include dbxout.h.
169 * config/picochip/picochip.c: Likewise.
171 2013-11-07 Cong Hou <congh@google.com>
173 PR tree-optimization/56764
174 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
175 Combine alias checks if it is possible to amortize the runtime
176 overhead. Return the number of alias checks after merging.
177 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
178 Use the function vect_create_cond_for_alias_checks () to check
179 the number of alias checks.
181 2013-11-07 Jeff Law <law@redhat.com>
183 * varpool.c (ctor_for_folding): Fix typo in comment.
185 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
187 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
188 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
190 2013-11-07 Jeff Law <law@redhat.com>
192 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
194 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
195 No longer check if we have __builtin_trap, assume it's
198 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
200 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
203 * config/arc/arc.h (LINK_COMMAND_SPEC): For
204 -ftree-parallelize-loops=*, link to libgomp and its dependencies.
205 * config/ia64/hpux.h (LIB_SPEC): Likewise.
206 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
207 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
208 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's
211 2013-11-07 Jakub Jelinek <jakub@redhat.com>
213 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
214 (determine_value_range): Add loop argument. Use get_range_info to
216 (bound_difference): Adjust caller.
218 2013-11-07 Richard Biener <rguenther@suse.de>
219 Jakub Jelinek <jakub@redhat.com>
221 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
222 latches from header PHI arguments from the latch edge.
224 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
227 * varasm.c (output_constant): Handle NULLPTR_TYPE.
229 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
231 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
232 misaligned_prologue_used when it has been set.
234 2013-11-07 Yury Gribov <y.gribov@samsung.com>
235 Jakub Jelinek <jakub@redhat.com>
238 * asan.c (get_mem_refs_of_builtin_call): Allow
239 integer literals as addresses in instrumented builtins.
241 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
243 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
244 Explain why plus_constant is not used.
246 2013-11-07 Richard Biener <rguenther@suse.de>
248 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
249 (canonicalize_value): ... this. Also handle stripping of
251 (get_value, set_lattice_value, get_value_for_expr): Adjust.
252 * gimple-fold.c (canonicalize_constructor_val): Strip
254 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
256 2013-11-07 Richard Biener <rguenther@suse.de>
258 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
260 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
263 * config/i386/i386.md (push peepholer/splitter): Use Pmode
264 with stack_pointer_rtx.
266 2013-11-07 Bin Cheng <bin.cheng@arm.com>
268 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
269 using operand_equal_p.
271 2013-11-07 Bin Cheng <bin.cheng@arm.com>
273 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
274 * tree-affine.c (get_inner_reference_aff): Return base.
275 * tree-affine.h (get_inner_reference_aff): Change prototype.
277 2013-11-06 Tobias Burnus <burnus@net-b.de>
279 * doc/invoke.texi (Wdate-time): Fix typo.
281 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
283 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
284 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
285 fp_arith_reg_operand instead of arith_reg_operand.
287 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
289 * config/sh/sh.md (adddi3): Remove empty constraints.
290 Remove can_create_pseudo_p and arith_reg_operand check.
291 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
294 2013-11-06 Jeff Law <law@redhat.com>
295 Tom Tromey <tromey@redhat.com>
297 * gdbinit.in: Disable strict type checking.
299 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
301 * tree-pass.h (make_pass_live_range_shrinkage): New external.
302 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
303 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
304 (rest_of_handle_live_range_shrinkage): Ditto
305 (class pass_live_range_shrinkage): Ditto.
306 (pass_data_live_range_shrinkage): Ditto.
307 (make_pass_live_range_shrinkage): Ditto.
308 * sched-int.h (initialize_live_range_shrinkage): New prototype.
309 (finish_live_range_shrinkage): Ditto.
310 * sched-deps.c (create_insn_reg_set): Make void return value.
311 * passes.def: Add pass_live_range_shrinkage.
312 * ira.c (update_equiv_regs): Don't move if
313 flag_live_range_shrinkage.
314 * haifa-sched.c (live_range_shrinkage_p): New.
315 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
317 (rank_for_schedule): Add code for pressure relief through live
319 (schedule_insn): Print more debug info.
320 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
321 through live range shrinkage.
322 * doc/invoke.texi (-flive-range-shrinkage): New.
323 * common.opt (flive-range-shrinkage): New.
325 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
328 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
329 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
330 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
331 RTXes that return in AVX256 register.
333 2013-11-06 Richard Biener <rguenther@suse.de>
335 PR tree-optimization/58653
336 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
337 (prepare_initializers_chain): Adjust.
339 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
341 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
342 (gimple_in_transaction): Use bb_in_transaction.
343 * basic-block.h (bb_in_transaction): Relocate here and rename.
344 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
346 2013-11-06 Richard Biener <rguenther@suse.de>
348 * tree.c (drop_tree_overflow): New function.
349 * tree.h (drop_tree_overflow): Declare.
350 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
351 * tree-vrp.c (range_int_cst_singleton_p): Use
352 is_overflow_infinity instead of testing TREE_OVERFLOW.
353 (extract_range_from_assert): Likewise.
354 (zero_nonzero_bits_from_vr): Likewise.
355 (extract_range_basic): Likewise.
356 (register_new_assert_for): Use drop_tree_overflow.
357 (vrp_visit_phi_node): Likewise.
359 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
361 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
362 checking for leaf functions without dynamic stack allocation.
363 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
364 (ia64_expand_prologue): Likewise.
365 * config/mips/mips.c (mips_expand_prologue): Likewise.
366 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
367 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
368 (sparc_flat_expand_prologue): Likewise.
370 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
372 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
373 (__ST3_LANE_FUNC): Likewise.
374 (__ST4_LANE_FUNC): Likewise.
376 2013-11-06 Nick Clifton <nickc@redhat.com>
378 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
379 name returned by msp430_mcu_name.
380 (LIB_SPEC): If a -T option has not been specified then set a
381 default, mcu-specific, linker script.
382 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
383 * config/msp430/msp430.c (msp430x_names): Likewise.
384 Alpha sort the names for ease of comparison.
385 (msp430_mcu_name): New function: Returns a string suitable for
386 use as a C preprocessor symbol based upon the name of the MCU
388 (msp430_option_override): Accept msp430x and msp430xv2 as generic
390 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
392 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
393 unless it is the last part of a spec substring.
394 * doc/invoke.texi (Spec Files): Document space insertion
397 2013-11-06 Christian Bruel <christian.bruel@st.com>
399 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
400 Factorize probabilities, Use adjust_address instead of
401 adjust_automodify_address when possible. Enable for optimize.
402 (sh_expand_strlen): New function.
403 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
404 * config/sh/sh.md (strlensi): New pattern.
405 (UNSPEC_BUILTIN_STRLEN): Define.
407 2013-11-06 Jakub Jelinek <jakub@redhat.com>
410 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
411 (expand_assignment): If *bitpos is negative, set *offset
412 and adjust *bitpos, so that it is not negative.
414 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
416 * config/i386/bdver3.md : Added two additional decoder units
417 to support issue rate of 4 and remodeled vector unit.
418 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
419 architectures is set to 4.
420 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
421 lookahead is set to 4 for BD architectures.
423 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
425 * config/rs6000/rs6000.c (rs6000_option_override_internal):
426 Remove restriction against use of VSX instructions when generating
427 code for little endian mode.
429 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
431 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
432 for both big and little endian.
433 (mulv8hi3): Swap input operands for merge high and merge low
434 instructions for little endian.
436 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
438 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
439 define_insn to define_expand that uses even patterns for big
440 endian and odd patterns for little endian.
441 (vec_widen_smult_even_v16qi): Likewise.
442 (vec_widen_umult_even_v8hi): Likewise.
443 (vec_widen_smult_even_v8hi): Likewise.
444 (vec_widen_umult_odd_v16qi): Likewise.
445 (vec_widen_smult_odd_v16qi): Likewise.
446 (vec_widen_umult_odd_v8hi): Likewise.
447 (vec_widen_smult_odd_v8hi): Likewise.
448 (altivec_vmuleub): New define_insn.
449 (altivec_vmuloub): Likewise.
450 (altivec_vmulesb): Likewise.
451 (altivec_vmulosb): Likewise.
452 (altivec_vmuleuh): Likewise.
453 (altivec_vmulouh): Likewise.
454 (altivec_vmulesh): Likewise.
455 (altivec_vmulosh): Likewise.
457 2013-11-05 Mike Stump <mikestump@comcast.net>
459 * Makefile.in (mostlyclean): Remove c-family objects.
461 2013-11-05 Ian Lance Taylor <iant@google.com>
463 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
464 If possible, add .cfi directives to record change to bx.
465 * config/i386/i386.c (ix86_emit_cfi): New function.
466 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
468 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
470 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
471 immediately after a label for a tablejump pattern.
473 * config/arm/arm.c (is_jump_table): Remove.
474 (create_fix_barrier): Use tablejump_p instead.
475 (arm_reorg): Likewise.
476 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
477 (thumb2_output_casesi): Likewise.
478 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
479 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
480 casesi_shift_media, casesi_load_media): Likewise.
481 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
482 * config/microblaze/microblaze.md: Likewise.
484 2013-11-05 Tobias Burnus <burnus@net-b.de>
486 * doc/invoke.texi (-Wdate-time): Document.
488 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
490 * double-int.c (lshift_double, rshift_double): Remove
491 SHIFT_COUNT_TRUNCATED handling.
493 2013-11-05 Jeff Law <law@redhat.com>
495 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
496 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
497 * gimple-ssa-isolate-paths.c: New file.
498 * gimple.c (check_loadstore): New function.
499 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
500 Verify OP is in the argument list and the argument corresponding
501 to OP is a pointer type. Use operand_equal_p rather than
502 pointer equality when testing if OP is on the nonnull list.
503 Use check_loadstore rather than count_ptr_derefs. Handle
504 GIMPLE_RETURN statements.
505 * tree-vrp.c (infer_nonnull_range): Remove.
506 * gimple.h (infer_nonnull_range): Declare.
507 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
508 * passes.def: Add pass_isolate_erroneous_paths.
509 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
510 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
511 * tree-ssa.c (struct count_ptr_d): Remove.
512 (count_ptr_derefs, count_uses_and_derefs): Remove.
513 * tree-ssa.h (count_uses_and_derefs): Remove.
515 2013-11-05 Jakub Jelinek <jakub@redhat.com>
517 PR rtl-optimization/58997
518 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
519 get_iv_value to be in iv->mode rather than iv->extend_mode.
520 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
521 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
522 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
525 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
527 * gimple.h: Move some prototypes to gimple-expr.h and add to include
529 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
530 Move to gimple-expr.h.
531 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
532 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
533 is_gimple_id, virtual_operand_p, is_gimple_addressable,
534 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
536 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
537 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
538 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
539 is_gimple_address, is_gimple_invariant_address,
540 is_gimple_ip_invariant_address, is_gimple_min_invariant,
541 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
542 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
543 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
544 gimple_decl_printable_name, useless_type_conversion_p,
545 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
547 * gimple-expr.c: New File.
548 (useless_type_conversion_p, gimple_set_body, gimple_body,
549 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
550 gimple_can_coalesce_p, extract_ops_from_tree_1,
551 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
552 is_gimple_address, is_gimple_invariant_address,
553 is_gimple_ip_invariant_address, is_gimple_min_invariant,
554 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
555 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
556 is_gimple_mem_ref_addr): Relocate here.
557 * Makefile.in (OBJS): Add gimple-expr.o.
559 2013-11-05 David Malcolm <dmalcolm@redhat.com>
561 * gengtype-parse.c (struct_field_seq): Support empty structs.
563 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
565 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
567 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
569 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
571 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
573 (LONG_DOUBLE_TYPE_SIZE): New define.
574 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
576 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
579 * doc/extend.texi [visibility ("visibility_type")]: Add example
580 about visibility attribute on namespace declaration.
582 2013-11-05 Richard Biener <rguenther@suse.de>
585 * passes.def (all_passes): Start with pass_fixup_cfg again.
587 2013-11-05 Richard Biener <rguenther@suse.de>
589 PR tree-optimization/58955
590 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
593 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
595 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
597 (vec_pack_ufix_trunc_v2df): Likewise.
599 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
602 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
603 pattern, instead of word_mode.
605 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
607 (set_storage_via_setmem): Likewise.
609 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
611 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
612 gimple.h and the rest of the condition in eliminate_build.
613 (eliminate_build): Call new routine.
614 * gimple.h (phi_ssa_name_p): Delete.
616 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
618 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
619 case of no prefix and reserving zero slots, because when that's the
620 case we'll never get here.
621 * vec.h (va_heap::reserve): Don't try and handle
622 vec_prefix::calculate_allocation returning zero because that should
625 2013-11-05 Richard Biener <rguenther@suse.de>
628 * tree-dfa.c (get_ref_base_and_extent): Merge common code
629 in MEM_REF and TARGET_MEM_REF handling. Make sure to
630 process trailing array detection before diving into the
631 view-converted object (and possibly apply some extra offset).
633 2013-11-05 Joseph Myers <joseph@codesourcery.com>
635 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
637 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
639 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
641 PR tree-optimization/58958
642 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
643 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
645 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
647 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
648 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
650 2013-11-05 Jakub Jelinek <jakub@redhat.com>
652 PR tree-optimization/58984
653 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
654 set *SIZE_P if non-NULL on success.
655 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
657 (ipcp_transform_function): Likewise. Punt if size of access
658 is different from TYPE_SIZE on v->value's type.
660 2013-11-05 Tobias Burnus <burnus@net-b.de>
662 * doc/invoke.texi (-fopenmp-simd): Document new option.
663 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
664 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
665 * tree.c (attribute_value_equal): Ditto.
667 2013-11-04 Wei Mi <wmi@google.com>
669 * sched-rgn.c (add_branch_dependences): Keep insns in
670 a SCHED_GROUP at the end of BB to remain their location.
672 2013-11-04 Wei Mi <wmi@google.com>
674 * config/i386/i386.c (memory_address_length): Extract a part
675 of code to rip_relative_addr_p.
676 (rip_relative_addr_p): New Function.
677 (ix86_macro_fusion_p): Ditto.
678 (ix86_macro_fusion_pair_p): Ditto.
679 * config/i386/i386.h: Add new tune features about macro-fusion.
680 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
681 * doc/tm.texi: Generated.
682 * doc/tm.texi.in: Ditto.
683 * haifa-sched.c (try_group_insn): New Function.
684 (group_insns_for_macro_fusion): Ditto.
685 (sched_init): Call group_insns_for_macro_fusion.
686 * target.def: Add two hooks: macro_fusion_p and
689 2013-11-04 Kostya Serebryany <kcc@google.com>
691 Update to match the changed asan API.
692 * asan.c (asan_function_start): New function.
693 (asan_emit_stack_protection): Update the string stored in the
694 stack red zone to match new API. Store the PC of the current
695 function in the red zone.
696 (asan_global_struct): Update the __asan_global definition to match
698 (asan_add_global): Ditto.
699 * asan.h (asan_function_start): New prototype.
700 * final.c (final_start_function): Call asan_function_start.
701 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
704 2013-11-04 Wei Mi <wmi@google.com>
706 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
707 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
708 * config/i386/i386.c (ix86_option_override_internal): Ditto.
709 (ix86_issue_rate): Ditto.
710 (ix86_adjust_cost): Ditto.
711 (ia32_multipass_dfa_lookahead): Ditto.
712 (ix86_sched_init_global): Ditto.
713 (get_builtin_code_for_version): Ditto.
714 * config/i386/i386.h (enum target_cpu_default): Ditto.
715 (enum processor_type): Ditto.
716 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
718 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
720 PR rtl-optimization/58967
721 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
722 !lra_in_progress for mode sizes bigger word.
724 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
726 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
727 arguments to merge instruction for little endian.
728 (vec_widen_umult_lo_v16qi): Likewise.
729 (vec_widen_smult_hi_v16qi): Likewise.
730 (vec_widen_smult_lo_v16qi): Likewise.
731 (vec_widen_umult_hi_v8hi): Likewise.
732 (vec_widen_umult_lo_v8hi): Likewise.
733 (vec_widen_smult_hi_v8hi): Likewise.
734 (vec_widen_smult_lo_v8hi): Likewise.
736 2013-11-04 Ian Lance Taylor <iant@google.com>
738 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
739 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
740 builtins that take pointers.
741 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
742 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
743 (run_gcc): Pass -fnon-call-exceptions.
745 2013-11-04 Jakub Jelinek <jakub@redhat.com>
747 * optabs.c (expand_vec_perm): Revert one incorrect line from
750 PR tree-optimization/58978
751 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
752 use_stmt by single_imm_use directly. Only call single_imm_use
755 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
757 PR rtl-optimization/58968
758 * lra-spills.c (return_regno_p): New function.
759 (lra_final_code_change): Use it.
761 2013-11-04 Joseph Myers <joseph@codesourcery.com>
763 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
764 * target.def (float_exceptions_rounding_supported_p): New hook.
765 * targhooks.c (default_float_exceptions_rounding_supported_p): New
767 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
768 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
770 * doc/tm.texi: Regenerate.
771 * config.gcc (powerpc*-*-linux*): Set extra_objs.
772 * config/rs6000/rs6000-linux.c: New file.
773 * config/rs6000/rs6000-protos.h
774 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
775 * config/rs6000/linux.h
776 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
777 * config/rs6000/linux64.h
778 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
779 * config/rs6000/t-linux (rs6000-linux.o): New rule.
780 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
782 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
784 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
785 Replace the define_insn_and_split with a define_insn and two
786 define_splits, with the split after reload re-permuting the source
787 register to its original value.
788 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
789 (*vsx_le_perm_store_v8hi): Likewise.
790 (*vsx_le_perm_store_v16qi): Likewise.
792 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
794 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
797 2013-11-04 Jakub Jelinek <jakub@redhat.com>
799 PR tree-optimization/58946
800 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
801 bbs with bbinfo[idx].op != NULL before all blocks with
802 bbinfo[idx].op == NULL.
804 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
806 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
807 (avr_dump_double_int_hex): Likewise.
808 (avr_log_vadump): Remove %D and %X handling.
809 * config/avr/avr.c (avr_double_int_push_digit): Delete.
810 (avr_map_op_t): Change map from double_int to unsigned int.
811 (avr_map_op): Update accordingly.
812 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
813 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
814 unsigned ints rather than double_ints.
816 2013-11-03 Marek Polacek <polacek@redhat.com>
818 Implement -fsanitize=vla-bound.
819 * opts.c (common_handle_option): Handle vla-bound.
820 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
821 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
822 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
824 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
826 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
829 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
831 * config/i386/constraints.md (Ts, Tv): New address constrains.
832 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
833 constraint for address_no_seg_operand.
834 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
835 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
836 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
837 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
838 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
839 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
840 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
841 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
842 constraint for vsib_address_operand.
844 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
846 * gcse.c (pre_delete): Remove references to regmove from comments.
847 * recog.c: (validate_replace_rtx_1): Likewise.
848 * config/rl78/rl78.c: Likewise.
849 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
850 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
851 * common/config/mmix/mmix-common.c: Likewise.
853 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
855 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
856 * gimplify.c (gimplify_compound_lval): Likewise.
857 * graphite-clast-to-gimple.c (gloog): Likewise.
858 * graphite-dependences.c (loop_is_parallel_p): Likewise.
859 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
860 (limit_scop); Likewise.
861 (build_scops): Likewise.
862 (dot_scop): Likewise.
863 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
864 (build_scop_drs): Likewise.
865 (insert_stmts): Likewise.
866 (insert_out_of_ssa_copy): Likewise.
867 (remove_phi): Likewise.
868 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
869 * hw-doloop.c (discover_loop): Likewise.
870 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
871 * tree-dfa.c (dump_enumerated_decls): Likewise.
872 * tree-if-conv.c (if_convertable_loop_p): Likewise.
873 * tree-inline.c (tree_function_versioning): Likewise.
874 * tree-loop-distribution.c (build_rdg): Likewise.
875 (rdg_flag_vertex_and_dependent): Likewise.
876 (distribute_loop): Likewise.
877 * tree-parloops.c (loop_parallel_p): Likewise.
878 (eliminate_local_variables): Likewise.
879 (separate_decls_in_region): Likewise.
880 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
881 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
882 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
883 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
884 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
885 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
886 (vectorizable_condition): Likewise.
888 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
890 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
891 * configure: Regenerate.
892 * config/i386/i386.md (*movdi_internal): Change
893 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
894 (*movdf_internal): Ditto.
895 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
896 * config/i386/sse.md (vec_concatv2di): Output interunit movq
897 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
899 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
901 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
904 2013-10-31 David Malcolm <dmalcolm@redhat.com>
906 Automated part of renaming of symtab_node_base to symtab_node.
908 Patch autogenerated by rename_symtab.py from
909 https://github.com/davidmalcolm/gcc-refactoring-scripts
910 revision 58bb219cc090b2f4516a9297d868c245495ee622
911 with ChangeLog entry fixed up by hand.
913 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
915 (cgraph_node_for_asm): Likewise.
916 * cgraph.h (symtab_node_base): Likewise.
917 (cgraph_node): Likewise.
918 (varpool_node): Likewise.
919 (is_a_helper <cgraph_node>::test): Likewise.
920 (is_a_helper <varpool_node>::test): Likewise.
921 (symtab_nodes): Likewise.
922 (symtab_register_node): Likewise.
923 (symtab_unregister_node): Likewise.
924 (symtab_remove_node): Likewise.
925 (symtab_get_node): Likewise.
926 (symtab_node_for_asm): Likewise.
927 (symtab_node_asm_name): Likewise.
928 (symtab_node_name): Likewise.
929 (symtab_insert_node_to_hashtable): Likewise.
930 (symtab_add_to_same_comdat_group): Likewise.
931 (symtab_dissolve_same_comdat_group_list): Likewise.
932 (dump_symtab_node): Likewise.
933 (debug_symtab_node): Likewise.
934 (dump_symtab_base): Likewise.
935 (verify_symtab_node): Likewise.
936 (verify_symtab_base): Likewise.
937 (symtab_used_from_object_file_p): Likewise.
938 (symtab_alias_ultimate_target): Likewise.
939 (symtab_resolve_alias): Likewise.
940 (fixup_same_cpp_alias_visibility): Likewise.
941 (symtab_for_node_and_aliases): Likewise.
942 (symtab_nonoverwritable_alias): Likewise.
943 (availability symtab_node_availability): Likewise.
944 (symtab_semantically_equivalent_p): Likewise.
945 (fixup_same_cpp_alias_visibility): Likewise.
946 (symtab_prevail_in_asm_name_hash): Likewise.
949 (varpool_first_variable): Likewise.
950 (varpool_next_variable): Likewise.
951 (varpool_first_static_initializer): Likewise.
952 (varpool_next_static_initializer): Likewise.
953 (varpool_first_defined_variable): Likewise.
954 (varpool_next_defined_variable): Likewise.
955 (cgraph_first_defined_function): Likewise.
956 (cgraph_next_defined_function): Likewise.
957 (cgraph_first_function): Likewise.
958 (cgraph_next_function): Likewise.
959 (cgraph_first_function_with_gimple_body): Likewise.
960 (cgraph_next_function_with_gimple_body): Likewise.
961 (symtab_alias_target): Likewise.
962 (symtab_real_symbol_p): Likewise.
963 (symtab_can_be_discarded): Likewise.
964 * cgraphbuild.c (mark_address): Likewise.
965 (mark_load): Likewise.
966 (mark_store): Likewise.
967 * cgraphunit.c (decide_is_symbol_needed): Likewise.
969 (enqueue_node): Likewise.
970 (referred_to_p): Likewise.
971 (cgraph_process_same_body_aliases): Likewise.
972 (analyze_functions): Likewise.
973 (handle_alias_pairs): Likewise.
974 (output_weakrefs): Likewise.
976 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
977 * ipa-inline-analysis.c (inline_write_summary): Likewise.
978 * ipa-prop.c (remove_described_reference): Likewise.
979 (try_decrement_rdesc_refcount): Likewise.
980 (ipa_edge_duplication_hook): Likewise.
981 * ipa-ref.c (ipa_record_reference): Likewise.
982 (ipa_maybe_record_reference): Likewise.
983 (ipa_clone_ref): Likewise.
984 (ipa_clone_references): Likewise.
985 (ipa_clone_referring): Likewise.
986 (ipa_find_reference): Likewise.
987 (ipa_remove_stmt_references): Likewise.
988 (ipa_clear_stmts_in_references): Likewise.
989 * ipa-ref.h (symtab_node_base): Likewise.
991 (ipa_record_reference): Likewise.
992 (ipa_maybe_record_reference): Likewise.
993 (ipa_clone_references): Likewise.
994 (ipa_clone_referring): Likewise.
995 (ipa_clone_ref): Likewise.
996 (ipa_find_reference): Likewise.
997 (ipa_remove_stmt_references): Likewise.
998 (ipa_clear_stmts_in_references): Likewise.
999 * ipa-reference.c (ipa_reference_write_optimization_summary):
1001 * ipa.c (enqueue_node): Likewise.
1002 (process_references): Likewise.
1003 (walk_polymorphic_call_targets): Likewise.
1004 (symtab_remove_unreachable_nodes): Likewise.
1005 (address_taken_from_non_vtable_p): Likewise.
1006 (comdat_can_be_unshared_p_1): Likewise.
1007 (comdat_can_be_unshared_p): Likewise.
1008 (can_replace_by_local_alias): Likewise.
1009 (function_and_variable_visibility): Likewise.
1010 * is-a.h: Likewise (within example in comment).
1011 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
1012 (lto_symtab_encoder_encode): Likewise.
1013 (lto_symtab_encoder_delete_node): Likewise.
1014 (lto_symtab_encoder_in_partition_p): Likewise.
1015 (lto_set_symtab_encoder_in_partition): Likewise.
1016 (output_refs): Likewise.
1017 (compute_ltrans_boundary): Likewise.
1018 (output_symtab): Likewise.
1019 (input_node): Likewise.
1020 (input_ref): Likewise.
1021 (input_edge): Likewise.
1022 (input_cgraph_1): Likewise.
1023 (input_refs): Likewise.
1024 (output_cgraph_opt_summary): Likewise.
1025 (input_node_opt_summary): Likewise.
1026 (input_cgraph_opt_section): Likewise.
1027 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1029 * lto-streamer-out.c (lto_output): Likewise.
1030 (output_symbol_p): Likewise.
1031 (produce_symtab): Likewise.
1032 * lto-streamer.h (lto_encoder_entry): Likewise.
1033 (lto_free_function_in_decl_state_for_node): Likewise.
1034 (lto_symtab_encoder_encode): Likewise.
1035 (lto_symtab_encoder_delete_node): Likewise.
1036 (lto_symtab_encoder_in_partition_p): Likewise.
1037 (lto_set_symtab_encoder_in_partition): Likewise.
1038 (lto_symtab_encoder_lookup): Likewise.
1039 (lsei_node): Likewise.
1040 (lto_symtab_encoder_deref): Likewise.
1041 * symtab.c (symtab_hash): Likewise.
1042 (assembler_name_hash): Likewise.
1043 (symtab_nodes): Likewise.
1044 (hash_node): Likewise.
1045 (eq_node): Likewise.
1046 (hash_node_by_assembler_name): Likewise.
1047 (eq_assembler_name): Likewise.
1048 (insert_to_assembler_name_hash): Likewise.
1049 (unlink_from_assembler_name_hash): Likewise.
1050 (symtab_prevail_in_asm_name_hash): Likewise.
1051 (symtab_register_node): Likewise.
1052 (symtab_insert_node_to_hashtable): Likewise.
1053 (symtab_unregister_node): Likewise.
1054 (symtab_get_node): Likewise.
1055 (symtab_remove_node): Likewise.
1056 (symtab_initialize_asm_name_hash): Likewise.
1057 (symtab_node_for_asm): Likewise.
1058 (symtab_add_to_same_comdat_group): Likewise.
1059 (symtab_dissolve_same_comdat_group_list): Likewise.
1060 (symtab_node_asm_name): Likewise.
1061 (symtab_node_name): Likewise.
1062 (dump_symtab_base): Likewise.
1063 (dump_symtab_node): Likewise.
1064 (dump_symtab): Likewise.
1065 (debug_symtab_node): Likewise.
1066 (verify_symtab_base): Likewise.
1067 (verify_symtab_node): Likewise.
1068 (verify_symtab): Likewise.
1069 (symtab_used_from_object_file_p): Likewise.
1070 (symtab_node_availability): Likewise.
1071 (symtab_alias_ultimate_target): Likewise.
1072 (fixup_same_cpp_alias_visibility): Likewise.
1073 (symtab_resolve_alias): Likewise.
1074 (symtab_for_node_and_aliases): Likewise.
1075 (symtab_for_node_and_aliases): Likewise.
1076 (symtab_nonoverwritable_alias_1): Likewise.
1077 (symtab_nonoverwritable_alias): Likewise.
1078 (symtab_semantically_equivalent_p): Likewise.
1079 * value-prof.c (init_node_map): Likewise.
1080 * varasm.c (find_decl): Likewise.
1081 * varpool.c (varpool_node_for_asm): Likewise.
1082 (varpool_remove_unreferenced_decls): Likewise.
1084 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1086 Manual part of renaming of symtab_node_base to symtab_node.
1088 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
1089 clashes with the preferred name for the base class.
1090 (const_symtab_node): Remove redundant typedef.
1092 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1094 * optabs.c (expand_vec_perm): Avoid vector mode punning
1095 SUBREGs in SET_DEST.
1096 * expmed.c (store_bit_field_1): Likewise.
1097 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
1098 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
1099 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
1100 * config/i386/i386.c (ix86_expand_vector_move_misalign,
1101 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
1102 ix86_expand_sse_unpack, ix86_expand_args_builtin,
1103 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
1104 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
1105 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
1106 expand_vec_perm_vpshufb2_vpermq,
1107 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
1108 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
1109 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
1110 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
1113 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
1115 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
1116 Use gcc_unreachable for unhandled modes. Do not check results of
1117 expand_simple_binop. If not expanded to target, move the result.
1119 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
1120 Shiva Chen <shiva0217@gmail.com>
1122 * config.gcc (nds32*-*-*): Add nds32 target.
1123 * config/nds32/nds32.c: New file.
1124 * config/nds32/nds32.h: New file.
1125 * config/nds32/nds32.md: New file.
1126 * config/nds32/constants.md: New file.
1127 * config/nds32/constraints.md: New file.
1128 * config/nds32/iterators.md: New file.
1129 * config/nds32/nds32-doubleword.md: New file.
1130 * config/nds32/nds32-intrinsic.md: New file.
1131 * config/nds32/nds32_intrinsic.h: New file.
1132 * config/nds32/nds32-modes.def: New file.
1133 * config/nds32/nds32-multiple.md: New file.
1134 * config/nds32/nds32.opt: New file.
1135 * config/nds32/nds32-opts.h: New file.
1136 * config/nds32/nds32-protos.h: New file.
1137 * config/nds32/nds32-peephole2.md: New file.
1138 * config/nds32/pipelines.md: New file.
1139 * config/nds32/predicates.md: New file.
1140 * config/nds32/t-mlibs: New file.
1141 * common/config/nds32: New directory and files.
1143 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
1144 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
1145 * doc/install.texi (Cross-Compiler-Specific Options): Document
1146 --with-nds32-lib for nds32 target.
1147 * doc/extend.texi (Function Attributes, Target Builtins): Document
1148 nds32 specific attributes.
1150 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
1152 * lra-constraints (process_alt_operands): Use the result
1153 elimination register for operand when matching constraints.
1155 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1157 * tree-vrp.c (maybe_set_nonzero_bits): New function.
1158 (remove_range_assertions): Call it.
1160 * tree.c (tree_ctz): New function.
1161 * tree.h (tree_ctz): New prototype.
1162 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
1163 first argument from tree to const_tree.
1164 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
1165 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
1166 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
1168 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
1169 * tree-vect-loop.c (vect_analyze_loop_operations,
1170 vect_transform_loop): Don't force scalar loop for bound just because
1171 number of iterations is unknown, only do it if it is not known to be
1172 a multiple of vectorization_factor.
1173 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
1175 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
1176 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
1177 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
1178 (set_nonzero_bits, get_nonzero_bits): New prototypes.
1179 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
1180 a default def isn't partially constant.
1181 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
1182 is known to be partially zero.
1183 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
1184 to see if a default def isn't partially constant.
1185 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
1186 creation of a range, if VR_RANGE, try to improve nonzero_bits from
1188 (set_nonzero_bits, get_nonzero_bits): New functions.
1190 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
1191 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
1192 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
1193 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
1194 uses but in the condition and ASSERT_EXPR and the other successor of
1195 the predecessor bb is __builtin_unreachable (), set_range_info of the
1196 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
1198 2013-10-31 Martin Jambor <mjambor@suse.cz>
1200 PR rtl-optimization/58934
1202 2013-10-30 Martin Jambor <mjambor@suse.cz>
1203 PR rtl-optimization/10474
1204 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1206 (interesting_dest_for_shprep): New function.
1207 (split_live_ranges_for_shrink_wrap): Likewise.
1208 (ira): Calculate dominance info and df analysis. Call
1209 split_live_ranges_for_shrink_wrap.
1211 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1212 Yury Gribov <y.gribov@samsung.com>
1215 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
1216 shadow pointer to avoid clobbering the main one.
1218 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1220 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
1222 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1225 * ira-color.c (update_costs_from_copies): Add new parameter. Use
1226 it for calling update_costs_from_allocno.
1227 (assign_hard_reg): Call restore_costs_from_copies only for
1228 !retry_p. Pass new argument to update_costs_from_copies.
1229 (color_pass): Pass new argument to update_costs_from_copies.
1230 (ira_mark_allocation_change): Ditto.
1232 2013-10-30 Sharad Singhai <singhai@google.com>
1235 * opts.c (common_handle_option): Remove deprecated option
1236 -ftree-vectorizer-verbose.
1237 * doc/invoke.texi (Debugging Options): Ditto.
1238 * opts-global.c (handle_common_deferred_options): Ditto.
1239 (dump_remap_tree_vectorizer_verbose): Delete.
1240 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
1242 2013-10-30 DJ Delorie <dj@redhat.com>
1244 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
1245 that take no arguments.
1247 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
1250 * reload1.c (update_eliminables_and_spill): New function, broken
1252 (reload): Use it. Check for frame size change after frame size
1253 alignment, and call update_eliminables_and_spill first if continue-ing.
1255 2013-10-30 Cong Hou <congh@google.com>
1258 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
1259 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
1260 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
1262 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
1264 PR rtl-optimization/58369
1265 * reload1.c (compute_reload_subreg_offset): New function.
1266 (choose_reload_regs): Use it to pass endian-correct
1267 offset to subreg_regno_offset.
1269 2013-10-30 Tobias Burnus <burnus@net-b.de>
1272 * tree-cfg.c (replace_loop_annotate): Replace warning by
1275 2013-10-30 Jason Merrill <jason@redhat.com>
1277 * configure.ac (loose_warn): Add -Wno-format if
1278 --disable-build-format-warnings.
1280 2013-10-30 David Malcolm <dmalcolm@redhat.com>
1282 * cgraphunit.c (analyze_functions): Split symtab_node declarations
1283 onto multiple lines to make things easier for rename_symtab.py.
1285 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
1286 (symtab_semantically_equivalent_p): Likewise.
1288 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1291 * lra.c (check_rtl): Remove address check before LRA work.
1293 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
1295 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
1296 POINTER_PLUS_EXPR in the defining statement.
1298 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1300 * regmove.c: Remove.
1301 * tree-pass.h (make_pass_regmove): Remove.
1302 * timevar.def (TV_REGMOVE): Remove.
1303 * passes.def (pass_regmove): Remove.
1304 * opts.c (default_options_table): Remove entry for regmove.
1305 * doc/passes.texi: Remove regmove pass description.
1306 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
1308 (-fdump-rtl-regmove): Ditto.
1309 * common.opt (foptimize-register-move, fregmove): Ignore.
1310 * Makefile.in (OBJS): Remove regmove.o.
1311 * regmove.c: Remove.
1312 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
1314 (struct ira_allocno) New member allocno_prefs.
1315 (ALLOCNO_PREFS): New macro.
1316 (ira_prefs, ira_prefs_num): New external vars.
1317 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
1318 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
1319 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
1321 (ira_add_allocno_copy_to_list): Remove prototype.
1322 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
1323 (ira_pref_iterator): New type.
1324 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
1325 (FOR_EACH_PREF): New macro.
1326 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
1327 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
1329 (ira_setup_alts): New function.
1330 (decrease_live_ranges_number): New function.
1331 (ira): Call the above function.
1332 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
1333 (ira_create_allocno): Initialize allocno prefs.
1334 (pref_pool, pref_vec): New static vars.
1335 (initiate_prefs, find_allocno_pref, ira_create_pref): New
1337 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
1338 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
1339 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
1340 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
1341 (ira_add_allocno_copy_to_list): Make static. Rename to
1342 add_allocno_copy_to_list.
1343 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
1344 swap_allocno_copy_ends_if_necessary.
1345 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
1346 ira_remove_allocno_prefs.
1347 (ira_flattening): Ditto.
1348 (ira_build): Call initiate_prefs, print_prefs.
1349 (ira_destroy): Call finish_prefs.
1350 * ira-color.c (struct update_cost_record): New.
1351 (struct allocno_color_data): Add new member update_cost_records.
1352 (update_cost_record_pool): New static var.
1353 (init_update_cost_records, get_update_cost_record): New functions.
1354 (free_update_cost_record_list, finish_update_cost_records): Ditto.
1355 (struct update_cost_queue_elem): Add member from.
1356 (initiate_cost_update): Call init_update_cost_records.
1357 (finish_cost_update): Call finish_update_cost_records.
1358 (queue_update_cost, get_next_update_cost): Add new param from.
1359 (Update_allocno_cost, update_costs_from_allocno): New functions.
1360 (update_costs_from_prefs): Ditto.
1361 (update_copy_costs): Rename to update_costs_from_copies.
1362 (restore_costs_from_copies): New function.
1363 (update_conflict_hard_regno_costs): Don't go back.
1364 (assign_hard_reg): Call restore_costs_from_copies. Add printing
1366 (pop_allocnos): Add priniting more debug info.
1367 (color_allocnos): Remove prefs for conflicting hard regs.
1368 Call update_costs_from_prefs.
1369 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
1370 (get_dup_num): Rename, modify, and move to ira.c
1371 (process_regs_for_copy): Add prefs.
1372 (add_insn_allocno_copies): Put src as first arg of
1373 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
1374 * ira-costs.c (record_reg_classes): Modify and move code into
1375 record_operands_costs.
1376 (find_costs_and_classes): Create prefs for the hard reg of small
1378 (process_bb_node_for_hard_reg_moves): Add prefs.
1380 2013-10-30 Richard Biener <rguenther@suse.de>
1383 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
1384 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
1386 (pre_and_rev_post_order_compute): ... which now wraps it.
1387 * graph.c (draw_cfg_nodes_no_loops): Use
1388 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
1391 2013-10-30 Christian Bruel <christian.bruel@st.com>
1393 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
1394 (sh_expand_cmpstr): Handle known align and schedule improvements.
1395 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
1396 * config/sh/sh.md (cmpstrnsi): New pattern.
1398 2013-10-30 Martin Jambor <mjambor@suse.cz>
1400 PR rtl-optimization/10474
1401 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1403 (interesting_dest_for_shprep): New function.
1404 (split_live_ranges_for_shrink_wrap): Likewise.
1405 (ira): Calculate dominance info and df analysis. Call
1406 split_live_ranges_for_shrink_wrap.
1408 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1411 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1413 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1415 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1416 * tree.h (POINTER_BOUNDS_P): New.
1417 (BOUNDED_TYPE_P): New.
1419 (pointer_bounds_type_node): New.
1420 * tree.c (build_common_tree_nodes): Initialize
1421 pointer_bounds_type_node.
1422 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1423 (gimple_call_num_nobnd_args): New.
1424 (gimple_call_nobnd_arg): New.
1425 (gimple_return_retbnd): New.
1426 (gimple_return_set_retbnd): New
1427 * gimple.c (gimple_build_return): Increase number of ops
1428 for return statement.
1429 (gimple_call_get_nobnd_arg_index): New.
1430 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1432 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1434 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1435 with "chkp ctor" and "bnd_legacy" attributes.
1436 * gimplify.c (gimplify_init_constructor): Avoid infinite
1437 loop during gimplification of bounds initializer.
1439 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1441 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1442 (handle_bnd_legacy): New.
1443 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1444 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1447 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1449 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1450 (BT_FN_PTR_CONST_PTR): New.
1451 (BT_FN_CONST_PTR_CONST_PTR): New.
1452 (BT_FN_PTR_CONST_PTR_SIZE): New.
1453 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1454 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1455 (BT_FN_VOID_CONST_PTR_SIZE): New.
1456 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1457 * chkp-builtins.def: New.
1458 * builtins.def: include chkp-builtins.def.
1459 (DEF_CHKP_BUILTIN): New.
1460 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1461 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1462 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1463 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1464 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1465 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1466 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1467 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1468 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1469 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1470 * common.opt (fcheck-pointer-bounds): New.
1471 * toplev.c (process_options): Check Pointer Bounds Checker is
1473 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1475 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1477 * target.def (builtin_chkp_function): New.
1478 (chkp_bound_type): New.
1479 (chkp_bound_mode): New.
1480 (fn_abi_va_list_bounds_size): New.
1481 (load_bounds_for_arg): New.
1482 (store_bounds_for_arg): New.
1483 * targhooks.h (default_load_bounds_for_arg): New.
1484 (default_store_bounds_for_arg): New.
1485 (default_fn_abi_va_list_bounds_size): New.
1486 (default_chkp_bound_type): New.
1487 (default_chkp_bound_mode): New.
1488 (default_builtin_chkp_function): New.
1489 * targhooks.c (default_load_bounds_for_arg): New.
1490 (default_store_bounds_for_arg): New.
1491 (default_fn_abi_va_list_bounds_size): New.
1492 (default_chkp_bound_type): New.
1493 (default_chkp_bound_mode); New.
1494 (default_builtin_chkp_function): New.
1495 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1496 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1497 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1498 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1499 (TARGET_CHKP_BOUND_TYPE): New.
1500 (TARGET_CHKP_BOUND_MODE): New.
1501 * doc/tm.texi: Regenerated.
1502 * langhooks.h (lang_hooks): Add chkp_supported field.
1503 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1504 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1506 2013-10-29 Andrew Pinski <apinski@cavium.com>
1508 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1509 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1510 fails, combining the branches anyways.
1511 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1512 increases the number of combinings.
1513 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1515 2013-10-29 Mike Stump <mikestump@comcast.net>
1517 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1518 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1519 (make_vector_mode): Increase namebuf to 16.
1520 (emit_insn_modes_h): When processing BImode, don't
1521 also match partial int modes.
1522 (emit_class_narrowest_mode): Likewise.
1524 * config/bfin/bfin-modes.def: Add precision to PDI.
1525 * config/m32c/m32c-modes.def: Add precision to PSI.
1526 * config/msp430/msp430-modes.def: Add precision to PSI.
1527 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1528 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1530 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1533 * config/sh/sh.md (*addc): Rename existing variations to ...
1534 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1535 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1536 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1537 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1539 2013-10-29 Teresa Johnson <tejohnson@google.com>
1542 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1543 insanities when updating probabilities.
1545 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1547 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1548 prettyprinter for cgraph_node to reflect the conversion of the
1549 symtable types to a C++ class hierarchy: it now *is* a
1550 symtab_node_base, rather than having one (named "symbol").
1552 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1554 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
1555 __cilkrts_pop_frame. If matched, then return true for built-in
1557 (expand_builtin): Added BUILT_IN_CILK_DETACH and
1558 BUILT_IN_CILK_POP_FRAME case.
1559 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
1560 (lhs_cilk_detect_spawn): Likewise.
1561 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
1562 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
1563 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
1564 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
1565 (LANG_HOOKS_CILKPLUS): Likewise.
1566 * tree.h (CILK_SPAWN_FN): Likewise.
1567 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
1568 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
1569 (OBJS): Added cilk-common.o.
1570 (BUILTINS_DEF): Added cilk-builtins.def.
1571 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
1572 (lhd_cilk_detect_spawn): Likewise.
1573 * langhooks.h (lang_hooks_for_cilkplus): New struct.
1574 (struct lang_hooks): Added new field called "cilkplus."
1575 * cilk-common.c: New file.
1577 * cilk-builtins.def: Likewise.
1578 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
1579 "__cilk" macro and set it to 200.
1580 * function.h (struct function::cilk_frame_decl): New field.
1581 (struct function::is_cilk_function): Likewise.
1582 (struct function::calls_cilk_spawn): Likewise.
1583 * gimplify.c (gimplify_call_expr): Added a check if the function call
1584 being gimplified is a spawn detach point. If so, then add pop_frame
1585 and detach function calls.
1586 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
1587 for gimplifying _Cilk_spawn and _Cilk_sync statements.
1588 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
1589 function. If so, added a _Cilk_sync and gimplified it.
1590 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
1591 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
1592 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
1594 (can_inline_edge_p): Prevent inling of spawnee function.
1595 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
1596 for functions that use Cilk keywords.
1597 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
1598 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
1599 CILK_SYNC_STMT cases.
1600 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
1602 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
1603 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1604 * passes.texi (Cilk Keywords): New section that describes the compiler
1605 code changes for handling Cilk Keywords.
1607 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1609 Patch autogenerated by refactor_symtab.py from
1610 https://github.com/davidmalcolm/gcc-refactoring-scripts
1611 revision 58bb219cc090b2f4516a9297d868c245495ee622
1613 * asan.c (asan_finish_file): Update for conversion of symtab types to
1614 a true class hierarchy.
1615 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1616 * cgraph.c (cgraph_get_body): Likewise.
1617 (cgraph_get_create_real_symbol_node): Likewise.
1618 (verify_cgraph_node): Likewise.
1619 (verify_edge_corresponds_to_fndecl): Likewise.
1620 (verify_edge_count_and_frequency): Likewise.
1621 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1622 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1623 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1624 (cgraph_node_cannot_return): Likewise.
1625 (cgraph_set_pure_flag_1): Likewise.
1626 (cgraph_set_const_flag_1): Likewise.
1627 (cgraph_set_nothrow_flag_1): Likewise.
1628 (cgraph_make_node_local_1): Likewise.
1629 (cgraph_for_node_and_aliases): Likewise.
1630 (cgraph_for_node_thunks_and_aliases): Likewise.
1631 (cgraph_node_can_be_local_p): Likewise.
1632 (cgraph_node_cannot_be_local_p_1): Likewise.
1633 (cgraph_function_body_availability): Likewise.
1634 (dump_cgraph_node): Likewise.
1635 (cgraph_rtl_info): Likewise.
1636 (cgraph_mark_address_taken_node): Likewise.
1637 (cgraph_remove_node): Likewise.
1638 (cgraph_release_function_body): Likewise.
1639 (cgraph_update_edges_for_call_stmt_node): Likewise.
1640 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1641 (cgraph_make_edge_direct): Likewise.
1642 (cgraph_resolve_speculation): Likewise.
1643 (cgraph_speculative_call_info): Likewise.
1644 (cgraph_turn_edge_to_speculative): Likewise.
1645 (cgraph_create_edge_1): Likewise.
1646 (cgraph_set_call_stmt): Likewise.
1647 (cgraph_node_for_asm): Likewise.
1648 (cgraph_add_thunk): Likewise.
1649 (cgraph_same_body_alias): Likewise.
1650 (cgraph_create_function_alias): Likewise.
1651 (cgraph_create_node): Likewise.
1652 (cgraph_create_empty_node): Likewise.
1653 (record_function_versions): Likewise.
1654 (used_from_object_file_p): Likewise.
1655 * cgraph.h (symtab_can_be_discarded): Likewise.
1656 (symtab_real_symbol_p): Likewise.
1657 (cgraph_mark_force_output_node): Likewise.
1658 (cgraph_edge_recursive_p): Likewise.
1659 (symtab_alias_target): Likewise.
1660 (varpool_all_refs_explicit_p): Likewise.
1661 (varpool_can_remove_if_no_refs): Likewise.
1662 (cgraph_only_called_directly_or_aliased_p): Likewise.
1663 (cgraph_next_function_with_gimple_body): Likewise.
1664 (cgraph_first_function_with_gimple_body): Likewise.
1665 (cgraph_function_with_gimple_body_p): Likewise.
1666 (cgraph_next_function): Likewise.
1667 (cgraph_first_function): Likewise.
1668 (cgraph_next_defined_function): Likewise.
1669 (cgraph_first_defined_function): Likewise.
1670 (varpool_next_defined_variable): Likewise.
1671 (varpool_first_defined_variable): Likewise.
1672 (varpool_next_static_initializer): Likewise.
1673 (varpool_first_static_initializer): Likewise.
1674 (varpool_next_variable): Likewise.
1675 (varpool_first_variable): Likewise.
1676 (varpool_node_name): Likewise.
1677 (varpool): Likewise.
1679 (is_a_helper <varpool_node>::test): Likewise.
1680 (is_a_helper <cgraph_node>::test): Likewise.
1681 (varpool_variable_node): Likewise.
1682 (cgraph_function_or_thunk_node): Likewise.
1683 (varpool_alias_target): Likewise.
1684 (cgraph_alias_target): Likewise.
1685 (cgraph_node_name): Likewise.
1686 (varpool_node_asm_name): Likewise.
1687 (cgraph_node_asm_name): Likewise.
1688 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1689 (cgraph_rebuild_references): Likewise.
1690 (rebuild_cgraph_edges): Likewise.
1691 (record_eh_tables): Likewise.
1692 (build_cgraph_edges): Likewise.
1693 (mark_store): Likewise.
1694 (mark_load): Likewise.
1695 (mark_address): Likewise.
1696 (record_type_list): Likewise.
1697 (record_reference): Likewise.
1698 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1699 (cgraph_materialize_clone): Likewise.
1700 (cgraph_function_versioning): Likewise.
1701 (cgraph_copy_node_for_versioning): Likewise.
1702 (update_call_expr): Likewise.
1703 (cgraph_find_replacement_node): Likewise.
1704 (cgraph_create_virtual_clone): Likewise.
1705 (cgraph_clone_node): Likewise.
1706 * cgraphunit.c (compile): Likewise.
1707 (output_weakrefs): Likewise.
1708 (output_in_order): Likewise.
1709 (expand_function): Likewise.
1710 (assemble_thunks_and_aliases): Likewise.
1711 (expand_thunk): Likewise.
1712 (mark_functions_to_output): Likewise.
1713 (handle_alias_pairs): Likewise.
1714 (analyze_functions): Likewise.
1715 (walk_polymorphic_call_targets): Likewise.
1716 (varpool_finalize_decl): Likewise.
1717 (process_function_and_variable_attributes): Likewise.
1718 (cgraph_process_same_body_aliases): Likewise.
1719 (analyze_function): Likewise.
1720 (cgraph_add_new_function): Likewise.
1721 (cgraph_finalize_function): Likewise.
1722 (referred_to_p): Likewise.
1723 (cgraph_reset_node): Likewise.
1724 (cgraph_process_new_functions): Likewise.
1725 (enqueue_node): Likewise.
1726 (decide_is_symbol_needed): Likewise.
1727 * coverage.c (coverage_compute_profile_id): Likewise.
1728 * dbxout.c (dbxout_expand_expr): Likewise.
1729 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1730 (reference_to_unused): Likewise.
1731 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1732 * gimplify.c (unvisit_body): Likewise.
1733 (unshare_body): Likewise.
1734 * ipa-cp.c (ipcp_generate_summary): Likewise.
1735 (ipcp_decision_stage): Likewise.
1736 (identify_dead_nodes): Likewise.
1737 (decide_whether_version_node): Likewise.
1738 (decide_about_value): Likewise.
1739 (perhaps_add_new_callers): Likewise.
1740 (create_specialized_node): Likewise.
1741 (update_profiling_info): Likewise.
1742 (ipcp_propagate_stage): Likewise.
1743 (estimate_local_effects): Likewise.
1744 (good_cloning_opportunity_p): Likewise.
1745 (devirtualization_time_bonus): Likewise.
1746 (propagate_constants_accross_call): Likewise.
1747 (initialize_node_lattices): Likewise.
1748 (ipcp_cloning_candidate_p): Likewise.
1749 (determine_versionability): Likewise.
1750 (print_all_lattices): Likewise.
1751 (print_lattice): Likewise.
1752 (ipcp_discover_new_direct_edges): Likewise.
1753 * ipa-devirt.c (ipa_devirt): Likewise.
1754 (likely_target_p): Likewise.
1755 (update_type_inheritance_graph): Likewise.
1756 (possible_polymorphic_call_target_p): Likewise.
1757 (dump_possible_polymorphic_call_targets): Likewise.
1758 (devirt_variable_node_removal_hook): Likewise.
1759 (record_binfo): Likewise.
1760 (maybe_record_node): Likewise.
1761 (build_type_inheritance_graph): Likewise.
1762 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1763 (inline_generate_summary): Likewise.
1764 (inline_analyze_function): Likewise.
1765 (do_estimate_growth): Likewise.
1766 (simple_edge_hints): Likewise.
1767 (estimate_node_size_and_time): Likewise.
1768 (estimate_edge_devirt_benefit): Likewise.
1769 (compute_inline_parameters): Likewise.
1770 (estimate_function_body_sizes): Likewise.
1771 (compute_bb_predicates): Likewise.
1772 (initialize_inline_failed): Likewise.
1773 (dump_inline_summary): Likewise.
1774 (dump_inline_edge_summary): Likewise.
1775 * ipa-inline-transform.c (inline_transform): Likewise.
1776 (preserve_function_body_p): Likewise.
1777 (save_inline_function_body): Likewise.
1778 (inline_call): Likewise.
1779 (clone_inlined_nodes): Likewise.
1780 (can_remove_node_now_p): Likewise.
1781 (can_remove_node_now_p_1): Likewise.
1782 * ipa-inline.c (early_inliner): Likewise.
1783 (early_inline_small_functions): Likewise.
1784 (inline_always_inline_functions): Likewise.
1785 (ipa_inline): Likewise.
1786 (flatten_function): Likewise.
1787 (inline_small_functions): Likewise.
1788 (speculation_useful_p): Likewise.
1789 (recursive_inlining): Likewise.
1790 (update_caller_keys): Likewise.
1791 (reset_edge_caches): Likewise.
1792 (update_edge_key): Likewise.
1793 (edge_badness): Likewise.
1794 (relative_time_benefit): Likewise.
1795 (want_inline_self_recursive_call_p): Likewise.
1796 (want_inline_small_function_p): Likewise.
1797 (want_early_inline_function_p): Likewise.
1798 (num_calls): Likewise.
1799 (can_early_inline_edge_p): Likewise.
1800 (can_inline_edge_p): Likewise.
1801 (report_inline_failed_reason): Likewise.
1802 * ipa-profile.c (ipa_profile): Likewise.
1803 (ipa_propagate_frequency): Likewise.
1804 (ipa_propagate_frequency_1): Likewise.
1805 (ipa_profile_generate_summary): Likewise.
1806 * ipa-prop.c (ipcp_transform_function): Likewise.
1807 (read_replacements_section): Likewise.
1808 (ipa_prop_read_section): Likewise.
1809 (ipa_modify_call_arguments): Likewise.
1810 (ipa_print_node_params): Likewise.
1811 (propagate_controlled_uses): Likewise.
1812 (update_indirect_edges_after_inlining): Likewise.
1813 (remove_described_reference): Likewise.
1814 (ipa_make_edge_direct_to_target): Likewise.
1815 (ipa_analyze_node): Likewise.
1816 (ipa_analyze_params_uses): Likewise.
1817 (ipa_compute_jump_functions): Likewise.
1818 (ipa_get_callee_param_type): Likewise.
1819 (ipa_print_node_jump_functions): Likewise.
1820 (ipa_initialize_node_params): Likewise.
1821 (ipa_populate_param_decls): Likewise.
1822 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1823 (write_agg_replacement_chain): Likewise.
1824 (ipa_write_node_info): Likewise.
1825 (ipa_edge_duplication_hook): Likewise.
1826 (try_decrement_rdesc_refcount): Likewise.
1827 * ipa-pure-const.c (propagate_nothrow): Likewise.
1828 (propagate_pure_const): Likewise.
1829 (pure_const_read_summary): Likewise.
1830 (pure_const_write_summary): Likewise.
1831 (analyze_function): Likewise.
1832 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1833 (ipa_ref_referring_ref_list): Likewise.
1834 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1835 (ipa_remove_stmt_references): Likewise.
1836 (ipa_find_reference): Likewise.
1837 (ipa_dump_referring): Likewise.
1838 (ipa_dump_references): Likewise.
1839 (ipa_record_reference): Likewise.
1840 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1841 (ipa_reference_write_optimization_summary): Likewise.
1842 (write_node_summary_p): Likewise.
1843 (propagate): Likewise.
1844 (read_write_all_from_decl): Likewise.
1845 (generate_summary): Likewise.
1846 (analyze_function): Likewise.
1847 (propagate_bits): Likewise.
1848 (ipa_reference_get_not_written_global): Likewise.
1849 (ipa_reference_get_not_read_global): Likewise.
1850 * ipa-split.c (execute_split_functions): Likewise.
1851 (split_function): Likewise.
1852 * ipa-utils.c (ipa_merge_profiles): Likewise.
1853 (dump_cgraph_node_set): Likewise.
1854 (ipa_reverse_postorder): Likewise.
1855 (ipa_edge_within_scc): Likewise.
1856 (ipa_get_nodes_in_cycle): Likewise.
1857 (ipa_free_postorder_info): Likewise.
1858 (ipa_reduced_postorder): Likewise.
1859 (searchc): Likewise.
1860 (recursive_call_p): Likewise.
1861 * ipa.c (ipa_cdtor_merge): Likewise.
1862 (record_cdtor_fn): Likewise.
1863 (function_and_variable_visibility): Likewise.
1864 (varpool_externally_visible_p): Likewise.
1865 (cgraph_externally_visible_p): Likewise.
1866 (comdat_can_be_unshared_p): Likewise.
1867 (comdat_can_be_unshared_p_1): Likewise.
1868 (address_taken_from_non_vtable_p): Likewise.
1869 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1870 (symtab_remove_unreachable_nodes): Likewise.
1871 (walk_polymorphic_call_targets): Likewise.
1872 (process_references): Likewise.
1873 (enqueue_node): Likewise.
1874 (has_addr_references_p): Likewise.
1875 (cgraph_non_local_node_p_1): Likewise.
1876 * is-a.h (varpool_analyze_node): Likewise.
1877 * lto-cgraph.c (input_symtab): Likewise.
1878 (merge_profile_summaries): Likewise.
1879 (input_cgraph_1): Likewise.
1880 (input_edge): Likewise.
1881 (input_varpool_node): Likewise.
1882 (input_node): Likewise.
1883 (input_overwrite_node): Likewise.
1884 (compute_ltrans_boundary): Likewise.
1885 (output_refs): Likewise.
1886 (lto_output_varpool_node): Likewise.
1887 (lto_output_node): Likewise.
1888 (reachable_from_other_partition_p): Likewise.
1889 (referenced_from_other_partition_p): Likewise.
1890 (lto_output_edge): Likewise.
1891 (output_node_opt_summary): Likewise.
1892 (add_node_to): Likewise.
1893 (reachable_from_this_partition_p): Likewise.
1894 (lto_set_symtab_encoder_in_partition): Likewise.
1895 (lto_symtab_encoder_in_partition_p): Likewise.
1896 (lto_set_symtab_encoder_encode_initializer): Likewise.
1897 (lto_symtab_encoder_encode_initializer_p): Likewise.
1898 (lto_set_symtab_encoder_encode_body): Likewise.
1899 (lto_symtab_encoder_encode_body_p): Likewise.
1900 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1902 * lto-streamer-in.c (lto_read_body): Likewise.
1903 (fixup_call_stmt_edges): Likewise.
1904 (fixup_call_stmt_edges_1): Likewise.
1905 * lto-streamer-out.c (produce_symtab): Likewise.
1906 (output_symbol_p): Likewise.
1907 (write_symbol): Likewise.
1908 (lto_output): Likewise.
1909 (copy_function): Likewise.
1910 (output_function): Likewise.
1911 * passes.c (function_called_by_processed_nodes_p): Likewise.
1912 (ipa_write_optimization_summaries): Likewise.
1913 (ipa_write_summaries): Likewise.
1914 (do_per_function_toporder): Likewise.
1915 (do_per_function): Likewise.
1916 (dump_passes): Likewise.
1917 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1918 (symtab_nonoverwritable_alias): Likewise.
1919 (symtab_nonoverwritable_alias_1): Likewise.
1920 (symtab_for_node_and_aliases): Likewise.
1921 (symtab_resolve_alias): Likewise.
1922 (fixup_same_cpp_alias_visibility): Likewise.
1923 (symtab_alias_ultimate_target): Likewise.
1924 (symtab_used_from_object_file_p): Likewise.
1925 (verify_symtab_base): Likewise.
1926 (dump_symtab_base): Likewise.
1927 (symtab_node_name): Likewise.
1928 (symtab_node_asm_name): Likewise.
1929 (symtab_dissolve_same_comdat_group_list): Likewise.
1930 (symtab_add_to_same_comdat_group): Likewise.
1931 (symtab_unregister_node): Likewise.
1932 (symtab_insert_node_to_hashtable): Likewise.
1933 (symtab_register_node): Likewise.
1934 (unlink_from_assembler_name_hash): Likewise.
1935 (insert_to_assembler_name_hash): Likewise.
1936 (eq_assembler_name): Likewise.
1937 (hash_node_by_assembler_name): Likewise.
1938 (eq_node): Likewise.
1939 (hash_node): Likewise.
1940 * toplev.c (wrapup_global_declaration_2): Likewise.
1941 * trans-mem.c (ipa_tm_execute): Likewise.
1942 (ipa_tm_transform_clone): Likewise.
1943 (ipa_tm_transform_transaction): Likewise.
1944 (ipa_tm_transform_calls_redirect): Likewise.
1945 (ipa_tm_insert_gettmclone_call): Likewise.
1946 (ipa_tm_insert_irr_call): Likewise.
1947 (ipa_tm_create_version): Likewise.
1948 (ipa_tm_create_version_alias): Likewise.
1949 (ipa_tm_mark_forced_by_abi_node): Likewise.
1950 (ipa_tm_mark_force_output_node): Likewise.
1951 (ipa_tm_diagnose_tm_safe): Likewise.
1952 (ipa_tm_mayenterirr_function): Likewise.
1953 (ipa_tm_scan_irr_function): Likewise.
1954 (ipa_tm_note_irrevocable): Likewise.
1955 (ipa_tm_scan_calls_clone): Likewise.
1956 (get_cg_data): Likewise.
1957 * tree-eh.c (tree_could_trap_p): Likewise.
1958 * tree-emutls.c (ipa_lower_emutls): Likewise.
1959 (create_emultls_var): Likewise.
1960 (lower_emutls_function_body): Likewise.
1961 (gen_emutls_addr): Likewise.
1962 (emutls_decl): Likewise.
1963 (new_emutls_decl): Likewise.
1964 * tree-inline.c (tree_function_versioning): Likewise.
1965 (optimize_inline_calls): Likewise.
1966 (expand_call_inline): Likewise.
1967 (estimate_num_insns): Likewise.
1968 (copy_bb): Likewise.
1969 (delete_unreachable_blocks_update_callgraph): Likewise.
1970 * tree-nested.c (gimplify_all_functions): Likewise.
1971 (create_nesting_tree): Likewise.
1972 (check_for_nested_with_variably_modified): Likewise.
1973 * tree-pretty-print.c (dump_function_header): Likewise.
1974 * tree-profile.c (tree_profiling): Likewise.
1975 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1976 (modify_function): Likewise.
1977 (convert_callers): Likewise.
1978 (convert_callers_for_node): Likewise.
1979 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1980 (associate_varinfo_to_alias): Likewise.
1981 (create_variable_info_for): Likewise.
1982 (get_constraint_for_ssa_var): Likewise.
1983 * tree-vectorizer.c (increase_alignment): Likewise.
1984 * tree.c (find_decls_types_in_var): Likewise.
1985 (find_decls_types_in_node): Likewise.
1986 (free_lang_data_in_decl): Likewise.
1987 * value-prof.c (gimple_ic_transform): Likewise.
1988 (gimple_ic): Likewise.
1989 (check_ic_target): Likewise.
1990 (init_node_map): Likewise.
1991 * varasm.c (decl_binds_to_current_def_p): Likewise.
1992 (default_binds_local_p_1): Likewise.
1993 (dump_tm_clone_pairs): Likewise.
1994 (assemble_alias): Likewise.
1995 (find_decl): Likewise.
1996 (mark_decl_referenced): Likewise.
1997 * varpool.c (varpool_for_node_and_aliases): Likewise.
1998 (varpool_extra_name_alias): Likewise.
1999 (varpool_create_variable_alias): Likewise.
2000 (add_new_static_var): Likewise.
2001 (varpool_finalize_named_section_flags): Likewise.
2002 (varpool_remove_unreferenced_decls): Likewise.
2003 (enqueue_node): Likewise.
2004 (varpool_assemble_decl): Likewise.
2005 (assemble_aliases): Likewise.
2006 (varpool_analyze_node): Likewise.
2007 (cgraph_variable_initializer_availability): Likewise.
2008 (varpool_add_new_variable): Likewise.
2009 (ctor_for_folding): Likewise.
2010 (dump_varpool_node): Likewise.
2011 (varpool_remove_initializer): Likewise.
2012 (varpool_remove_node): Likewise.
2013 (varpool_node_for_decl): Likewise.
2014 (varpool_create_empty_node): Likewise.
2015 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
2016 (ix86_get_function_versions_dispatcher): Likewise.
2018 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2020 * cgraph.h (symtab_node_base): Convert to a class;
2021 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
2022 chain_next/prev from symtab_node_def.
2023 (cgraph_node): Inherit from symtab_node; add GTY option
2024 tag ("SYMTAB_FUNCTION").
2025 (varpool_node): Inherit from symtab_node; add GTY option
2026 tag ("SYMTAB_VARIABLE").
2027 (symtab_node_def): Remove.
2028 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
2029 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
2030 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
2031 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
2033 * ipa-ref.h (symtab_node_def): Drop.
2034 (symtab_node): Change underlying type from symtab_node_def to
2036 (const_symtab_node): Likwise.
2038 * is-a.h: Update examples in comment.
2040 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
2041 (assembler_name_hash): Likewise.
2043 2013-10-29 Martin Liska <marxin.liska@gmail.com>
2045 * doc/tree-ssa.texi (gimple_phi_result): Document.
2046 (gimple_phi_num_args, gimple_phi_arg): Likewise.
2047 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
2048 (PHI_RESULT, PHI_NUM_ARGS): Remove.
2049 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
2051 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2053 * expr.h: Revert change and include tree-core.h.
2054 * rtl.h: Revert change and don't include tree-core.h.
2056 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2058 * config/darwin.c: Include gimple.h.
2059 * config/i386/winnt.c: Likewise.
2061 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
2063 PR tree-optimization/19831
2064 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
2066 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2068 * tree-outof-ssa.h: Remove include files.
2069 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
2071 * tree-ssa-coalesce.c: Likewise.
2072 * cfgexpand.c: Likewise.
2073 * tree-ssa-ter.c: Likewise.
2074 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
2075 * lto-streamer.h: Likewise.
2076 * cgraphbuild.c: Add gimple.h to include list.
2077 * data-streamer-in.c: Likewise.
2078 * ipa-cp.c: Likewise.
2079 * tree-streamer.c: Likewise.
2080 * lto-compress.c: Likewise.
2081 * ipa-reference.c: Likewise.
2082 * data-streamer-out.c: Likewise.
2083 * lto-cgraph.c: Likewise.
2084 * cgraphclones.c: Likewise.
2085 * ipa-utils.c: Likewise.
2086 * data-streamer.c: Likewise.
2087 * ipa-split.c: Likewise.
2088 * lto-section-in.c: Likewise.
2089 * tree-streamer-out.c: Likewise.
2090 * ipa-prop.c: Likewise.
2091 * tree-streamer-in.c: Likewise.
2092 * symtab.c: Likewise.
2093 * opts-global.c: Likewise.
2094 * lto-opts.c: Likewise.
2095 * lto-section-out.c: Likewise.
2096 * lto-streamer.c: Likewise.
2097 * rtl.h: Add tree-core.h to include list.
2098 * expr.h: Remove tree-core.h from include list.
2099 * gimple.h: Likewise.
2100 * ipa-utils.h: Likewise.
2101 * streamer-hooks.h: Likewise.
2102 * streamer-hooks.c: Include input.h.
2104 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2106 * config/arm/arm.c (cortexa7_extra_costs): New table.
2107 (arm_cortex_a7_tune): New.
2108 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
2110 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
2112 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
2114 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2116 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
2117 to use autogenerated markers for a class-hierarchy, every class
2118 must have a GTY marker.
2119 * gengtype.h (struct type): Add linked list of subclasses to
2120 the "s" member of the union.
2121 (add_subclass): New decl.
2122 * gengtype-state.c (read_state_struct_type): Set up subclass
2124 * gengtype.c (get_ultimate_base_class): New.
2125 (add_subclass): New.
2126 (new_structure): Set up subclass linked list.
2127 (set_gc_used_type): Propagate usage information to subclasses.
2128 (output_mangled_typename): Use get_ultimate_base_class.
2129 (walk_subclasses): Use the subclass linked list, avoiding an
2130 O(N^2) when writing out all types.
2131 (walk_type): Issue an error if the base class is missing a tag,
2132 rather than generating bogus C code. Add a gcc_unreachable
2133 default case, in case people omit tags from concrete subclasses,
2134 or get the values wrong.
2135 (write_func_for_structure): Issue an error for subclasses for
2136 which the base doesn't have a "desc", since otherwise the
2137 autogenerated routines for the base would silently fail to visit
2138 any subclass fields.
2139 (write_root): Use get_ultimate_base_class, tweaking constness of
2140 tp to match that function's signature.
2142 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2144 * doc/gty.texi (GTY Options): Add note about inheritance to
2145 description of desc and tag.
2146 (Inheritance and GTY): New.
2148 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2150 * gengtype-parse.c (opts_have): Drop "static" so that
2151 we can use this from gengtype.c.
2152 * gengtype.c (set_gc_used_type): Mark any base class as used;
2153 update field traversal to visit inherited fields.
2154 (output_mangled_typename): Convert references to classes within
2155 an inheritance hierarchy to reference the ultimate base class,
2156 since only it will have gt_ functions.
2157 (get_string_option): New.
2158 (walk_subclasses): New.
2159 (walk_type): Treat GTY structs that have a "desc" as being the
2160 root of an inheritance hierarchy. Generate a switch on it
2161 within the marking function which walks all subclasses, adding
2162 cases for them via walk_subclasses. For subclasses, visit all
2163 fields of the type (including inherited ones).
2164 (write_func_for_structure): Don't write fns for subclasses, only
2165 for the ultimate base class within an inheritance hierarchy.
2166 Subclasses-marking will be handled by the base class marking functions.
2167 (write_types): Likewise.
2168 (write_local_func_for_structure): Likewise.
2169 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
2170 a "tag" option (and are thus concrete subclasses).
2171 (write_root): Use the marker function for the ultimate base class.
2172 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
2173 (opts_have): Add declaration.
2175 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
2177 * lra-spills.c (lra_final_code_change): Remove useless move insns
2178 originated from moves of pseudos.
2180 2013-10-28 Jeff Law <law@redhat.com>
2182 * lower-subreg.c (resolve_simple_move): Fix comment typo.
2184 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
2186 * df-scan.c (df_collection_rec): Adjust.
2187 (copy_defs): New constant.
2188 (copy_uses): Likewise.
2189 (copy_eq_uses): Likewise.
2190 (copy_mw): Likewise.
2191 (copy_all): Likewise.
2192 (df_insn_rescan): Adjust.
2193 (df_notes_rescan): Likewise.
2194 (df_swap_refs): Likewise.
2195 (df_sort_and_compress_refs): Likewise.
2196 (df_sort_and_compress_mws): Likewise.
2197 (df_install_refs): Likewise.
2198 (df_install_mws): Likewise.
2199 (df_refs_add_to_chains): Add flags parameter controlling which vectors
2201 (df_bb_refs_record): Adjust.
2202 (df_record_entry_block_defs): Likewise.
2203 (df_record_exit_block_defs): Likewise.
2204 (df_refs_verify): Likewise.
2205 (df_mws_verify): Likewise.
2206 (df_insn_refs_verify): Likewise.
2207 (df_bb_verify): Likewise.
2208 * ipa-pure-const.c (finish_state): Remove.
2209 (propagate): Adjust.
2210 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
2211 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
2212 var-tracking.c: Adjust.
2213 * vec.c (stack_vecs): Remove.
2214 (register_stack_vec): Likewise.
2215 (stack_vec_register_index): Likewise.
2216 (unregister_stack_vec): Likewise.
2217 * vec.h (struct va_stack): Remove.
2218 (struct vec<T, A, vl_ptr>): Specialize as
2219 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
2220 allocation strategy compatable with the vl_ptr layout.
2221 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
2222 specialization anyway.
2223 (class stack_vec): New class.
2224 (vec_stack_alloc): Remove.
2225 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
2227 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2228 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2229 Sergey Lega <sergey.s.lega@intel.com>
2230 Anna Tikhonova <anna.tikhonova@intel.com>
2231 Ilya Tocar <ilya.tocar@intel.com>
2232 Andrey Turetskiy <andrey.turetskiy@intel.com>
2233 Ilya Verbin <ilya.verbin@intel.com>
2234 Kirill Yukhin <kirill.yukhin@intel.com>
2235 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2237 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
2238 (*prefetch_avx512pf_<mode>): New.
2239 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
2240 (avx512f_maskcmp<mode>3): Ditto.
2241 (vashrv16si3): Ditto.
2243 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2244 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2245 Sergey Lega <sergey.s.lega@intel.com>
2246 Anna Tikhonova <anna.tikhonova@intel.com>
2247 Ilya Tocar <ilya.tocar@intel.com>
2248 Andrey Turetskiy <andrey.turetskiy@intel.com>
2249 Ilya Verbin <ilya.verbin@intel.com>
2250 Kirill Yukhin <kirill.yukhin@intel.com>
2251 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2253 * config/i386/i386.md (any_truncate): New.
2254 (trunsuffix): Ditto.
2255 * config/i386/predicates.md (const_8_to_9_operand): New.
2256 (const_10_to_11_operand): Ditto.
2257 (const_12_to_13_operand): Ditto.
2258 (const_14_to_15_operand): Ditto.
2259 (const_16_to_19_operand): Ditto.
2260 (const_20_to_23_operand): Ditto.
2261 (const_24_to_27_operand): Ditto.
2262 (const_28_to_31_operand): Ditto.
2263 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
2264 (cvtusi2<ssescalarmodesuffix>32): New.
2265 (cvtusi2<ssescalarmodesuffix>64): Ditto.
2266 (ufloatv16siv16sf2): Ditto.
2267 (avx512f_fix_notruncv16sfv16si): Ditto.
2268 (avx512f_ufix_notruncv16sfv16si): Ditto.
2269 (avx512f_vcvtss2usi): Ditto.
2270 (avx512f_vcvtss2usiq): Ditto.
2271 (avx512f_vcvttss2usi): Ditto.
2272 (avx512f_vcvttss2usiq): Ditto.
2273 (avx512f_vcvtsd2usi): Ditto.
2274 (avx512f_vcvtsd2usiq): Ditto.
2275 (avx512f_vcvttsd2usi): Ditto.
2276 (avx512f_vcvttsd2usiq): Ditto.
2277 (ufloatv8siv8df): Ditto.
2278 (avx512f_cvtdq2pd512_2): Ditto.
2279 (avx512f_cvtpd2dq512): Ditto.
2280 (avx512f_ufix_notruncv8dfv8si): Ditto.
2281 (avx512f_cvtpd2ps512): Ditto.
2282 (vec_unpacks_lo_v16sf): Ditto.
2283 (vec_unpacks_hi_v16sf): Ditto.
2284 (vec_unpacks_float_hi_v16si): Ditto.
2285 (vec_unpacks_float_lo_v16si): Ditto.
2286 (avx512f_unpckhps512): Ditto.
2287 (avx512f_unpcklps512): Ditto.
2288 (avx512f_movshdup512): Ditto.
2289 (avx512f_movsldup512): Ditto.
2290 (vec_extract_lo_v32hi): Ditto.
2291 (vec_extract_hi_v32hi): Ditto.
2292 (vec_extract_lo_v64qi): Ditto.
2293 (vec_extract_hi_v64qi): Ditto.
2294 (avx512f_unpckhpd512): Ditto.
2295 (avx512f_movddup512): Ditto.
2296 (avx512f_unpcklpd512): Ditto.
2297 (*avx512f_unpcklpd512): Ditto.
2298 (avx512f_shufps512_1): Ditto.
2299 (avx512f_shufpd512_1): Ditto.
2300 (avx512f_interleave_highv8di): Ditto.
2301 (avx512f_interleave_lowv8di): Ditto.
2302 (PMOV_DST_MODE): Ditto.
2303 (pmov_src_mode): Ditto.
2304 (pmov_src_lower): Ditto.
2306 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
2307 (*avx512f_<code>v8div16qi2): Ditto.
2308 (*avx512f_<code>v8div16qi2_store): Ditto.
2309 (vec_widen_umult_even_v16si): Ditto.
2310 (*vec_widen_umult_even_v16si): Ditto.
2311 (vec_widen_smult_even_v16si): Ditto.
2312 (*vec_widen_smult_even_v16si): Ditto.
2313 (avx512f_interleave_highv16si): Ditto.
2314 (avx512f_interleave_lowv16si): Ditto.
2315 (avx512f_<code>v16qiv16si2): Ditto.
2316 (avx512f_<code>v16hiv16si2): Ditto.
2317 (avx512f_<code>v8qiv8di2): Ditto.
2318 (avx512f_<code>v8hiv8di2): Ditto.
2319 (avx512f_<code>v8siv8di2): Ditto.
2320 (avx512cd_maskb_vec_dupv8di): Ditto.
2321 (avx512cd_maskw_vec_dupv16si): Ditto.
2322 (avx512f_vcvtph2ps512): Ditto.
2323 (avx512f_vcvtps2ph512): Ditto.
2324 (VEC_EXTRACT_MODE): Extened with wider modes.
2325 (VEC_PERM_AVX2): Ditto.
2326 (VEC_PERM_CONST): Ditto.
2328 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
2330 * config/arc/arc.c (arc_ccfsm_post_advance):
2331 Add comment about TYPE_RETURN.
2333 2013-10-28 Bin Cheng <bin.cheng@arm.com>
2335 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
2336 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
2337 (strip_offset): Convert offset to unsigned number.
2339 2013-10-27 Tom de Vries <tom@codesourcery.com>
2341 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
2342 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
2344 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
2346 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
2347 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
2348 (TARGET_BIG_ENDIAN): New macro.
2349 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
2350 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
2351 * config/sh/sh.c: Likewise.
2352 * config/sh/sh.h: Likewise.
2354 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
2356 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
2357 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
2359 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2361 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
2363 (cmpstr_t, cmpstrsi): Fix formatting.
2365 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2368 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
2369 addressing, do not use general_operand for memory operands.
2371 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
2374 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2375 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2377 2013-10-26 Jeff Law <law@redhat.com>
2379 * predict.c (PRED_MUDFLAP): Remove.
2380 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
2382 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
2383 (OBJS): Remove tree-nomudflap.o
2384 (GTFILES): Remove tree-mudflap.c
2385 * builtins.c (expand_builtin_alloc): Remove mudflap support.
2386 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2387 (mfwrap_spec, mflib_spec): Likewise.
2388 (cpp_unique_options, cc1_options, static_specs): Likewise.
2389 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
2390 * passes.def: Likewise.
2391 * toplev.c (compile_file, process_options): Likewise.
2392 * tree-inline.c (copy_tree_r): Likewise.
2393 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
2394 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
2395 (build_constant_desc, output_constant_def_contents): Likewise.
2396 (categorize_decl_for_section): Likewise.
2397 * tree-mudflap.c: Removed.
2398 * tree-mudflap.h: Removed.
2399 * tree-nomudflap.c: Removed.
2400 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
2401 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
2402 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2403 * config/sol2.h (MFLIB_SPEC): Likewise.
2404 * doc/install.texi: Remove mudflap references.
2405 * doc/passes.texi: Similarly.
2406 * doc/sourcebuild.texi: Similarly.
2407 * doc/invoke.texi: Remove mudlfap related options.
2409 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2411 PR rtl-optimization/58759
2412 * lra-constraints.c (lra_constraints): Remove wrong condition to
2413 remove insn setting up an equivalent pseudo.
2415 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2417 * config/rs6000/rs6000-protos.h
2418 (rs6000_secondary_memory_needed_mode): New prototype.
2419 * config/rs6000/rs6000.c: Include ira.h.
2420 (TARGET_LRA_P): Redefine.
2421 (rs6000_legitimate_offset_address_p): Call
2422 legitimate_constant_pool_address_p in strict mode for LRA.
2423 (rs6000_legitimate_address_p): Ditto.
2424 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2425 (rs6000_emit_move): Add LRA version of code to generate load/store
2427 (rs6000_secondary_memory_needed_mode): New.
2428 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2429 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2430 constants, memory, and FP registers.
2431 (rs6000_lra_p): New.
2432 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2433 * config/rs6000/rs6000.opt (mlra): New option.
2434 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2436 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2438 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2439 has_single_use () and not do the conversion if has_single_use ()
2440 returns false for the multiplication result.
2442 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2444 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2445 to "NODE", since this works on a "tree", not an
2447 (CONSTANT_CLASS_P): Likewise.
2450 (INDIRECT_REF_P): Likewise.
2451 (REFERENCE_CLASS_P): Likewise.
2452 (COMPARISON_CLASS_P): Likewise.
2453 (UNARY_CLASS_P): Likewise.
2454 (BINARY_CLASS_P): Likewise.
2455 (STATEMENT_CLASS_P): Likewise.
2456 (VL_EXP_CLASS_P): Likewise.
2457 (EXPRESSION_CLASS_P): Likewise.
2458 (IS_TYPE_OR_DECL_P): Likewise.
2460 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2462 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2463 ADDR_EXPR in the defining statement.
2465 2013-10-25 Richard Biener <rguenther@suse.de>
2467 PR tree-optimization/58626
2468 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2469 anti_dd, output_dd and input_dd.
2470 (struct rdg_edge): Remove level and relation members.
2471 (RDGE_LEVEL, RDGE_RELATION): Remove.
2472 (dot_rdg_1): Adjust.
2473 (create_rdg_edge_for_ddr): Remove.
2474 (create_rdg_edges_for_scalar): Adjust.
2475 (create_edge_for_control_dependence): Likewise.
2476 (create_rdg_edges): Split into ...
2477 (create_rdg_flow_edges): ... this
2478 (create_rdg_cd_edges): ... and this.
2480 (build_rdg): Likewise, do not compute data dependences or
2482 (pg_add_dependence_edges): New function.
2484 (distribute_loop): First apply all non-dependence based
2485 partition mergings. Then compute dependences between partitions
2486 and merge and order partitions according to them.
2488 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2490 PR rtl-optimization/58831
2491 * alias.c (init_alias_analysis): At the beginning of each iteration,
2492 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2494 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2496 * recog.c (search_ofs): New static variable moved from...
2497 (peep2_find_free_register): ...here.
2498 (peephole2_optimize): Initialize it.
2500 2013-10-25 Tobias Burnus <burnus@net-b.de>
2502 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2504 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2506 * config/i386/i386.h (TARGET_MPX): New define.
2507 (TARGET_MPX_P): Ditto.
2509 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2511 * config/i386/constraints.md (B): New.
2514 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2515 * config/i386/i386-modes.def (BND32): New.
2517 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2518 * config/i386/i386.c (isa_opts): Add mmpx.
2519 (regclass_map): Add bound registers.
2520 (dbx_register_map): Likewise.
2521 (dbx64_register_map): Likewise.
2522 (svr4_dbx_register_map): Likewise.
2524 (ix86_option_override_internal): Support MPX ISA.
2525 (ix86_conditional_register_usage): Support bound registers.
2526 (print_reg): Likewise.
2527 (ix86_code_end): Add MPX bnd prefix.
2528 (output_set_got): Likewise.
2529 (ix86_output_call_insn): Likewise.
2530 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2531 (ix86_print_operand_punct_valid_p): Likewise.
2532 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2534 (ix86_class_likely_spilled_p): Add bound regs support.
2535 (ix86_hard_regno_mode_ok): Likewise.
2536 (x86_order_regs_for_local_alloc): Likewise.
2537 (ix86_bnd_prefixed_insn_p): New.
2538 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2539 (FIXED_REGISTERS): Add bound registers.
2540 (CALL_USED_REGISTERS): Likewise.
2541 (REG_ALLOC_ORDER): Likewise.
2542 (HARD_REGNO_NREGS): Likewise.
2544 (VALID_BND_REG_MODE): New.
2545 (FIRST_BND_REG): New.
2546 (LAST_BND_REG): New.
2547 (reg_class): Add BND_REGS.
2548 (REG_CLASS_NAMES): Likewise.
2549 (REG_CLASS_CONTENTS): Likewise.
2551 (ANY_BND_REG_P): New.
2553 (HI_REGISTER_NAMES): Add bound registers.
2554 * config/i386/i386.md (UNSPEC_BNDMK): New.
2555 (UNSPEC_BNDMK_ADDR): New.
2556 (UNSPEC_BNDSTX): New.
2557 (UNSPEC_BNDLDX): New.
2558 (UNSPEC_BNDLDX_ADDR): New.
2559 (UNSPEC_BNDCL): New.
2560 (UNSPEC_BNDCU): New.
2561 (UNSPEC_BNDCN): New.
2562 (UNSPEC_MPX_FENCE): New.
2565 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
2566 (length_immediate): Likewise.
2567 (prefix_0f): Likewise.
2569 (prefix_rep): Check for bnd prefix.
2570 (length_nobnd): New.
2571 (length): Use length_nobnd if specified.
2576 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
2579 (simple_return_internal): Likewise.
2580 (simple_return_pop_internal): Likewise.
2581 (*indirect_jump): Add MPX bnd prefix.
2582 (*tablejump_1): Likewise.
2583 (simple_return_internal_long): Likewise.
2584 (simple_return_indirect_internal): Likewise.
2588 (*mov<mode>_internal_mpx): New.
2589 (<mode>_<bndcheck>): New.
2590 (*<mode>_<bndcheck>): New.
2595 * config/i386/predicates.md (lea_address_operand): Rename to...
2596 (address_no_seg_operand): ... this.
2597 (address_mpx_no_base_operand): New.
2598 (address_mpx_no_index_operand): New.
2599 (bnd_mem_operator): New.
2600 * config/i386/i386.opt (mmpx): New.
2601 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
2602 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
2604 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2606 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2607 * tree.def (POINTER_BOUNDS_TYPE): New.
2608 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2609 (POINTER_BOUNDS_MODE): New.
2610 (make_pointer_bounds_mode): New.
2611 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2612 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2613 (layout_type): Support POINTER_BOUNDS_TYPE.
2614 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2615 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2616 (type_contains_placeholder_1): Likewise.
2617 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2618 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2619 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2621 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2623 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2624 mode of address computation when calling memory_address_addr_space.
2626 2013-08-24 Richard Henderson <rth@twiddle.net>
2629 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2630 instead of create_convert_operand_to.
2631 (maybe_emit_sync_lock_test_and_set): Likewise.
2632 (expand_atomic_compare_and_swap): Likewise.
2633 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2635 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2637 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2638 Change param_is to use the struct and not the pointer to the struct.
2640 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2642 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2643 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2644 Move to targhooks.c.
2645 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2646 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2647 Relocate from gimplify.c.
2648 * targhooks.h: Add 2 prototypes.
2649 * tree.h: Delete 2 prototypes.
2651 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2653 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2654 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2656 2013-10-24 Cong Hou <congh@google.com>
2658 * convert.c (convert_to_real): Guard those unsafe math function
2659 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2662 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2665 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2667 (cgraph_create_edge): Use the new argument.
2668 (cgraph_create_indirect_edge): Likewise.
2670 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2672 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2674 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2675 changes statep->state.
2677 2013-10-24 Tobias Burnus <burnus@net-b.de>
2680 * tree-cfg.c (replace_loop_annotate): New function.
2681 (execute_build_cfg): Call it.
2682 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2683 * internal-fn.c (expand_ANNOTATE): New function.
2684 * internal-fn.def (ANNOTATE): Define as new internal function.
2685 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2686 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2687 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2688 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2689 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2691 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2692 Marcus Shawcroft <marcus.shawcroft@arm.com>
2694 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2695 Special case reload SP+C into none GENERAL_REGS.
2697 2013-10-24 Michael Matz <matz@suse.de>
2699 * gengtype.c (is_file_equal): Check that files will be same length.
2701 2013-10-25 Christian Bruel <christian.bruel@st.com>
2703 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2704 * config/sh/t-sh (sh-mem.o): New rule.
2705 * config/sh/sh-mem.cc (expand_block_move): Moved here.
2706 (sh_expand_cmpstr): New function.
2707 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2708 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2709 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2710 (rotlhi3_8): Rename.
2712 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2714 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2715 * Makefile.in (CXXDEPMODE): Assign and change users.
2716 (CCDEPMODE): Delete.
2717 * configure: Regenerate.
2719 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2721 * gengtype-parse.c (require_without_advance): New.
2722 (type): For GTY-marked types that are not GTY((user)), parse any
2723 base classes, requiring them to be single-inheritance, and not
2724 be templates. For non-GTY-marked types and GTY((user)),
2725 continue to skip over any C++ inheritance specification.
2726 * gengtype-state.c (state_writer::write_state_struct_type):
2727 Write base class of type (if any).
2728 (read_state_struct_type): Read base class of type (if any).
2729 * gengtype.c (new_structure): Add a "base_class" parameter.
2730 (create_optional_field_): Update for new parameter to new_structure.
2731 (adjust_field_rtx_def): Likewise.
2732 (adjust_field_tree_exp): Likewise.
2733 * gengtype.h (struct type): Add "base_class" field to the s
2735 (new_structure): Add "base" parameter.
2737 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2740 * config/i386/i386.c (ix86_option_override_internal):
2741 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2742 (ix86_valid_target_attribute_tree):
2743 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2744 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2746 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2748 * tree-ssa-loop.h: Remove include files.
2749 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2750 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2751 * graphite-clast-to-gimple.c: Likewise.
2752 * graphite-scop-detection.c: Likewise.
2753 * graphite-sese-to-poly.c: Likewise.
2754 * ipa-inline-analysis.c: Likewise.
2755 * ipa-pure-const.c: Likewise.
2756 * loop-init.c: Likewise.
2757 * passes.c: Likewise.
2758 * predict.c: Likewise.
2759 * tree-cfg.c: Likewise.
2760 * tree-cfgcleanup.c: Likewise.
2761 * tree-chrec.c: Likewise.
2762 * tree-data-ref.c: Likewise.
2763 * tree-loop-distribution.c: Likewise.
2764 * tree-parloops.c: Likewise.
2765 * tree-predcom.c: Likewise.
2766 * tree-scalar-evolution.c: Likewise.
2767 * tree-ssa-address.c: Likewise.
2768 * tree-ssa.c: Likewise.
2769 * tree-ssa-dce.c: Likewise.
2770 * tree-ssa-loop.c: Likewise.
2771 * tree-ssa-loop-im.c: Likewise.
2772 * tree-ssa-loop-ivcanon.c: Likewise.
2773 * tree-ssa-loop-ivopts.c: Likewise.
2774 * tree-ssa-loop-manip.c: Likewise.
2775 * tree-ssa-loop-niter.c: Likewise.
2776 * tree-ssa-loop-prefetch.c: Likewise.
2777 * tree-ssa-loop-unswitch.c: Likewise.
2778 * tree-ssa-reassoc.c: Likewise.
2779 * tree-vect-data-refs.c: Likewise.
2780 * tree-vect-loop.c: Likewise.
2781 * tree-vect-loop-manip.c: Likewise.
2782 * tree-vectorizer.c: Likewise.
2783 * tree-vect-stmts.c: Likewise.
2784 * tree-vrp.c: Likewise.
2786 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2788 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2790 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2792 PR tree-optimization/58775
2793 PR tree-optimization/58791
2794 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2795 (insert_stmt_after): Rewritten, don't move the stmt, but really
2797 (get_stmt_uid_with_default): Remove.
2798 (build_and_add_sum): Use insert_stmt_after and
2799 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2800 (update_range_test): Set uid on stmts added by
2801 force_gimple_operand_gsi. Don't immediately modify statements
2802 in inter-bb optimization, just update oe->op values.
2803 (optimize_range_tests): Return bool whether any changed have been made.
2804 (update_ops): New function.
2805 (struct inter_bb_range_test_entry): New type.
2806 (maybe_optimize_range_tests): Perform statement changes here.
2807 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2808 ensure_ops_are_available): Remove.
2809 (find_insert_point): Rewritten.
2810 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2811 return LHS of the (new resp. old) stmt. Don't call
2812 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2813 instead of last, move new stmt at the right place.
2814 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2815 (negate_value): Likewise. Set uids.
2816 (break_up_subtract_bb): Initialize uids.
2817 (reassociate_bb): Adjust rewrite_expr_tree caller.
2818 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2820 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2823 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2824 TARGET_32BIT final condition.
2825 (mulsi3_internal2 and splitter): Same.
2827 2013-10-23 Jeff Law <law@redhat.com>
2829 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2830 through joiner blocks with abnormal outgoing edges.
2832 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2833 Add parameter JOINERS, to allow/disallow threading through joiner
2835 (thread_block): New. Call thread_block_1.
2836 (mark_threaded_blocks): Remove code to filter out certain cases
2837 of threading through joiner blocks.
2838 (thread_through_all_blocks): Document how we can have a dangling
2839 edge AUX field and clear it.
2841 2013-10-23 Ian Lance Taylor <iant@google.com>
2843 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2844 (C++ Dialect Options): Likewise.
2845 (Optimize Options): Likewise.
2847 2013-10-23 Tom de Vries <tom@codesourcery.com>
2849 PR tree-optimization/58805
2850 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2852 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2854 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2855 sequence based on get_range_info returned range.
2857 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2859 * tree-ssa.h: Remove all #include's
2860 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2861 * alias.c: Move required includes from tree-ssa.h.
2863 * builtins.c: Likewise.
2864 * calls.c: Likewise.
2865 * cfgexpand.c: Likewise.
2866 * cfghooks.c: Likewise.
2867 * cfgloop.c: Likewise.
2868 * cfgloopmanip.c: Likewise.
2869 * cgraph.c: Likewise.
2870 * cgraphbuild.c: Likewise.
2871 * cgraphclones.c: Likewise.
2872 * cgraphunit.c: Likewise.
2874 * except.c: Likewise.
2876 * final.c: Likewise.
2877 * fold-const.c: Likewise.
2878 * ggc-page.c: Likewise.
2879 * gimple-builder.c: Likewise.
2880 * gimple-fold.c: Likewise.
2881 * gimple-iterator.c: Likewise.
2882 * gimple-low.c: Likewise.
2883 * gimple-pretty-print.c: Likewise.
2884 * gimple-ssa-strength-reduction.c: Likewise.
2885 * gimple-streamer-in.c: Likewise.
2886 * gimple-streamer-out.c: Likewise.
2887 * gimplify.c: Likewise.
2888 * graphite-blocking.c: Likewise.
2889 * graphite-clast-to-gimple.c: Likewise.
2890 * graphite-dependences.c: Likewise.
2891 * graphite-interchange.c: Likewise.
2892 * graphite-optimize-isl.c: Likewise.
2893 * graphite-poly.c: Likewise.
2894 * graphite-scop-detection.c: Likewise.
2895 * graphite-sese-to-poly.c: Likewise.
2896 * graphite.c: Likewise.
2897 * ipa-cp.c: Likewise.
2898 * ipa-inline-analysis.c: Likewise.
2899 * ipa-inline-transform.c: Likewise.
2900 * ipa-inline.c: Likewise.
2901 * ipa-prop.c: Likewise.
2902 * ipa-pure-const.c: Likewise.
2903 * ipa-reference.c: Likewise.
2904 * ipa-split.c: Likewise.
2905 * ipa-utils.c: Likewise.
2906 * loop-init.c: Likewise.
2907 * lto-cgraph.c: Likewise.
2908 * lto-section-in.c: Likewise.
2909 * lto-section-out.c: Likewise.
2910 * lto-streamer-in.c: Likewise.
2911 * lto-streamer-out.c: Likewise.
2912 * lto-streamer.c: Likewise.
2913 * omp-low.c: Likewise.
2914 * passes.c: Likewise.
2915 * predict.c: Likewise.
2916 * print-tree.c: Likewise.
2917 * profile.c: Likewise.
2919 * targhooks.c: Likewise.
2920 * tracer.c: Likewise.
2921 * trans-mem.c: Likewise.
2922 * tree-call-cdce.c: Likewise.
2923 * tree-cfg.c: Likewise.
2924 * tree-cfgcleanup.c: Likewise.
2925 * tree-chrec.c: Likewise.
2926 * tree-complex.c: Likewise.
2927 * tree-data-ref.c: Likewise.
2928 * tree-dfa.c: Likewise.
2929 * tree-eh.c: Likewise.
2930 * tree-emutls.c: Likewise.
2931 * tree-if-conv.c: Likewise.
2932 * tree-inline.c: Likewise.
2933 * tree-into-ssa.c: Likewise.
2934 * tree-loop-distribution.c: Likewise.
2935 * tree-mudflap.c: Likewise.
2936 * tree-nested.c: Likewise.
2937 * tree-nrv.c: Likewise.
2938 * tree-object-size.c: Likewise.
2939 * tree-outof-ssa.c: Likewise.
2940 * tree-parloops.c: Likewise.
2941 * tree-phinodes.c: Likewise.
2942 * tree-predcom.c: Likewise.
2943 * tree-pretty-print.c: Likewise.
2944 * tree-profile.c: Likewise.
2945 * tree-scalar-evolution.c: Likewise.
2946 * tree-sra.c: Likewise.
2947 * tree-ssa-address.c: Likewise.
2948 * tree-ssa-alias.c: Likewise.
2949 * tree-ssa-ccp.c: Likewise.
2950 * tree-ssa-coalesce.c: Likewise.
2951 * tree-ssa-copy.c: Likewise.
2952 * tree-ssa-copyrename.c: Likewise.
2953 * tree-ssa-dce.c: Likewise.
2954 * tree-ssa-dom.c: Likewise.
2955 * tree-ssa-dse.c: Likewise.
2956 * tree-ssa-forwprop.c: Likewise.
2957 * tree-ssa-ifcombine.c: Likewise.
2958 * tree-ssa-live.c: Likewise.
2959 * tree-ssa-loop-ch.c: Likewise.
2960 * tree-ssa-loop-im.c: Likewise.
2961 * tree-ssa-loop-ivcanon.c: Likewise.
2962 * tree-ssa-loop-ivopts.c: Likewise.
2963 * tree-ssa-loop-manip.c: Likewise.
2964 * tree-ssa-loop-niter.c: Likewise.
2965 * tree-ssa-loop-prefetch.c: Likewise.
2966 * tree-ssa-loop-unswitch.c: Likewise.
2967 * tree-ssa-loop.c: Likewise.
2968 * tree-ssa-math-opts.c: Likewise.
2969 * tree-ssa-operands.c: Likewise.
2970 * tree-ssa-phiopt.c: Likewise.
2971 * tree-ssa-phiprop.c: Likewise.
2972 * tree-ssa-pre.c: Likewise.
2973 * tree-ssa-propagate.c: Likewise.
2974 * tree-ssa-reassoc.c: Likewise.
2975 * tree-ssa-sccvn.c: Likewise.
2976 * tree-ssa-sink.c: Likewise.
2977 * tree-ssa-strlen.c: Likewise.
2978 * tree-ssa-structalias.c: Likewise.
2979 * tree-ssa-tail-merge.c: Likewise.
2980 * tree-ssa-ter.c: Likewise.
2981 * tree-ssa-threadedge.c: Likewise.
2982 * tree-ssa-threadupdate.c: Likewise.
2983 * tree-ssa-uncprop.c: Likewise.
2984 * tree-ssa-uninit.c: Likewise.
2985 * tree-ssa.c: Likewise.
2986 * tree-ssanames.c: Likewise.
2987 * tree-stdarg.c: Likewise.
2988 * tree-streamer-in.c: Likewise.
2989 * tree-switch-conversion.c: Likewise.
2990 * tree-tailcall.c: Likewise.
2991 * tree-vect-data-refs.c: Likewise.
2992 * tree-vect-generic.c: Likewise.
2993 * tree-vect-loop-manip.c: Likewise.
2994 * tree-vect-loop.c: Likewise.
2995 * tree-vect-patterns.c: Likewise.
2996 * tree-vect-slp.c: Likewise.
2997 * tree-vect-stmts.c: Likewise.
2998 * tree-vectorizer.c: Likewise.
2999 * tree-vrp.c: Likewise.
3002 * value-prof.c: Likewise.
3003 * var-tracking.c: Likewise.
3004 * varpool.c: Likewise.
3005 * vtable-verify.c: Likewise.
3007 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3009 * config/tilegx/tilegx.c: Include "tree.h".
3011 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3013 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
3014 the info, not after it.
3015 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
3016 dump_ssaname_info call.
3017 (pp_gimple_stmt_1): Adjust caller.
3018 (dump_phi_nodes): Likewise. Don't print "# " here.
3020 2013-10-22 Jan Hubicka <jh@suse.cz>
3022 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
3024 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
3025 * i386.c (expand_small_movmem_or_setmem): New function.
3026 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
3028 (alg_usable_p): Add support for value ranges; cleanup.
3029 (ix86_expand_set_or_movmem): Add support for misaligned moves.
3031 2013-10-22 Sterling Augustine <saugustine@google.com>
3033 * doc/invoke.texi: Document -ggnu-pubnames.
3034 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
3036 * dwarf2out.c: Include gdb/gdb-index.h.
3037 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
3038 debug_generate_pub_sections.
3039 (is_java, output_pubtables, output_pubname): New functions.
3040 (include_pubname_in_output): Handle debug_generate_pub_sections at
3042 (size_of_pubnames): Use new local space_for_flags based on
3043 debug_generate_pub_sections.
3044 (output_pubnames): Unify pubnames and pubtypes output logic.
3045 Genericize comments. Call output_pubname.
3046 (dwarf2out_finish): Move logic to output_pubnames and call it.
3048 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
3051 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
3052 Remove CCCmode handling.
3053 <case LTU>: Return 'c' suffix for CCCmode.
3054 <case GEU>: Return 'nc' suffix for CCCmode.
3055 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
3056 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
3057 (*sub<mode>3_cc_overflow): Ditto.
3058 (*subsi3_zext_cc_overflow): Ditto.
3060 2013-10-22 Steve Ellcey <sellcey@mips.com>
3062 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
3063 (AND (NOT OP1) (NOT OP2)).
3065 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
3067 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3068 meaning of merge-high and merge-low masks for little endian; avoid
3069 use of vector-pack masks for little endian for mismatched modes.
3071 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3073 * config/tilepro/tilepro.c: Include "tree.h".
3075 2013-10-22 Andreas Schwab <schwab@suse.de>
3077 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
3080 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
3082 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
3083 [Ed Smith-Rowland]): New entries.
3084 ([Stephen M. Webb]): Update.
3086 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
3088 * tree-ssa-ter.h: Remove duplicate copy of file contents.
3090 2013-10-21 Marek Polacek <polacek@redhat.com>
3093 * fold-const.c (fold_range_test): Return 0 if the type is not
3096 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
3098 * system.h: Move hwint.h include further down.
3099 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
3100 gcc_checking_asserts.
3101 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
3103 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3105 Fix volatile issues in optimize_bit_field_compare.
3106 * fold-const.c (optimize_bit_field_compare): Bail out if
3107 lvolatilep or rvolatilep.
3109 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3111 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
3112 and get_inner_reference returning different pmode for non-volatile
3113 bit-field members dependent on flag_strict_volatile_bitfields.
3114 * stor-layout.c (layout_decl): Remove special handling of
3115 flag_strict_volatile_bitfields.
3116 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
3117 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
3119 2013-10-21 Paulo Matos <pmatos@broadcom.com>
3121 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
3124 2013-10-21 Jeff Law <law@redhat.com>
3126 * tree-ssa-threadedge.c (thread_through_normal_block): New
3127 argument VISITED. Remove VISISTED as a local variable. When we
3128 have a threadable jump, verify the destination of the jump has not
3130 (thread_across_edge): Allocate VISITED bitmap once at function
3131 scope and use it throughout. Make sure to set appropriate bits in
3132 VISITED for E (start of jump thread path).
3133 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
3134 through a joiner if any edge on the path has a recorded jump thread.
3136 2013-10-21 Ian Lance Taylor <iant@google.com>
3138 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
3139 don't imply that attributes can solve all problems.
3140 (Directory Options): Fix typo.
3142 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3144 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
3145 extend and extend_add.
3147 2013-10-21 Richard Biener <rguenther@suse.de>
3149 PR tree-optimization/58794
3150 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
3151 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
3153 2013-10-21 Richard Biener <rguenther@suse.de>
3156 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
3157 to (T) X for sign-changing conversions (or no conversion).
3159 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
3161 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
3163 2013-10-20 Jan Hubicka <jh@suse.cz>
3165 * config/i386/i386-tune.def: Add comment; organize into categories
3167 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3169 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
3170 argument. Update function comment.
3171 (expand_set_or_movmem_via_rep): New function combining
3172 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
3173 (expand_movmem_via_rep_mov): Remove.
3174 expand_setmem_via_rep_stos): Remove.
3175 (expand_movmem_epilogue): Update calls correspondingly.
3176 (expand_setmem_epilogue_via_loop): Likewise.
3178 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
3179 (expand_set_or_movmem_prologue): New function combining
3180 expand_movmem_prologue and expand_setmem_prologue.
3181 (expand_movmem_prologue): Remove.
3182 (expand_setmem_prologue): Remove.
3183 (expand_set_or_movmem_constant_prologue): New function combining
3184 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
3185 (expand_constant_movmem_prologue): Remove.
3186 (expand_constant_setmem_prologue): Remove.
3187 (promote_duplicated_reg): Allow vector-const0 value.
3188 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
3189 and ix86_expand_setmem.
3190 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
3191 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
3193 2013-10-21 Diego Novillo <dnovillo@google.com>
3195 * asan.c: Include tree.h
3196 * bb-reorder.c: Likewise.
3197 * cfgcleanup.c: Likewise.
3198 * cfgloopmanip.c: Likewise.
3199 * data-streamer-in.c: Likewise.
3200 * data-streamer-out.c: Likewise.
3201 * data-streamer.c: Likewise.
3202 * dwarf2cfi.c: Likewise.
3203 * graphite-blocking.c: Likewise.
3204 * graphite-clast-to-gimple.c: Likewise.
3205 * graphite-dependences.c: Likewise.
3206 * graphite-interchange.c: Likewise.
3207 * graphite-optimize-isl.c: Likewise.
3208 * graphite-poly.c: Likewise.
3209 * graphite-scop-detection.c: Likewise.
3210 * graphite-sese-to-poly.c: Likewise.
3211 * graphite.c: Likewise.
3212 * ipa-devirt.c: Likewise.
3213 * ipa-profile.c: Likewise.
3216 * loop-init.c: Likewise.
3217 * loop-unroll.c: Likewise.
3218 * lower-subreg.c: Likewise.
3219 * lto/lto-object.c: Likewise.
3220 * recog.c: Likewise.
3221 * reginfo.c: Likewise.
3222 * tree-loop-distribution.c: Likewise.
3223 * tree-parloops.c: Likewise.
3224 * tree-ssa-strlen.c: Likewise.
3225 * tree-streamer.c: Likewise.
3226 * value-prof.c: Likewise.
3227 * target-globals.c: Likewise.
3228 * expr.h: Include tree-core.h instead of tree.h.
3229 * gimple.h: Likewise.
3230 * ipa-prop.h: Likewise.
3231 * ipa-utils.h: Likewise.
3232 * lto-streamer.h: Likewise.
3233 * streamer-hooks.h: Likewise.
3234 * ipa-reference.h: Include cgraph.h instead of tree.h.
3235 * cgraph.h: Include basic-block.h instead of tree.h.
3236 * tree-streamer.h: Do not include tree.h.
3237 * genattrtab.c (write_header): Generate inclusion of
3239 * genautomata.c (main): Likewise.
3240 * genemit.c: Likewise.
3241 * genopinit.c: Likewise.
3242 * genoutput.c (output_prologue): Likewise.
3243 * genpeep.c: Likewise.
3245 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3247 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
3249 (vec_unpacku_hi_v8hi): Likewise.
3250 (vec_unpacku_lo_v16qi): Likewise.
3251 (vec_unpacku_lo_v8hi): Likewise.
3253 2013-10-20 Jan Hubicka <jh@suse.cz>
3255 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
3256 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
3257 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
3259 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
3261 * config/mips/mips.h (ISA_HAS_WSBH): Define.
3262 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
3264 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
3266 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
3269 * system.h: Undef m_slot.
3271 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3273 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
3274 all elements for both endian flavors.
3276 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3279 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
3280 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
3281 and SI_REG for 64bit SYSV ABI targets.
3283 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3285 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
3286 to multi_reg_return. Clarify that we are skipping USEs of multiple
3287 return registers. Use bool type where appropriate.
3289 2013-10-18 Jan Hubicka <jh@suse.cz>
3291 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
3293 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
3294 (X86_TUNE_PUSH_MEMORY): Likewise.
3295 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
3296 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
3297 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
3299 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
3300 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
3302 (ix86_option_override_internal): Update to use tune features instead
3305 2013-10-18 Cong Hou <congh@google.com>
3307 PR tree-optimization/58508
3308 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
3309 statement that contains data refs with zero-step.
3311 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3313 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
3314 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
3315 * gimple-low.c (gimple_check_call_arg,
3316 gimple_check_call_matching_types): Move to cgraph.c.
3317 * gimple-low.h: Remove prototype.
3318 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
3319 Relocate from gimple-low.c.
3320 * cgraph.h: Add prototype. Don't include basic-block.h.
3321 * gimplify.c: Add gimple-low to include list.
3322 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
3323 * tree-eh.c: Add gimple-low to include list.
3324 * tree-nested.c: Likewise.
3325 * cfgexpand.c: Add tree-ssa-address.h to include list.
3327 * gimple-fold.c: Likewise.
3328 * gimple-ssa-strength-reduction.c: Likewise.
3329 * trans-mem.c: Likewise.
3330 * tree-mudflap.c: Likewise.
3331 * tree-ssa-loop-ivopts.c: Likewise.
3332 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
3333 (degenerate_phi_result): Move to tree-phinodes.c.
3334 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
3335 * tree-ssa-threadedge.c: Likewise.
3336 * tree-vrp.c: Likewise.
3337 * tree-phinodes.c (degenerate_phi_result): Relocate here.
3338 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
3339 * tree-phinodes.h (degenerate_phi_result): Add prototype.
3340 * tree-ssa-copy.c: Include tree-ssa-dom.h.
3341 * tree-ssa-forwprop.c: Likewise.
3342 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
3343 pass_data_cleanup_cfg_post_optimizing,
3344 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
3345 * tree-optimize.c: Delete File.
3346 * graphite.c: Include tree-cfgcleanup.h.
3347 * passes.c: Likewise.
3348 * tree-cfg.c: Likewise.
3349 * tree-profile.c: Likewise.
3350 * tree-ssa-dse.c: Likewise.
3351 * tree-ssa-loop-ivcanon.c: Likewise.
3352 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
3353 * tree-outof-ssa.c: Include sbitmap.h.
3354 * tree-ssa-live.c: Likewise.
3355 * tree-ssa-propagate.c: Likewise.
3356 * tree-ssa-structalias.c: Likewise.
3357 * tree-stdarg.c: Likewise.
3358 * Makefile.in (OBJS): Delete tree-optimize.o.
3359 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
3360 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
3361 * varasm.c: Include basic-block.h.
3362 * cfgloop.h: Include function.h instead of basic-block.h
3363 (bb_loop_depth): Move to cfgloop.c.
3364 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
3366 2013-10-18 Teresa Johnson <tejohnson@google.com>
3368 * predict.c (probably_never_executed): Compare frequency-based
3369 count to number of training runs.
3370 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
3372 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3374 * config/arm/arm.c (cortexa9_extra_costs): New table.
3375 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
3377 2013-10-18 Jeff Law <law@redhat.com>
3379 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
3381 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
3382 (register_jump_thread): Add "registered_jump_thread" debug
3384 * dbgcnt.def (registered_jump_thread): New debug counter.
3386 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3388 * config/rs6000/rs6000.c: Include cgraph.h.
3390 2013-10-18 Teresa Johnson <tejohnson@google.com>
3392 * tree-ssa-tail-merge.c (replace_block_by): Update edge
3393 weights during merging.
3395 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3397 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
3398 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
3399 * tree-ssa-operands.h: Remove prototype.
3400 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
3401 * gimple.c (virtual_operand_p): Relocate from gimple.c.
3402 * gimple.h: Add prototype.
3403 * gimple-ssa.h: Include tree-ssa-operands.h.
3404 * tree-dump.c: Add tree-cfg.h to include list.
3405 * tree-ssa-alias.c: Add ipa-reference.h to include list.
3406 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
3407 * config/i386/i386.c: Don't include tree-flow.h.
3408 * config/rs6000/rs6000.c: Likewise.
3410 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3412 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3414 2013-10-18 Richard Biener <rguenther@suse.de>
3416 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3417 or TYPE_UNSIGNED of integral types.
3418 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3419 NULL_TREE for zero-precision integral types.
3421 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3423 * config/aarch64/arm_neon.h
3424 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3426 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3429 * opth-gen.awk: Define target_flags_explicit.
3431 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3433 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3434 fields to the reg_addr array that describes the valid addressing
3435 mode for any register, general purpose registers, floating point
3436 registers, and Altivec registers.
3437 (FIRST_RELOAD_REG_CLASS): Likewise.
3438 (LAST_RELOAD_REG_CLASS): Likewise.
3439 (struct reload_reg_map_type): Likewise.
3440 (reload_reg_map_type): Likewise.
3441 (RELOAD_REG_VALID): Likewise.
3442 (RELOAD_REG_MULTIPLE): Likewise.
3443 (RELOAD_REG_INDEXED): Likewise.
3444 (RELOAD_REG_OFFSET): Likewise.
3445 (RELOAD_REG_PRE_INCDEC): Likewise.
3446 (RELOAD_REG_PRE_MODIFY): Likewise.
3447 (reg_addr): Likewise.
3448 (mode_supports_pre_incdec_p): New helper functions to say whether
3449 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3450 (mode_supports_pre_modify_p): Likewise.
3451 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3452 print the valid address mode bits for each mode.
3453 (rs6000_debug_print_mode): Likewise.
3454 (rs6000_debug_reg_global): Likewise.
3455 (rs6000_setup_reg_addr_masks): New function to set up the address
3456 mask bits for each type.
3457 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3458 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3459 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3460 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3461 PRE_MODIFY are supported.
3462 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3463 registers, instead of {src,dest}_av_p.
3464 (rs6000_print_options_internal): Tweak the debug output slightly.
3466 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3468 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3471 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3473 * tree-flow.h (struct omp_region): Move to omp-low.c.
3474 Remove omp_ prototypes and variables.
3475 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3476 (copy_var_decl): Relocate prototype from tree-flow.h.
3477 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3478 * tree.h: Move prototype to omp-low.h.
3479 * omp-low.h: New File. Relocate prototypes here.
3480 * omp-low.c (struct omp_region): Make local here.
3481 (root_omp_region): Make static.
3482 (copy_var_decl) Move to gimple.c.
3483 (new_omp_region): Make static.
3484 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3485 * tree-cfg.c: Include omp-low.h.
3486 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3487 * gimplify.c: Include omp-low.h.
3488 * tree-parloops.c: Likewise.
3490 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3492 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3493 operands are in memory, prefer to force non-matched operand 1 to
3496 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3499 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3500 restrict TImode addresses to single indirect registers if both
3501 -mquad-memory and -mvsx-timode are used.
3502 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3503 we should emit load/store quad. Remove using %y for quad memory
3506 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3507 constraints to allow load/store quad on machines where TImode is
3508 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3509 for TImode to load integer constants.
3511 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3513 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3515 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3517 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3518 handling of STACK_REG.
3520 2013-10-17 Richard Biener <rguenther@suse.de>
3522 PR tree-optimization/58143
3523 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3525 (rewrite_to_defined_overflow): Likewise.
3526 (move_computations_dom_walker::before_dom): Rewrite stmts
3527 with undefined signed overflow that are not always executed
3528 into unsigned arithmetic.
3530 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3533 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3534 explicit isa flag to be an options variable, instead of using
3535 global_options_set. Remove define from rs6000.h.
3536 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3538 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3539 Initialize rs6000_isa_flags_explicit.
3540 (rs6000_function_specific_save): Add gcc_options* parameter, so
3541 that the powerpc builds after the 2013-10-15 changes.
3542 (rs6000_function_specific_restore): Likewise.
3544 2013-10-16 DJ Delorie <dj@redhat.com>
3546 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3547 op is a REG before checking REGNO.
3548 (rl78_alloc_physical_registers): Verify pattern is a SET before
3551 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3553 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
3555 (vec_unpacks_lo_v4sf): Likewise.
3556 (vec_unpacks_float_hi_v4si): Likewise.
3557 (vec_unpacks_float_lo_v4si): Likewise.
3558 (vec_unpacku_float_hi_v4si): Likewise.
3559 (vec_unpacku_float_lo_v4si): Likewise.
3561 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3563 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
3564 (vsx_concat_v2sf): Likewise.
3566 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
3568 * config/aarch64/aarch64.md
3569 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
3571 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
3573 PR tree-optimization/58697
3574 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
3575 estimated_loop_iterations_int.
3576 (max_stmt_executions_int): Call get_max_loop_iterations_int.
3577 (get_max_loop_iterations_int): New. HWINT version of
3578 get_max_loop_iterations.
3579 * cfgloop.h: Add prototypes.
3580 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
3581 * loop-unroll.c (decide_peel_once_rolling): Call
3582 get_estimated_loop_iterations_int.
3583 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
3584 * tree-ssa-loop-niter.h: Tweak prototypes.
3586 2013-10-16 David Malcolm <dmalcolm@redhat.com>
3588 * gengtype-parse.c (struct_field_seq): Ignore access-control
3589 keywords ("public:" etc).
3591 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3593 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
3594 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
3596 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
3598 * config/arm/arm.opt (mlra): New option.
3599 * config/arm/arm.c (arm_lra_p): New function.
3600 (TARGET_LRA_P): Define.
3602 2013-10-16 Paulo Matos <pmatos@broadcom.com>
3604 * tree-core.h (tree_code_name): Remove.
3605 * tree.h (get_tree_code_name): New prototype.
3606 * tree.c (tree_code_name): Make static.
3607 (get_tree_code_name): New function.
3608 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3609 tree_class_check_failed, tree_range_check_failed,
3610 tree_not_class_check_failed, omp_clause_check_failed,
3611 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3612 wrapper get_tree_code_name instead of calling tree_code_name directly.
3613 * tree-vrp.c (dump_asserts_for): Likewise.
3614 * tree-dump.c (dequeue_and_dump): Likewise.
3615 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3616 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3617 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3618 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3619 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3620 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3621 dump_gimple_omp_for): Likewise.
3622 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3623 * tree-ssa-pre.c (print_pre_expr): Likewise.
3624 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3625 * print-tree.c (print_node_brief, print_node): Likewise.
3626 * gimple.c (gimple_check_failed): Likewise.
3627 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3628 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3629 * config/mep/mep.c (mep_validate_vliw): Likewise.
3630 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3631 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3633 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3635 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3638 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3640 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3641 (MULTILIB_MATCHES): Add multilib for -march=v8.
3643 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3646 * optc-save-gen.awk: Add extra parameter to the save and restore
3648 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3649 * tree.c (build_optimization_node): New parameter. Add extra parameter
3650 to call to cl_optimization_save.
3651 (build_target_option_node): New parameter. Add extra parameter
3652 to call to cl_target_option_save.
3653 * tree.h (build_optimization_node): New parameter.
3654 (build_target_option_node): New parameter.
3655 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3656 build_optimization_node and build_target_option_node.
3657 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3658 (handle_pragma_push_options): Ditto.
3659 * toplev.c (process_options): Ditto.
3660 * opts.c (init_options_struct): Check for opts_set non-null.
3661 * target.def (target_option.save): New parameter.
3662 (target_option.restore): New parameter.
3663 * tm.texi: Generate.
3664 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3665 (ix86_pragma_target_parse): Ditto.
3666 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3668 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3669 to build_optimization_node and build_target_option_node.
3670 (rs6000_valid_attribute_p): Ditto.
3671 (rs6000_pragma_target_parse): Ditto.
3672 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3674 * config/i386/i386.h:
3675 TARGET_64BIT_P: New Macro
3676 TARGET_MMX_P: New Macro.
3677 TARGET_3DNOW_P: New Macro.
3678 TARGET_3DNOW_A_P: New Macro.
3679 TARGET_SSE_P: New Macro.
3680 TARGET_SSE2_P: New Macro.
3681 TARGET_SSE3_P: New Macro.
3682 TARGET_SSSE3_P: New Macro.
3683 TARGET_SSE4_1_P: New Macro.
3684 TARGET_SSE4_2_P: New Macro.
3685 TARGET_AVX_P: New Macro.
3686 TARGET_AVX2_P: New Macro.
3687 TARGET_AVX512F_P: New Macro.
3688 TARGET_AVX512PF_P: New Macro.
3689 TARGET_AVX512ER_P: New Macro.
3690 TARGET_AVX512CD_P: New Macro.
3691 TARGET_FMA_P: New Macro.
3692 TARGET_SSE4A_P: New Macro.
3693 TARGET_FMA4_P: New Macro.
3694 TARGET_XOP_P: New Macro.
3695 TARGET_LWP_P: New Macro.
3696 TARGET_ABM_P: New Macro.
3697 TARGET_BMI_P: New Macro.
3698 TARGET_BMI2_P: New Macro.
3699 TARGET_LZCNT_P: New Macro.
3700 TARGET_TBM_P: New Macro.
3701 TARGET_POPCNT_P: New Macro.
3702 TARGET_SAHF_P: New Macro.
3703 TARGET_MOVBE_P: New Macro.
3704 TARGET_CRC32_P: New Macro.
3705 TARGET_AES_P: New Macro.
3706 TARGET_PCLMUL_P: New Macro.
3707 TARGET_CMPXCHG16B_P: New Macro.
3708 TARGET_FSGSBASE_P: New Macro.
3709 TARGET_RDRND_P: New Macro.
3710 TARGET_F16C_P: New Macro.
3711 TARGET_RTM_P: New Macro.
3712 TARGET_HLE_P: New Macro.
3713 TARGET_RDSEED_P: New Macro.
3714 TARGET_PRFCHW_P: New Macro.
3715 TARGET_ADX_P: New Macro.
3716 TARGET_FXSR_P: New Macro.
3717 TARGET_XSAVE_P: New Macro.
3718 TARGET_XSAVEOPT_P: New Macro.
3719 TARGET_LP64_P: New Macro.
3720 TARGET_X32_P: New Macro.
3721 TARGET_FPMATH_DEFAULT_P: New Macro.
3722 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3723 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3725 Change ix86_tune_string to access opts->x_ix86_tune_string.
3726 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3727 Change ix86_arch_string to access opts->x_ix86_arch_string.
3728 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3729 Change ix86_pmode to access opts->x_ix86_pmode.
3730 Change ix86_abi to access opts->x_ix86_abi.
3731 Change ix86_cmodel to access opts->x_ix86_cmodel.
3732 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3733 Change ix86_isa_flags_explicit to access
3734 opts->x_ix86_isa_flags_explicit.
3735 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3736 Change ix86_regparm to access opts->x_ix86_regparm.
3737 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3738 Change ix86_preferred_stack_boundary_arg to access
3739 opts->x_ix86_preferred_stack_boundary_arg.
3740 Change ix86_force_align_arg_pointer to access
3741 opts->x_ix86_force_align_arg_pointer.
3742 Change ix86_incoming_stack_boundar_arg to access
3743 opts->x_ix86_incoming_stack_boundar_arg.
3744 Change ix86_fpmath to access opts->x_ix86_fpmath.
3745 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3746 Change ix86_recip_name to access opts->x_ix86_recip_name.
3747 Change ix86_stack_protector_guard to access
3748 opts->x_ix86_stack_protector_guard.
3749 Change ix86_tune_memcpy_strategy to access
3750 opts->x_ix86_tune_memcpy_strategy.
3751 Change ix86_tune_memset_strategy to access
3752 opts->x_ix86_tune_memset_strategy.
3753 Change global_options to access opts.
3754 Change global_options_set to access opts_set.
3755 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3756 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3757 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3758 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3759 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3760 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3761 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3762 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3763 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3764 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3765 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3766 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3767 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3768 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3769 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3770 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3771 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3772 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3773 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3774 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3775 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3776 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3777 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3778 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3779 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3780 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3781 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3782 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3783 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3784 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3785 Change TARGET_AES to TARGET_AES_P (opts->...).
3786 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3787 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3788 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3789 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3790 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3791 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3792 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3793 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3794 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3795 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3796 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3797 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3798 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3799 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3800 Change TARGET_X32 to TARGET_X32_P (opts->...).
3801 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3802 Change TARGET_FLOAT_RETURNS_IN_80387 to
3803 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3804 (ix86_function_specific_save): New parameter. Use opts-> fields
3805 to replace global fields.
3806 (ix86_function_specific_restore): Ditto.
3807 (ix86_valid_target_attribute_inner_p): New parameters.
3809 Fix call to ix86_handle_option and set_option.
3810 (ix86_valid_target_attribute_tree): New parameters.
3811 Change global_options to access opts.
3812 Change global_options_set to access opts_set.
3813 Fix call to ix86_valid_target_attribute_inner_p.
3814 Change ix86_tune_string to access opts->x_ix86_tune_string.
3815 Change ix86_arch_string to access opts->x_ix86_arch_string.
3816 Change ix86_fpmath to access opts->x_ix86_fpmath
3817 Fix call to ix86_option_override_internal.
3818 Fix call to ix86_add_new_builtins.
3819 Fix calls to build_optimization_node and build_target_option_node.
3820 (ix86_valid_target_attribute_p): Remove access to global_options.
3821 Use new gcc_options structure func_options.
3822 Fix call to ix86_valid_target_attribute_tree.
3823 Fix call to build_optimization_node.
3824 (get_builtin_code_for_version): Fix call to
3825 ix86_valid_target_attribute_tree.
3827 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3829 * Makefile.in (PICFLAG): New.
3830 (enable_host_shared): New.
3831 (INTERNAL_CFLAGS): Use PICFLAG.
3832 (LIBIBERTY): Use pic build of libiberty.a if configured with
3833 --enable-host-shared.
3834 * configure.ac: Add --enable-host-shared, setting up new
3836 * configure: Regenerate.
3837 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3838 (--enable-host-shared): New option.
3840 2013-10-15 Richard Biener <rguenther@suse.de>
3842 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3843 for built-in functions.
3845 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3847 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3848 (optimize_range_tests_1): New function,
3849 extracted from optimize_range_tests.
3850 (optimize_range_tests_xor): Similarly.
3851 (optimize_range_tests_diff): New function.
3852 (optimize_range_tests): Use optimize_range_tests_1.
3854 2013-10-15 Cong Hou <congh@google.com>
3856 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3857 requirement of the reduction pattern so that one operand of the
3858 reduction operation can come from outside of the loop.
3860 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3862 * config/arm/neon-schedgen.ml: Remove.
3863 * config/arm/cortex-a9-neon.md: Remove comment regarding
3866 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3868 * config/arm/types: Remove old neon types.
3870 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3872 * config/arm/cortex-a7.md
3873 (cortex_a7_neon_type): New.
3874 (cortex_a7_neon_mul): Update for new types.
3875 (cortex_a7_neon_mla): Likewise.
3876 (cortex_a7_neon): Likewise.
3878 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3880 * config/arm/cortex-a15-neon.md
3881 (cortex_a15_neon_type): New,
3883 (cortex_a15_neon_int_1): Remove.
3884 (cortex_a15_neon_int_2): Likewise.
3885 (cortex_a15_neon_int_3): Likewise.
3886 (cortex_a15_neon_int_4): Likewise.
3887 (cortex_a15_neon_int_5): Likewise.
3888 (cortex_a15_neon_vqneg_vqabs): Likewise.
3889 (cortex_a15_neon_vmov): Likewise.
3890 (cortex_a15_neon_vaba): Likewise.
3891 (cortex_a15_neon_vaba_qqq): Likewise.
3892 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3893 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3894 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3896 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3897 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3898 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3899 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3900 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3901 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3902 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3903 (cortex_a15_neon_shift_1): Likewise.
3904 (cortex_a15_neon_shift_2): Likewise.
3905 (cortex_a15_neon_shift_3): Likewise.
3906 (cortex_a15_neon_vshl_ddd): Likewise.
3907 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3908 (cortex_a15_neon_vsra_vrsra): Likewise.
3909 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3910 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3911 (cortex_a15_neon_bp_3cycle): Likewise.
3912 (cortex_a15_neon_ldm_2): Likewise.
3913 (cortex_a15_neon_stm_2): Likewise.
3914 (cortex_a15_neon_mcr): Likewise.
3915 (cortex_a15_neon_mrc): Likewise.
3916 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3917 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3918 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3919 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3920 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3921 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3922 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3923 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3924 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3925 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3926 (cortex_a15_neon_bp_simple): Likewise.
3927 (cortex_a15_neon_bp_2cycle): Likewise.
3928 (cortex_a15_neon_bp_3cycle): Likewise.
3929 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3930 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3931 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3932 (cortex_a15_neon_vld2_4_regs): Likewise.
3933 (cortex_a15_neon_vld3_vld4): Likewise.
3934 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3935 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3936 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3937 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3938 (cortex_a15_neon_vst3_vst4): Rename to...
3939 (cortex_a15_neon_vst4): This, update for new attributes.
3940 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3941 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3942 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3943 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3944 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3945 (cortex_a15_neon_ldm_2): Likewise.
3946 (cortex_a15_neon_stm_2): Likewise.
3947 (cortex_a15_neon_mcr): Likewise.
3948 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3949 (cortex_a15_neon_mrc): Likewise.
3950 (cortex_a15_neon_mrrc): Likewise.
3952 (cortex_a15_neon_abd): New.
3953 (cortex_a15_neon_abd_q): Likewise.
3954 (cortex_a15_neon_aba): Likewise.
3955 (cortex_a15_neon_aba_q): Likewise.
3956 (cortex_a15_neon_acc): Likewise.
3957 (cortex_a15_neon_acc_q): Likewise.
3958 (cortex_a15_neon_arith_basic): Likewise.
3959 (cortex_a15_neon_arith_complex): Likewise.
3960 (cortex_a15_neon_multiply): Likewise.
3961 (cortex_a15_neon_multiply_q): Likewise.
3962 (cortex_a15_neon_mla): Likewise.
3963 (cortex_a15_neon_mla_q): Likewise.
3964 (cortex_a15_neon_sat_mla_long): Likewise.
3965 (cortex_a15_neon_shift_acc): Likewise.
3966 (cortex_a15_neon_shift_imm_basic): Likewise.
3967 (cortex_a15_neon_shift_imm_complex): Likewise.
3968 (cortex_a15_neon_shift_reg_basic): Likewise.
3969 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3970 (cortex_a15_neon_shift_reg_complex): Likewise.
3971 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3972 (cortex_a15_neon_fp_negabs): Likewise
3973 (cortex_a15_neon_fp_arith): Likewise
3974 (cortex_a15_neon_fp_arith_q): Likewise
3975 (cortex_a15_neon_fp_cvt_int): Likewise
3976 (cortex_a15_neon_fp_cvt_int_q): Likewise
3977 (cortex_a15_neon_fp_cvt_16): Likewise
3978 (cortex_a15_neon_fp_mul): Likewise
3979 (cortex_a15_neon_fp_mul_q): Likewise
3980 (cortex_a15_neon_fp_mla): Likewise
3981 (cortex_a15_neon_fp_mla_q): Likewise
3982 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3983 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3984 (cortex_a15_neon_bitops): Likewise.
3985 (cortex_a15_neon_bitops_q): Likewise.
3986 (cortex_a15_neon_from_gp): Likewise.
3987 (cortex_a15_neon_from_gp_q): Likewise.
3988 (cortex_a15_neon_tbl3_tbl4): Likewise.
3989 (cortex_a15_neon_zip_q): Likewise.
3990 (cortex_a15_neon_to_gp): Likewise.
3991 (cortex_a15_neon_load_a): Likewise.
3992 (cortex_a15_neon_load_b): Likewise.
3993 (cortex_a15_neon_load_c): Likewise.
3994 (cortex_a15_neon_load_d): Likewise.
3995 (cortex_a15_neon_load_e): Likewise.
3996 (cortex_a15_neon_load_f): Likewise.
3997 (cortex_a15_neon_store_a): Likewise.
3998 (cortex_a15_neon_store_b): Likewise.
3999 (cortex_a15_neon_store_c): Likewise.
4000 (cortex_a15_neon_store_d): Likewise.
4001 (cortex_a15_neon_store_e): Likewise.
4002 (cortex_a15_neon_store_f): Likewise.
4003 (cortex_a15_neon_store_g): Likewise.
4004 (cortex_a15_neon_store_h): Likewise.
4005 (cortex_a15_vfp_to_from_gp): Likewise.
4007 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4009 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
4011 (cortex_a9_neon_vshl_ddd): Remove.
4012 (cortex_a9_neon_vst3_vst4): Likewise.
4013 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4015 (cortex_a9_neon_bit_ops_q): New.
4017 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
4018 (cortex_a9_neon_int_2): Likewise.
4019 (cortex_a9_neon_int_3): Likewise.
4020 (cortex_a9_neon_int_4): Likewise.
4021 (cortex_a9_neon_int_5): Likewise.
4022 (cortex_a9_neon_vqneg_vqabs): Likewise.
4023 (cortex_a9_neon_vmov): Likewise.
4024 (cortex_a9_neon_vaba): Likewise.
4025 (cortex_a9_neon_vaba_qqq): Likewise.
4026 (cortex_a9_neon_shift_1): Likewise.
4027 (cortex_a9_neon_shift_2): Likewise.
4028 (cortex_a9_neon_shift_3): Likewise.
4029 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4030 (cortex_a9_neon_vsra_vrsra): Likewise.
4031 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4032 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4033 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4035 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4036 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4037 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4039 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4040 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4041 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4042 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4043 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4044 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4045 (cortex_a9_neon_fp_vsum): Likewise.
4046 (cortex_a9_neon_fp_vmul_ddd): Likewise.
4047 (cortex_a9_neon_fp_vmul_qqd): Likewise.
4048 (cortex_a9_neon_fp_vmla_ddd): Likewise.
4049 (cortex_a9_neon_fp_vmla_qqq): Likewise.
4050 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
4051 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
4052 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4053 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4054 (cortex_a9_neon_bp_simple): Likewise.
4055 (cortex_a9_neon_bp_2cycle): Likewise.
4056 (cortex_a9_neon_bp_3cycle): Likewise.
4057 (cortex_a9_neon_ldr): Likewise.
4058 (cortex_a9_neon_str): Likewise.
4059 (cortex_a9_neon_vld1_1_2_regs): Likewise.
4060 (cortex_a9_neon_vld1_3_4_regs): Likewise.
4061 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4062 (cortex_a9_neon_vld2_4_regs): Likewise.
4063 (cortex_a9_neon_vld3_vld4): Likewise.
4064 (cortex_a9_neon_vld1_vld2_lane): Likewise.
4065 (cortex_a9_neon_vld3_vld4_lane): Likewise.
4066 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4067 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4068 (cortex_a9_neon_vst1_3_4_regs): Likewise.
4069 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
4070 (cortex_a9_neon_vst1_vst2_lane): Likewise.
4071 (cortex_a9_neon_vst3_vst4_lane): Likewise.
4072 (cortex_a9_neon_mcr): Likewise.
4073 (cortex_a9_neon_mcr_2_mcrr): Likewise.
4074 (cortex_a9_neon_mrc): Likewise.
4075 (cortex_a9_neon_mrrc): Likewise.
4077 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4079 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
4081 (cortex_a8_neon_vshl_ddd): Remove.
4082 (cortex_a8_neon_vst3_vst4): Likewise.
4083 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4085 (cortex_a8_neon_bit_ops_q): New.
4087 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
4088 (cortex_a8_neon_int_2): Likewise..
4089 (cortex_a8_neon_int_3): Likewise.
4090 (cortex_a8_neon_int_5): Likewise.
4091 (cortex_a8_neon_vqneg_vqabs): Likewise.
4092 (cortex_a8_neon_int_4): Likewise.
4093 (cortex_a8_neon_vaba): Likewise.
4094 (cortex_a8_neon_vaba_qqq): Likewise.
4095 (cortex_a8_neon_shift_1): Likewise.
4096 (cortex_a8_neon_shift_2): Likewise.
4097 (cortex_a8_neon_shift_3): Likewise.
4098 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4099 (cortex_a8_neon_vsra_vrsra): Likewise.
4100 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4101 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4102 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4104 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4105 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4106 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4108 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4109 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4110 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4111 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4112 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4113 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4114 (cortex_a8_neon_fp_vsum): Likewise.
4115 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4116 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4117 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4118 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4119 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4120 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4121 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4122 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4123 (cortex_a8_neon_bp_simple): Likewise.
4124 (cortex_a8_neon_bp_2cycle): Likewise.
4125 (cortex_a8_neon_bp_3cycle): Likewise.
4126 (cortex_a8_neon_ldr): Likewise.
4127 (cortex_a8_neon_str): Likewise.
4128 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4129 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4130 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4131 (cortex_a8_neon_vld2_4_regs): Likewise.
4132 (cortex_a8_neon_vld3_vld4): Likewise.
4133 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4134 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4135 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4136 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4137 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4138 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4139 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4140 (cortex_a8_neon_mcr): Likewise.
4141 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4142 (cortex_a8_neon_mrc): Likewise.
4143 (cortex_a8_neon_mrrc): Likewise.
4145 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4147 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
4149 * config/aarch64/aarch64-simd.md (neon_type): Remove.
4150 (aarch64_simd_dup<mode>): Add "type" attribute.
4151 (aarch64_dup_lane<mode>): Likewise.
4152 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
4153 (*aarch64_simd_mov<mode>): Likewise.
4154 (aarch64_simd_mov_from_<mode>low): Likewise.
4155 (aarch64_simd_mov_from_<mode>high): Likewise.
4156 (orn<mode>3): Likewise.
4157 (bic<mode>3): Likewise.
4158 (add<mode>3): Likewise.
4159 (sub<mode>3): Likewise.
4160 (mul<mode>3): Likewise.
4161 (*aarch64_mul3_elt<mode>): Likewise.
4162 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4163 (*aarch64_mul3_elt_to_128df): Likewise.
4164 (*aarch64_mul3_elt_to_64v2df): Likewise.
4165 (neg<mode>2): Likewise.
4166 (abs<mode>2): Likewise.
4167 (abd<mode>_3): Likewise.
4168 (aba<mode>_3): Likewise.
4169 (fabd<mode>_3): Likewise.
4170 (*fabd_scalar<mode>3): Likewise.
4171 (and<mode>3): Likewise.
4172 (ior<mode>3): Likewise.
4173 (xor<mode>3): Likewise.
4174 (one_cmpl<mode>2): Likewise.
4175 (aarch64_simd_vec_set<mode>): Likewise.
4176 (aarch64_simd_lshr<mode>): Likewise.
4177 (aarch64_simd_ashr<mode>): Likewise.
4178 (aarch64_simd_imm_shl<mode>): Likewise.
4179 (aarch64_simd_reg_sshl<mode): Likewise.
4180 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
4181 (aarch64_simd_reg_shl<mode>_signed): Likewise.
4182 (aarch64_simd_vec_setv2di): Likewise.
4183 (aarch64_simd_vec_set<mode>): Likewise.
4184 (aarch64_mla<mode>): Likewise.
4185 (*aarch64_mla_elt<mode>): Likewise.
4186 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4187 (aarch64_mls<mode>): Likewise.
4188 (*aarch64_mls_elt<mode>): Likewise.
4189 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4190 (<su><maxmin><mode>3): Likewise.
4191 (move_lo_quad_<mode>): Likewise.
4192 (aarch64_simd_move_hi_quad_<mode>): Likewise.
4193 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
4194 (vec_pack_trunc_<mode>): Likewise.
4195 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
4196 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
4197 (*aarch64_<su>mlal_lo<mode>): Likewise.
4198 (*aarch64_<su>mlal_hi<mode>): Likewise.
4199 (*aarch64_<su>mlsl_lo<mode>): Likewise.
4200 (*aarch64_<su>mlsl_hi<mode>): Likewise.
4201 (*aarch64_<su>mlal<mode>): Likewise.
4202 (*aarch64_<su>mlsl<mode>): Likewise.
4203 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
4204 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
4205 (add<mode>3): Likewise.
4206 (sub<mode>3): Likewise.
4207 (mul<mode>3): Likewise.
4208 (div<mode>3): Likewise.
4209 (neg<mode>2): Likewise.
4210 (abs<mode>2): Likewise.
4211 (fma<mode>4): Likewise.
4212 (*aarch64_fma4_elt<mode>): Likewise.
4213 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4214 (*aarch64_fma4_elt_to_128df): Likewise.
4215 (*aarch64_fma4_elt_to_64v2df): Likewise.
4216 (fnma<mode>4): Likewise.
4217 (*aarch64_fnma4_elt<mode>): Likewise.
4218 (*aarch64_fnma4_elt_<vswap_width_name><mode>
4219 (*aarch64_fnma4_elt_to_128df): Likewise.
4220 (*aarch64_fnma4_elt_to_64v2df): Likewise.
4221 (<frint_pattern><mode>2): Likewise.
4222 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
4223 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
4224 (vec_unpacks_lo_v4sf): Likewise.
4225 (aarch64_float_extend_lo_v2df): Likewise.
4226 (vec_unpacks_hi_v4sf): Likewise.
4227 (aarch64_float_truncate_lo_v2sf): Likewise.
4228 (aarch64_float_truncate_hi_v4sf): Likewise.
4229 (aarch64_vmls<mode>): Likewise.
4230 (<su><maxmin><mode>3): Likewise.
4231 (<maxmin_uns><mode>3): Likewise.
4232 (reduc_<sur>plus_<mode>): Likewise.
4233 (reduc_<sur>plus_v2di): Likewise.
4234 (reduc_<sur>plus_v2si): Likewise.
4235 (reduc_<sur>plus_<mode>): Likewise.
4236 (aarch64_addpv4sf): Likewise.
4237 (clz<mode>2): Likewise.
4238 (reduc_<maxmin_uns>_<mode>): Likewise.
4239 (reduc_<maxmin_uns>_v2di): Likewise.
4240 (reduc_<maxmin_uns>_v2si): Likewise.
4241 (reduc_<maxmin_uns>_<mode>): Likewise.
4242 (reduc_<maxmin_uns>_v4sf): Likewise.
4243 (aarch64_simd_bsl<mode>_internal): Likewise.
4244 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
4245 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
4246 (aarch64_get_lane<mode>): Likewise.
4247 (*aarch64_combinez<mode>): Likewise.
4248 (aarch64_combine<mode>): Likewise.
4249 (aarch64_simd_combine<mode>): Likewise.
4250 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
4251 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
4252 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
4253 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
4254 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
4255 (aarch64_<sur>h<addsub><mode>): Likewise.
4256 (aarch64_<sur><addsub>hn<mode>): Likewise.
4257 (aarch64_<sur><addsub>hn2<mode>): Likewise.
4258 (aarch64_pmul<mode>): Likewise.
4259 (aarch64_<su_optab><optab><mode>): Likewise.
4260 (aarch64_<sur>qadd<mode>): Likewise.
4261 (aarch64_sqmovun<mode>): Likewise.
4262 (aarch64_<sur>qmovn<mode>): Likewise.
4263 (aarch64_s<optab><mode>): Likewise.
4264 (aarch64_sq<r>dmulh<mode>): Likewise.
4265 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4266 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
4267 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4268 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
4269 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4270 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4271 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
4272 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
4273 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
4274 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
4275 (aarch64_sqdmull<mode>): Likewise.
4276 (aarch64_sqdmull_lane<mode>_internal): Likewise.
4277 (aarch64_sqdmull_n<mode>): Likewise.
4278 (aarch64_sqdmull2<mode>_internal): Likewise.
4279 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
4280 (aarch64_sqdmull2_n<mode>_internal): Likewise.
4281 (aarch64_<sur>shl<mode>): Likewise.
4282 (aarch64_<sur>q<r>shl<mode>
4283 (aarch64_<sur>shll_n<mode>): Likewise.
4284 (aarch64_<sur>shll2_n<mode>): Likewise.
4285 (aarch64_<sur>shr_n<mode>): Likewise.
4286 (aarch64_<sur>sra_n<mode>): Likewise.
4287 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
4288 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
4289 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
4290 (aarch64_cm<optab><mode>): Likewise.
4291 (aarch64_cm<optab>di): Likewise.
4292 (aarch64_cm<optab><mode>): Likewise.
4293 (aarch64_cm<optab>di): Likewise.
4294 (aarch64_cmtst<mode>): Likewise.
4295 (aarch64_cmtstdi): Likewise.
4296 (aarch64_cm<optab><mode>): Likewise.
4297 (*aarch64_fac<optab><mode>): Likewise.
4298 (aarch64_addp<mode>): Likewise.
4299 (aarch64_addpdi): Likewise.
4300 (sqrt<mode>2): Likewise.
4301 (vec_load_lanesoi<mode>): Likewise.
4302 (vec_store_lanesoi<mode>): Likewise.
4303 (vec_load_lanesci<mode>): Likewise.
4304 (vec_store_lanesci<mode>): Likewise.
4305 (vec_load_lanesxi<mode>): Likewise.
4306 (vec_store_lanesxi<mode>): Likewise.
4307 (*aarch64_mov<mode>): Likewise.
4308 (aarch64_ld2<mode>_dreg): Likewise.
4309 (aarch64_ld2<mode>_dreg): Likewise.
4310 (aarch64_ld3<mode>_dreg): Likewise.
4311 (aarch64_ld3<mode>_dreg): Likewise.
4312 (aarch64_ld4<mode>_dreg): Likewise.
4313 (aarch64_ld4<mode>_dreg): Likewise.
4314 (aarch64_tbl1<mode>): Likewise.
4315 (aarch64_tbl2v16qi): Likewise.
4316 (aarch64_combinev16qi): Likewise.
4317 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
4318 (aarch64_st2<mode>_dreg): Likewise.
4319 (aarch64_st2<mode>_dreg): Likewise.
4320 (aarch64_st3<mode>_dreg): Likewise.
4321 (aarch64_st3<mode>_dreg): Likewise.
4322 (aarch64_st4<mode>_dreg): Likewise.
4323 (aarch64_st4<mode>_dreg): Likewise.
4324 (*aarch64_simd_ld1r<mode>): Likewise.
4325 (aarch64_frecpe<mode>): Likewise.
4326 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
4327 (aarch64_frecps<mode>): Likewise.
4329 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4331 * config/arm/iterators.md (V_elem_ch): New.
4333 (VQH_type): Likewise.
4334 * config/arm/arm.md (is_neon_type): New.
4335 (conds): Use is_neon_type.
4336 (anddi3_insn): Update type attribute.
4337 (xordi3_insn): Likewise.
4338 (one_cmpldi2): Likewise.
4339 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
4340 * config/arm/neon.md (neon_mov): Update type attribute.
4341 (*movmisalign<mode>_neon_store): Likewise.
4342 (*movmisalign<mode>_neon_load): Likewise.
4343 (vec_set<mode>_internal): Likewise.
4344 (vec_set<mode>_internal): Likewise.
4345 (vec_setv2di_internal): Likewise.
4346 (vec_extract<mode>): Likewise.
4347 (vec_extract<mode>): Likewise.
4348 (vec_extractv2di): Likewise.
4349 (*add<mode>3_neon): Likewise.
4350 (adddi3_neon): Likewise.
4351 (*sub<mode>3_neon): Likewise.
4352 (subdi3_neon): Likewise.
4353 (fma<VCVTF:mode>4): Likewise.
4354 (fma<VCVTF:mode>4_intrinsic): Likewise.
4355 (*fmsub<VCVTF:mode>4): Likewise.
4356 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4357 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4358 (ior<mode>3): Likewise.
4359 (and<mode>3): Likewise.
4360 (orn<mode>3_neon): Likewise.
4361 (orndi3_neon): Likewise.
4362 (bic<mode>3_neon): Likewise.
4363 (bicdi3_neon): Likewise.
4364 (xor<mode>3): Likewise.
4365 (one_cmpl<mode>2): Likewise.
4366 (abs<mode>2): Likewise.
4367 (neg<mode>2): Likewise.
4368 (negdi2_neon): Likewise.
4369 (*umin<mode>3_neon): Likewise.
4370 (*umax<mode>3_neon): Likewise.
4371 (*smin<mode>3_neon): Likewise.
4372 (*smax<mode>3_neon): Likewise.
4373 (vashl<mode>3): Likewise.
4374 (vashr<mode>3_imm): Likewise.
4375 (vlshr<mode>3_imm): Likewise.
4376 (ashl<mode>3_signed): Likewise.
4377 (ashl<mode>3_unsigned): Likewise.
4378 (neon_load_count): Likewise.
4379 (ashldi3_neon_noclobber): Likewise.
4380 (ashldi3_neon): Likewise.
4381 (signed_shift_di3_neon): Likewise.
4382 (unsigned_shift_di3_neon): Likewise.
4383 (ashrdi3_neon_imm_noclobber): Likewise.
4384 (lshrdi3_neon_imm_noclobber): Likewise.
4385 (<shift>di3_neon): Likewise.
4386 (widen_ssum<mode>3): Likewise.
4387 (widen_usum<mode>3): Likewise.
4388 (quad_halves_<code>v4si): Likewise.
4389 (quad_halves_<code>v4sf): Likewise.
4390 (quad_halves_<code>v8hi): Likewise.
4391 (quad_halves_<code>v16qi): Likewise.
4392 (reduc_splus_v2di): Likewise.
4393 (neon_vpadd_internal<mode>): Likewise.
4394 (neon_vpsmin<mode>): Likewise.
4395 (neon_vpsmax<mode>): Likewise.
4396 (neon_vpumin<mode>): Likewise.
4397 (neon_vpumax<mode>): Likewise.
4398 (*ss_add<mode>_neon): Likewise.
4399 (*us_add<mode>_neon): Likewise.
4400 (*ss_sub<mode>_neon): Likewise.
4401 (*us_sub<mode>_neon): Likewise.
4402 (neon_vadd<mode>_unspec): Likewise.
4403 (neon_vaddl<mode>): Likewise.
4404 (neon_vaddw<mode>): Likewise.
4405 (neon_vhadd<mode>): Likewise.
4406 (neon_vqadd<mode>): Likewise.
4407 (neon_vaddhn<mode>): Likewise.
4408 (neon_vmul<mode>): Likewise.
4409 (neon_vfms<VCVTF:mode>): Likewise.
4410 (neon_vmlal<mode>): Likewise.
4411 (neon_vmls<mode>): Likewise.
4412 (neon_vmlsl<mode>): Likewise.
4413 (neon_vqdmulh<mode>): Likewise.
4414 (neon_vqdmlal<mode>): Likewise.
4415 (neon_vqdmlsl<mode>): Likewise.
4416 (neon_vmull<mode>): Likewise.
4417 (neon_vqdmull<mode>): Likewise.
4418 (neon_vsub<mode>_unspec): Likewise.
4419 (neon_vsubl<mode>): Likewise.
4420 (neon_vsubw<mode>): Likewise.
4421 (neon_vqsub<mode>): Likewise.
4422 (neon_vhsub<mode>): Likewise.
4423 (neon_vsubhn<mode>): Likewise.
4424 (neon_vceq<mode>): Likewise.
4425 (neon_vcge<mode>): Likewise.
4426 (neon_vcgeu<mode>): Likewise.
4427 (neon_vcgt<mode>): Likewise.
4428 (neon_vcgtu<mode>): Likewise.
4429 (neon_vcle<mode>): Likewise.
4430 (neon_vclt<mode>): Likewise.
4431 (neon_vcage<mode>): Likewise.
4432 (neon_vcagt<mode>): Likewise.
4433 (neon_vtst<mode>): Likewise.
4434 (neon_vabd<mode>): Likewise.
4435 (neon_vabdl<mode>): Likewise.
4436 (neon_vaba<mode>): Likewise.
4437 (neon_vabal<mode>): Likewise.
4438 (neon_vmax<mode>): Likewise.
4439 (neon_vmin<mode>): Likewise.
4440 (neon_vpaddl<mode>): Likewise.
4441 (neon_vpadal<mode>): Likewise.
4442 (neon_vpmax<mode>): Likewise.
4443 (neon_vpmin<mode>): Likewise.
4444 (neon_vrecps<mode>): Likewise.
4445 (neon_vrsqrts<mode>): Likewise.
4446 (neon_vqabs<mode>): Likewise.
4447 (neon_vqneg<mode>): Likewise.
4448 (neon_vcls<mode>): Likewise.
4449 (clz<mode>2): Likewise.
4450 (popcount<mode>2): Likewise.
4451 (neon_vrecpe<mode>): Likewise.
4452 (neon_vrsqrte<mode>): Likewise.
4453 (neon_vget_lane<mode>_sext_internal): Likewise.
4454 (neon_vget_lane<mode>_zext_internal): Likewise.
4455 (neon_vdup_n<mode>): Likewise.
4456 (neon_vdup_n<mode>): Likewise.
4457 (neon_vdup_nv2di): Likewise.
4458 (neon_vdup_lane<mode>_interal): Likewise.
4459 (*neon_vswp<mode>): Likewise.
4460 (neon_vcombine<mode>): Likewise.
4461 (float<mode><V_cvtto>2): Likewise.
4462 (floatuns<mode><V_cvtto>2): Likewise.
4463 (fix_trunc<mode><V_cvtto>2): Likewise.
4464 (fixuns_trunc<mode><V_cvtto>2
4465 (neon_vcvt<mode>): Likewise.
4466 (neon_vcvt<mode>): Likewise.
4467 (neon_vcvtv4sfv4hf): Likewise.
4468 (neon_vcvtv4hfv4sf): Likewise.
4469 (neon_vcvt_n<mode>): Likewise.
4470 (neon_vcvt_n<mode>): Likewise.
4471 (neon_vmovn<mode>): Likewise.
4472 (neon_vqmovn<mode>): Likewise.
4473 (neon_vqmovun<mode>): Likewise.
4474 (neon_vmovl<mode>): Likewise.
4475 (neon_vmul_lane<mode>): Likewise.
4476 (neon_vmul_lane<mode>): Likewise.
4477 (neon_vmull_lane<mode>): Likewise.
4478 (neon_vqdmull_lane<mode>): Likewise.
4479 (neon_vqdmulh_lane<mode>): Likewise.
4480 (neon_vqdmulh_lane<mode>): Likewise.
4481 (neon_vmla_lane<mode>): Likewise.
4482 (neon_vmla_lane<mode>): Likewise.
4483 (neon_vmlal_lane<mode>): Likewise.
4484 (neon_vqdmlal_lane<mode>): Likewise.
4485 (neon_vmls_lane<mode>): Likewise.
4486 (neon_vmls_lane<mode>): Likewise.
4487 (neon_vmlsl_lane<mode>): Likewise.
4488 (neon_vqdmlsl_lane<mode>): Likewise.
4489 (neon_vext<mode>): Likewise.
4490 (neon_vrev64<mode>): Likewise.
4491 (neon_vrev32<mode>): Likewise.
4492 (neon_vrev16<mode>): Likewise.
4493 (neon_vbsl<mode>_internal): Likewise.
4494 (neon_vshl<mode>): Likewise.
4495 (neon_vqshl<mode>): Likewise.
4496 (neon_vshr_n<mode>): Likewise.
4497 (neon_vshrn_n<mode>): Likewise.
4498 (neon_vqshrn_n<mode>): Likewise.
4499 (neon_vqshrun_n<mode>): Likewise.
4500 (neon_vshl_n<mode>): Likewise.
4501 (neon_vqshl_n<mode>): Likewise.
4502 (neon_vqshlu_n<mode>): Likewise.
4503 (neon_vshll_n<mode>): Likewise.
4504 (neon_vsra_n<mode>): Likewise.
4505 (neon_vsri_n<mode>): Likewise.
4506 (neon_vsli_n<mode>): Likewise.
4507 (neon_vtbl1v8qi): Likewise.
4508 (neon_vtbl2v8qi): Likewise.
4509 (neon_vtbl3v8qi): Likewise.
4510 (neon_vtbl4v8qi): Likewise.
4511 (neon_vtbl1v16qi): Likewise.
4512 (neon_vtbl2v16qi): Likewise.
4513 (neon_vcombinev16qi): Likewise.
4514 (neon_vtbx1v8qi): Likewise.
4515 (neon_vtbx2v8qi): Likewise.
4516 (neon_vtbx3v8qi): Likewise.
4517 (neon_vtbx4v8qi): Likewise.
4518 (*neon_vtrn<mode>_insn): Likewise.
4519 (*neon_vzip<mode>_insn): Likewise.
4520 (*neon_vuzp<mode>_insn): Likewise.
4521 (neon_vld1<mode>): Likewise.
4522 (neon_vld1_lane<mode>): Likewise.
4523 (neon_vld1_lane<mode>): Likewise.
4524 (neon_vld1_dup<mode>): Likewise.
4525 (neon_vld1_dup<mode>): Likewise.
4526 (neon_vld1_dupv2di): Likewise.
4527 (neon_vst1<mode>): Likewise.
4528 (neon_vst1_lane<mode>): Likewise.
4529 (neon_vst1_lane<mode>): Likewise.
4530 (neon_vld2<mode>): Likewise.
4531 (neon_vld2<mode>): Likewise.
4532 (neon_vld2_lane<mode>): Likewise.
4533 (neon_vld2_lane<mode>): Likewise.
4534 (neon_vld2_dup<mode>): Likewise.
4535 (neon_vst2<mode>): Likewise.
4536 (neon_vst2<mode>): Likewise.
4537 (neon_vst2_lane<mode>): Likewise.
4538 (neon_vst2_lane<mode>): Likewise.
4539 (neon_vld3<mode>): Likewise.
4540 (neon_vld3qa<mode>): Likewise.
4541 (neon_vld3qb<mode>): Likewise.
4542 (neon_vld3_lane<mode>): Likewise.
4543 (neon_vld3_lane<mode>): Likewise.
4544 (neon_vld3_dup<mode>): Likewise.
4545 (neon_vst3<mode>): Likewise.
4546 (neon_vst3qa<mode>): Likewise.
4547 (neon_vst3qb<mode>): Likewise.
4548 (neon_vst3_lane<mode>): Likewise.
4549 (neon_vst3_lane<mode>): Likewise.
4550 (neon_vld4<mode>): Likewise.
4551 (neon_vld4qa<mode>): Likewise.
4552 (neon_vld4qb<mode>): Likewise.
4553 (neon_vld4_lane<mode>): Likewise.
4554 (neon_vld4_lane<mode>): Likewise.
4555 (neon_vld4_dup<mode>): Likewise.
4556 (neon_vst4<mode>): Likewise.
4557 (neon_vst4qa<mode>): Likewise.
4558 (neon_vst4qb<mode>): Likewise.
4559 (neon_vst4_lane<mode>): Likewise.
4560 (neon_vst4_lane<mode>): Likewise.
4561 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4562 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4563 (neon_vec_<US>mult_lo_<mode>): Likewise.
4564 (neon_vec_<US>mult_hi_<mode>): Likewise.
4565 (neon_vec_<US>shiftl_<mode>): Likewise.
4566 (neon_unpack<US>_<mode>): Likewise.
4567 (neon_vec_<US>mult_<mode>): Likewise.
4568 (vec_pack_trunc_<mode>): Likewise.
4569 (neon_vec_pack_trunc_<mode>): Likewise.
4570 (neon_vabd<mode>_2): Likewise.
4571 (neon_vabd<mode>_3): Likewise.
4573 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4575 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
4576 (load_pair): Update type attribute.
4577 (store_pair): Update type attribute.
4578 * config/aarch64/iterators.md (q): New.
4580 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4582 * config/arm/types.md: Add new types for Neon insns.
4584 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4585 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4586 Sergey Lega <sergey.s.lega@intel.com>
4587 Anna Tikhonova <anna.tikhonova@intel.com>
4588 Ilya Tocar <ilya.tocar@intel.com>
4589 Andrey Turetskiy <andrey.turetskiy@intel.com>
4590 Ilya Verbin <ilya.verbin@intel.com>
4591 Kirill Yukhin <kirill.yukhin@intel.com>
4592 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4594 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
4595 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
4596 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
4598 (srcp14<mode>): Ditto.
4599 (rsqrt14<mode>): Ditto.
4600 (rsqrt14<mode>): Ditto.
4601 (avx512f_vmscalef<mode>): Ditto.
4602 (avx512f_scalef<mode>): Ditto.
4603 (avx512f_getexp<mode>): Ditto.
4604 (avx512f_sgetexp<mode>): Ditto.
4605 (avx512f_fixupimm<mode>): Ditto.
4606 (avx512f_sfixupimm<mode>): Ditto.
4607 (avx512f_rndscale<mode>): Ditto.
4608 (*avx512er_exp2<mode>): Ditto.
4609 (*avx512er_rcp28<mode>): Ditto.
4610 (avx512er_rsqrt28<mode>): Ditto.
4611 (avx512f_getmant<mode>): Ditto.
4612 (avx512f_getmant<mode>): Ditto.
4613 (avx512f_rndscale<mode>): Fix formatting.
4615 2013-10-15 Martin Jambor <mjambor@suse.cz>
4617 * ipa-utils.h (ipa_edge_within_scc): Declare.
4618 * ipa-cp.c (edge_within_scc): Moved...
4619 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4621 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4622 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4623 Sergey Lega <sergey.s.lega@intel.com>
4624 Anna Tikhonova <anna.tikhonova@intel.com>
4625 Ilya Tocar <ilya.tocar@intel.com>
4626 Andrey Turetskiy <andrey.turetskiy@intel.com>
4627 Ilya Verbin <ilya.verbin@intel.com>
4628 Kirill Yukhin <kirill.yukhin@intel.com>
4629 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4631 * config/i386/predicates.md (const_8_to_15_operand): New.
4632 (const_16_to_31_operand): Ditto.
4633 * config/i386/sse.md (V8FI): New.
4635 (reduc_splus_v8df): Ditto.
4636 (reduc_splus_v16sf): Ditto.
4637 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4638 (vec_extract_hi_<mode>): Ditto.
4639 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4640 (vec_set_lo_<mode>): Ditto.
4641 (vec_set_hi_<mode>): Ditto.
4642 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4643 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4644 (avx512f_pshufd_1): Ditto.
4645 (avx512f_broadcast<mode>): Ditto.
4646 (avx512f_broadcast<mode>): Ditto.
4647 (define_split): Split vec_extract_lo into move.
4648 (ssequartermode): Ditto.
4649 (ssedoublemode): Extened with wider modes.
4650 (vec_extract_lo_<mode>): Ditto.
4652 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4653 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4654 Sergey Lega <sergey.s.lega@intel.com>
4655 Anna Tikhonova <anna.tikhonova@intel.com>
4656 Ilya Tocar <ilya.tocar@intel.com>
4657 Andrey Turetskiy <andrey.turetskiy@intel.com>
4658 Ilya Verbin <ilya.verbin@intel.com>
4659 Kirill Yukhin <kirill.yukhin@intel.com>
4660 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4662 * config/i386/predicates.md (register_or_constm1_operand): New.
4663 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4664 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4665 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4666 UNSPEC_SCATTER_PREFETCH
4668 (avx512f_ucmp<mode>3): Ditto.
4669 (avx512f_vternlog<mode>): Ditto.
4670 (avx512f_align<mode>): Ditto.
4671 (<shift_insn><mode>3): Ditto.
4672 (avx512f_<rotate>v<mode>): Ditto.
4673 (avx512f_<rotate><mode>): Ditto.
4674 (avx512f_eq<mode>3): Ditto.
4675 (avx512f_eq<mode>3_1): Ditto.
4676 (avx512f_gt<mode>3): Ditto.
4677 (avx512f_testm<mode>3): Ditto.
4678 (avx512f_testnm<mode>3): Ditto.
4679 (avx512pf_gatherpf<mode>): Ditto.
4680 (*avx512pf_gatherpf<mode>_mask): Ditto.
4681 (*avx512pf_gatherpf<mode>): Ditto.
4682 (avx512pf_scatterpf<mode>): Ditto.
4683 (*avx512pf_scatterpf<mode>_mask): Ditto.
4684 (*avx512pf_scatterpf<mode>): Ditto.
4685 (avx512f_vec_dup_gpr<mode>): Ditto.
4686 (clz<mode>2): Ditto.
4687 (conflict<mode>): Ditto.
4688 (REDUC_SMINMAX_MODE): Extened with wider modes.
4689 (reduc_<code>_<mode>): Ditto.
4690 (vlshr<mode>3): Ditto.
4691 (vashl<mode>3): Ditto.
4693 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4694 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4695 Sergey Lega <sergey.s.lega@intel.com>
4696 Anna Tikhonova <anna.tikhonova@intel.com>
4697 Ilya Tocar <ilya.tocar@intel.com>
4698 Andrey Turetskiy <andrey.turetskiy@intel.com>
4699 Ilya Verbin <ilya.verbin@intel.com>
4700 Kirill Yukhin <kirill.yukhin@intel.com>
4701 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4703 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4706 (avx512fmaskmode): Ditto.
4707 (bcstscalarsuff): Ditto.
4708 (avx512f_blendm<mode>): Ditto.
4709 (cmp_imm_predicate): Ditto.
4710 (avx512f_cmp<mode>3): Ditto.
4711 (avx512f_vec_dup<mode>): Ditto.
4712 (avx512f_vec_dup_mem<mode>): Ditto.
4713 (avx512f_vpermi2var<mode>3): Ditto.
4714 (avx512f_vpermt2var<mode>3): Ditto.
4715 (vec_init<mode>): Ditto.
4716 (avx512f_gathersi<mode>): Ditto.
4717 (*avx512f_gathersi<mode>): Ditto.
4718 (*avx512f_gathersi<mode>_2): Ditto.
4719 (avx512f_gatherdi<mode>): Ditto.
4720 (*avx512f_gatherdi<mode>): Ditto.
4721 (*avx512f_gatherdi<mode>_2): Ditto.
4722 (avx512f_scattersi<mode>): Ditto.
4723 (*avx512f_scattersi<mode>): Ditto.
4724 (avx512f_scatterdi<mode>): Ditto.
4725 (*avx512f_scatterdi<mode>): Ditto.
4726 (sseintprefix): Extened with wider modes.
4727 (VEC_GATHER_IDXSI): Ditto.
4728 (VEC_GATHER_IDXDI): Ditto.
4729 (VEC_GATHER_SRCDI): Ditto.
4731 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4732 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4734 * config/arm/t-aprofile: New file.
4735 * config.gcc: Handle --with-multilib-list option.
4737 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4739 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4742 2013-10-15 Richard Biener <rguenther@suse.de>
4744 * tree-loop-distribution.c (build_empty_rdg): Inline into
4746 (rdg_flag_vertex): Inline into single user.
4747 (rdg_flag_vertex_and_dependent): Likewise.
4748 (build_rdg_partition_for_vertex): Remove processed bitmap.
4749 (rdg_build_partitions): Simplify.
4751 2013-10-15 Richard Biener <rguenther@suse.de>
4753 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4754 Restructure forwarding through conversions and copies to
4755 avoid performing copy-propagation the wrong way. Adjust
4756 recursion invocations.
4757 (forward_propagate_addr_expr): Add argument stating if we
4758 are recursing from a single-use.
4759 (ssa_forward_propagate_and_combine): Adjust.
4761 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4763 * dumpfile.h (gcc::dump_manager): New class, to hold state
4764 relating to dumpfile management.
4765 (get_dump_file_name): Remove in favor of method of dump_manager.
4766 (dump_initialized_p): Likewise.
4767 (dump_start): Likewise.
4768 (dump_finish): Likewise.
4769 (dump_switch_p): Likewise.
4770 (dump_register): Likewise.
4771 (get_dump_file_info): Likewise.
4772 * context.c (gcc::context::context): Construct the dump_manager
4774 * context.h (gcc::context::get_dumps): New.
4775 (gcc::context::m_dumps): New.
4776 * coverage.c (coverage_init): Port to dump_manager API.
4777 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4778 (extra_dump_files_in_use): Likewise.
4779 (extra_dump_files_alloced): Likewise.
4780 (gcc::dump_manager::dump_manager): New.
4781 (dump_register): Convert to...
4782 (gcc::dump_manager::dump_register): ...method, replacing
4783 function-static next_dump with m_next_dump field.
4784 (get_dump_file_info): Convert to...
4785 (gcc::dump_manager::get_dump_file_info): ...method.
4786 (get_dump_file_name): Convert to...
4787 (gcc::dump_manager::get_dump_file_name): ...method.
4788 (dump_start): Convert to...
4789 (gcc::dump_manager::dump_start): ...method.
4790 (dump_finish): Convert to...
4791 (gcc::dump_manager::dump_finish): ...method.
4792 (dump_begin): Replace body with...
4793 (gcc::dump_manager::dump_begin): ...new method.
4794 (dump_phase_enabled_p): Convert to...
4795 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4796 (dump_phase_enabled_p): Convert to...
4797 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4798 (dump_initialized_p): Convert to...
4799 (gcc::dump_manager::dump_initialized_p): ...method.
4800 (dump_flag_name): Replace body with...
4801 (gcc::dump_manager::dump_flag_name): ...new method.
4802 (dump_enable_all): Convert to...
4803 (gcc::dump_manager::dump_enable_all): ...new method.
4804 (opt_info_enable_passes): Convert to...
4805 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4806 (dump_switch_p_1): Convert to...
4807 (gcc::dump_manager::dump_switch_p_1): ...new method.
4808 (dump_switch_p): Convert to...
4809 (gcc::dump_manager::dump_switch_p): ...new method.
4810 (opt_info_switch_p): Port to dump_manager API.
4811 (enable_rtl_dump_file): Likewise.
4812 * opts-global.c (handle_common_deferred_options): Port to new
4814 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4815 (pass_manager::register_one_dump_file): Likewise.
4816 (pass_manager::register_pass): Likewise.
4817 (pass_init_dump_file): Likewise.
4818 (pass_fini_dump_file): Likewise.
4819 * statistics.c (statistics_early_init): Likewise.
4821 2013-10-14 Richard Biener <rguenther@suse.de>
4823 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4824 iterative_hash_canonical_type, gimple_canonical_type_hash,
4825 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4826 gimple_register_canonical_type, print_gimple_types_stats,
4827 free_gimple_type_tables): Move to lto/lto.c
4828 (gt-gimple.h): Do not include.
4829 * gimple.h (gimple_register_canonical_type,
4830 print_gimple_types_stats, free_gimple_type_tables): Remove.
4831 * Makefile.in (GTFILES): Remove gimple.c.
4833 2013-10-14 Travis Snoozy <quandary@remstate.com>
4836 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4837 scanning for msp430x targets.
4839 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4842 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4843 (registers_ok_for_ldd_peep): Move around.
4844 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4845 * config/sparc/sparc.md (widening peepholes): Use it.
4847 2013-10-14 Richard Biener <rguenther@suse.de>
4851 * gimple.c (iterative_hash_canonical_type): Make sure to
4852 record the hash into the correct hashtable slot.
4854 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4856 PR rtl-optimization/58662
4857 * combine.c (try_combine): Take into account death nodes on I2 when
4858 splitting a PARALLEL of two independent SETs. Fix dump message.
4860 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4863 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4864 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4865 * config/sh/t-sh (sh_treg_combine.o): New entry.
4866 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4867 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4868 (register_sh_passes): New function. Register sh_treg_combine pass.
4869 (sh_option_override): Invoke it.
4870 (sh_canonicalize_comparison): Handle op0_preserve_value.
4871 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4872 opportunities. Canonicalize branch condition.
4873 (nott): Allow only if pseudos can be created for non-SH2A.
4875 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4878 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4879 (ix86_expand_movmem): Replace copy_addr_to_reg with
4880 ix86_copy_addr_to_reg.
4881 (ix86_expand_setmem): Likewise.
4883 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4885 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4886 provided by ix86_fp_compare_mode instead of CCFPUmode.
4888 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4890 * config/aarch64/arm_neon.h
4891 (vtbx<1,3>_<psu>8): Fix register constriants.
4893 2013-10-11 Jeff Law <law@redhat.com>
4895 PR tree-optimization/58640
4896 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4897 threading paths that cross over two loop entry points.
4899 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4901 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4902 handle vector float as well.
4903 (*vsx_le_perm_load_v4si): Likewise.
4904 (*vsx_le_perm_store_v2di): Likewise.
4905 (*vsx_le_perm_store_v4si): Likewise.
4907 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4909 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4910 directly to circumvent subtract from splat{31} workaround.
4911 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4913 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4914 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4915 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4916 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4917 separate big and little endian logic.
4918 (*altivec_vperm_<mode>_internal): New define_insn.
4919 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4920 separate big and little endian logic.
4921 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4922 (vec_permv16qi): Add little endian logic.
4924 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4926 * doc/extend.texi (returns_nonnull): Remove arguments.
4928 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4929 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4930 Sergey Lega <sergey.s.lega@intel.com>
4931 Anna Tikhonova <anna.tikhonova@intel.com>
4932 Ilya Tocar <ilya.tocar@intel.com>
4933 Andrey Turetskiy <andrey.turetskiy@intel.com>
4934 Ilya Verbin <ilya.verbin@intel.com>
4935 Kirill Yukhin <kirill.yukhin@intel.com>
4936 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4938 * config/i386/sse.md (VI48F_256_512): New.
4939 (avx2_permvar<mode>): Change to ...
4940 (<avx2_avx512f>_permvar<mode>): This.
4942 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4943 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4944 Sergey Lega <sergey.s.lega@intel.com>
4945 Anna Tikhonova <anna.tikhonova@intel.com>
4946 Ilya Tocar <ilya.tocar@intel.com>
4947 Andrey Turetskiy <andrey.turetskiy@intel.com>
4948 Ilya Verbin <ilya.verbin@intel.com>
4949 Kirill Yukhin <kirill.yukhin@intel.com>
4950 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4952 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4953 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4954 * config/i386/sse.md (VI4_AVX): New.
4955 (sf2simodelower): Ditto.
4956 (sse2_cvtps2dq): Change to ...
4957 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4959 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4960 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4961 Sergey Lega <sergey.s.lega@intel.com>
4962 Anna Tikhonova <anna.tikhonova@intel.com>
4963 Ilya Tocar <ilya.tocar@intel.com>
4964 Andrey Turetskiy <andrey.turetskiy@intel.com>
4965 Ilya Verbin <ilya.verbin@intel.com>
4966 Kirill Yukhin <kirill.yukhin@intel.com>
4967 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4969 * config/i386/sse.md (V_512): New.
4971 (vcond<V_512:mode><VF_512:mode>): Ditto.
4972 (vcond<V_512:mode><VI_512:mode>): Ditto.
4973 (vcondu<V_512:mode><VI_512:mode>): Ditto.
4975 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4976 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4977 Sergey Lega <sergey.s.lega@intel.com>
4978 Anna Tikhonova <anna.tikhonova@intel.com>
4979 Ilya Tocar <ilya.tocar@intel.com>
4980 Andrey Turetskiy <andrey.turetskiy@intel.com>
4981 Ilya Verbin <ilya.verbin@intel.com>
4982 Kirill Yukhin <kirill.yukhin@intel.com>
4983 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4985 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4986 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4988 (fma<mode>4): Removed condition.
4989 (fms<mode>4): Ditto.
4990 (fnma<mode>4): Ditto.
4991 (fnms<mode>4): Ditto.
4992 (fma4i_fmadd_<mode>): Ditto.
4993 (*fma_fmadd_<mode>): Ditto.
4994 (*fma_fmsub_<mode>): Ditto.
4995 (*fma_fnmadd_<mode>): Ditto.
4996 (*fma_fnmsub_<mode>): Ditto.
4997 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4998 (*fma_fmaddsub_<mode>): Ditto.
4999 (*fma_fmsubadd_<mode>): Ditto.
5000 (*fmai_fmadd_<mode>): Ditto.
5001 (*fmai_fmsub_<mode>): Ditto.
5002 (*fmai_fnmadd_<mode>): Ditto.
5003 (*fmai_fnmsub_<mode>): Ditto.
5005 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5006 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5007 Sergey Lega <sergey.s.lega@intel.com>
5008 Anna Tikhonova <anna.tikhonova@intel.com>
5009 Ilya Tocar <ilya.tocar@intel.com>
5010 Andrey Turetskiy <andrey.turetskiy@intel.com>
5011 Ilya Verbin <ilya.verbin@intel.com>
5012 Kirill Yukhin <kirill.yukhin@intel.com>
5013 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5015 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
5016 (VI124_256): Changed to ...
5017 (VI124_256_48_512): This.
5018 (ssepackmode): Extended with wider modes.
5019 (<code><mode>3): Changed iterator.
5020 (*avx2_<code><mode>3): Ditto.
5021 (vec_pack_trunc_<mode>): Ditto.
5023 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5024 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5025 Sergey Lega <sergey.s.lega@intel.com>
5026 Anna Tikhonova <anna.tikhonova@intel.com>
5027 Ilya Tocar <ilya.tocar@intel.com>
5028 Andrey Turetskiy <andrey.turetskiy@intel.com>
5029 Ilya Verbin <ilya.verbin@intel.com>
5030 Kirill Yukhin <kirill.yukhin@intel.com>
5031 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5033 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
5034 (VI8F_256_512): Ditto.
5035 (abs<mode>2): Changed iterator.
5036 (avx2_perm<mode>): Changed to ...
5037 (<avx2_avx512f>_perm<mode>): This.
5038 (avx2_perm<mode>_1): Changed to ...
5039 (<avx2_avx512f>_perm<mode>_1): This.
5041 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5042 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5043 Sergey Lega <sergey.s.lega@intel.com>
5044 Anna Tikhonova <anna.tikhonova@intel.com>
5045 Ilya Tocar <ilya.tocar@intel.com>
5046 Andrey Turetskiy <andrey.turetskiy@intel.com>
5047 Ilya Verbin <ilya.verbin@intel.com>
5048 Kirill Yukhin <kirill.yukhin@intel.com>
5049 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5051 * config/i386/sse.md (VI48_AVX512F): New.
5052 (VI48_AVX2): Changed to ...
5053 (VI48_AVX2_48_AVX512F): This.
5054 (avx2_ashrv<mode>): Changed to ...
5055 (<avx2_avx512f>_ashrv<mode>): This.
5056 (avx2_<shift_insn>v<mode>): Changed to ...
5057 (<avx2_avx512f>_<shift_insn>v<mode>): This.
5059 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5060 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5061 Sergey Lega <sergey.s.lega@intel.com>
5062 Anna Tikhonova <anna.tikhonova@intel.com>
5063 Ilya Tocar <ilya.tocar@intel.com>
5064 Andrey Turetskiy <andrey.turetskiy@intel.com>
5065 Ilya Verbin <ilya.verbin@intel.com>
5066 Kirill Yukhin <kirill.yukhin@intel.com>
5067 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5069 * config/i386/sse.md (VI4_AVX512F): New.
5070 (VI8_AVX2_AVX512F): Ditto.
5071 (mul<mode>3): Extended with wider modes.
5072 (*<sse4_1_avx2>_mul<mode>3): Ditto.
5073 (mul<mode>3): Ditto.
5074 (vec_widen_<s>mult_odd_<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/sse.md (VI2_AVX512F): New.
5087 (VI124_AVX512F): Ditto.
5088 (sseunpackmode): Extended with wider modes.
5089 (sseunpackfltmode): Ditto.
5090 (vec_unpacks_float_hi_<mode>): Ditto.
5091 (vec_unpacks_float_lo_<mode>): Ditto.
5092 (vec_unpacku_float_hi_<mode>): Ditto.
5093 (vec_unpacku_float_lo_<mode>): Ditto.
5094 (vec_unpacks_lo_<mode>): Ditto.
5095 (vec_unpacks_hi_<mode>): Ditto.
5096 (vec_unpacku_lo_<mode>): Ditto.
5097 (vec_unpacku_hi_<mode>): Ditto.
5099 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5100 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5101 Sergey Lega <sergey.s.lega@intel.com>
5102 Anna Tikhonova <anna.tikhonova@intel.com>
5103 Ilya Tocar <ilya.tocar@intel.com>
5104 Andrey Turetskiy <andrey.turetskiy@intel.com>
5105 Ilya Verbin <ilya.verbin@intel.com>
5106 Kirill Yukhin <kirill.yukhin@intel.com>
5107 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5109 * config/i386/i386.md (multdiv): New.
5110 (multdiv_mnemonic): Ditto.
5111 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
5112 (<sse>_vm<multdiv_mnemonic><mode>3): This.
5113 (<sse>_vmdiv<mode>3): Removed.
5115 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5116 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5117 Sergey Lega <sergey.s.lega@intel.com>
5118 Anna Tikhonova <anna.tikhonova@intel.com>
5119 Ilya Tocar <ilya.tocar@intel.com>
5120 Andrey Turetskiy <andrey.turetskiy@intel.com>
5121 Ilya Verbin <ilya.verbin@intel.com>
5122 Kirill Yukhin <kirill.yukhin@intel.com>
5123 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5125 * config/i386/sse.md (V): Extended with wider modes.
5127 (ssehalfvecmode): Ditto.
5129 (ssepackfltmode): Ditto.
5130 (avx_vec_concat<mode>): Ditto.
5131 (V_256_512): New iterator.
5132 (VF2_512_256): Ditto.
5133 (si2dfmode): New attribute.
5134 (si2dfmodelower): Ditto.
5136 (concat_tg_mode): Ditto.
5137 (floatv4siv4df2): Changed to ...
5138 (float<si2dfmodelower><mode>2): This.
5139 (avx_cvtps2pd256): Changed to ...
5140 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
5141 (vec_pack_trunc_v4df): Changed to ...
5142 (vec_pack_trunc_<mode>): This.
5143 (avx_vpermil<mode>): Changed to ...
5144 (<sse2_avx_avx512f>_vpermil<mode>): This.
5145 (<fixsuffix>fix_truncv8dfv8si2): New.
5146 (vec_pack_sfix_trunc_v8df): Ditto.
5147 (avx512f_rndscale<mode>): Ditto.
5148 (avx512f_roundpd512): Ditto.
5149 (vec_pack_ufix_trunc_<mode>): Updated iterator.
5151 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5152 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5153 Sergey Lega <sergey.s.lega@intel.com>
5154 Anna Tikhonova <anna.tikhonova@intel.com>
5155 Ilya Tocar <ilya.tocar@intel.com>
5156 Andrey Turetskiy <andrey.turetskiy@intel.com>
5157 Ilya Verbin <ilya.verbin@intel.com>
5158 Kirill Yukhin <kirill.yukhin@intel.com>
5159 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5161 * config/i386/i386.md (any_fix): New iterator.
5162 (fixsuffix): New attribute.
5163 * config/i386/sse.md (VF1): Extened with wider modes.
5167 (sseintvecmodelower): Ditto.
5168 (ssescalarmode): Ditto.
5169 (ssescalarnum): Ditto.
5171 (ssexmmmode): Ditto.
5172 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
5173 (<sse>_rcp<mode>2): Change iterator.
5174 (rsqrt<mode>2): Ditto.
5175 (<sse>_rsqrt<mode>2): Ditto.
5176 (avx2_vec_dup<mode>): Ditto.
5177 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
5178 (round<mode>2_sfix): Ditto.
5179 (avx2_pbroadcast<mode>): Ditto.
5180 (*andnot<mode>3): Handle XI mode.
5181 (*<code><mode>3): Ditto.
5182 (AVXTOSSEMODE): Removed.
5183 (avx_vpermil<mode>): Changed to ...
5184 (<sse2_avx_avx512f>_vpermil<mode>): This.
5186 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5187 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5188 Sergey Lega <sergey.s.lega@intel.com>
5189 Anna Tikhonova <anna.tikhonova@intel.com>
5190 Ilya Tocar <ilya.tocar@intel.com>
5191 Andrey Turetskiy <andrey.turetskiy@intel.com>
5192 Ilya Verbin <ilya.verbin@intel.com>
5193 Kirill Yukhin <kirill.yukhin@intel.com>
5194 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5196 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
5197 (<sse>_comi): Ditto.
5198 (<sse>_ucomi): Ditto.
5199 (sse_cvtss2siq_2): Ditto.
5200 (sse2_cvtsd2si): Ditto.
5201 (sse2_cvtsd2siq): Ditto.
5202 (sse2_cvttsd2si): Ditto.
5203 (sse2_cvttsd2siq): Ditto.
5204 (<shift_insn><mode>3): Ditto.
5205 (sse2_cvtsi2sdq): Update constraint and prefix.
5206 (sse_cvtsi2ss): Update prefix.
5207 (sse_cvtsi2ssq): Ditto.
5209 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5211 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
5212 ignore internal calls.
5214 2013-10-11 Richard Biener <rguenther@suse.de>
5216 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
5217 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
5218 TREE_OVERFLOW is set.
5220 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
5222 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
5224 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
5225 * gimple.h: Likewise.
5227 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
5229 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
5232 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
5234 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
5235 * configure: Regenerate.
5237 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5239 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
5240 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
5241 formatting fixes, use pp_colon instead of pp_character (..., ':'),
5242 similarly pp_right_paren.
5243 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
5244 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
5245 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
5246 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
5248 (omp_declare_simd_clauses_equal,
5249 omp_remove_redundant_declare_simd_attrs): New functions.
5250 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
5251 (walk_tree_1): Handle new OpenMP 4.0 clauses.
5252 * tree.h (OMP_LOOP_CHECK): Define.
5253 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
5254 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
5255 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
5256 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
5257 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
5258 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
5259 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
5260 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
5261 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
5262 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
5263 OMP_CLAUSE_SIMDLEN_EXPR): Define.
5264 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
5265 (omp_remove_redundant_declare_simd_attrs): New prototype.
5266 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
5267 GIMPLE_OMP_TEAMS): New codes.
5268 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
5269 * omp-low.c (struct omp_context): Add cancel_label and cancellable
5271 (target_nesting_level): New variable.
5272 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
5273 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
5274 for collapse > 1 static schedule unless ordered.
5275 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
5276 (determine_parallel_type): Adjust get_ws_args_for caller.
5277 (install_var_field): Handle mask & 4 for double indirection.
5278 (scan_sharing_clauses): Ignore shared clause on teams construct.
5279 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
5280 (create_omp_child_function): If inside target or declare target
5281 constructs, set "omp declare target" attribute on the child function.
5282 (find_combined_for): New function.
5283 (scan_omp_parallel): Handle combined loops.
5284 (scan_omp_target, scan_omp_teams): New functions.
5285 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
5286 restrictions and set ctx->cancellable for cancellable constructs.
5287 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
5288 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
5289 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
5290 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
5291 (omp_clause_aligned_alignment): New function.
5292 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
5293 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
5294 on teams constructs. Handle user defined reductions and new
5296 (lower_reduction_clauses): Don't set placeholder to address of ref
5297 if it has already the right type.
5298 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
5299 (expand_parallel_call): Use the new non-_start suffixed builtins,
5300 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
5301 and GOMP_parallel_end after the call.
5302 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
5303 (expand_omp_for_init_counts): Handle combined loops.
5304 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
5306 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
5307 end of cancellable loops.
5308 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
5309 Likewise. Handle collapse > 1 loops.
5310 (expand_omp_simd): Handle combined loops.
5311 (expand_omp_for): Add inner_stmt argument, adjust callers of
5312 expand_omp_for* functions, use expand_omp_for_static*chunk even
5313 for collapse > 1 unless ordered.
5314 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
5315 of cancellable sections.
5316 (expand_omp_single): Remove need_barrier variable, just rely on
5317 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
5318 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
5319 (expand_omp_atomic_load, expand_omp_atomic_store,
5320 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
5321 (expand_omp_target): New function.
5322 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
5323 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
5324 (build_omp_regions_1): Immediately close region for
5325 GF_OMP_TARGET_KIND_UPDATE.
5326 (maybe_add_implicit_barrier_cancel): New function.
5327 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
5329 (lower_omp_single): Likewise. Add clobber after the barrier.
5330 (lower_omp_taskgroup): New function.
5331 (lower_omp_for): Handle combined loops. Adjust
5332 lower_rec_input_clauses caller. Handle cancellation.
5333 (lower_depend_clauses): New function.
5334 (lower_omp_taskreg): Lower depend clauses. Adjust
5335 lower_rec_input_clauses caller. Add clobber after the call. Handle
5337 (lower_omp_target, lower_omp_teams): New functions.
5338 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
5339 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
5340 and GOMP_cancellation_point calls.
5341 (lower_omp): Fold stmts inside of target region.
5342 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
5343 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5344 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
5345 (BT_FN_VOID_OMPFN_PTR_UINT,
5346 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
5347 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
5348 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
5349 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
5350 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
5351 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
5352 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
5353 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
5354 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
5355 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
5356 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
5357 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
5358 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
5359 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
5360 BUILT_IN_GOMP_PARALLEL_END.
5361 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
5362 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5363 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
5364 GF_OMP_FOR_KIND_DISTRIBUTE.
5365 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
5366 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
5367 (dump_gimple_omp_return): Print lhs if it has any.
5368 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
5369 gimple_omp_atomic_seq_cst_p.
5370 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
5371 and GIMPLE_OMP_TEAMS.
5372 * langhooks.c (lhd_omp_mappable_type): New function.
5373 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
5374 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
5376 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
5377 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
5378 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
5379 (struct gimplify_omp_ctx): Add combined_loop field.
5380 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
5381 on stmts inside of target region.
5382 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
5383 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
5384 ORT_TARGET and ORT_TARGET_DATA.
5385 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
5387 (omp_notice_threadprivate_variable): Complain about threadprivate
5388 variables in target region.
5389 (omp_notice_variable): Complain about vars with non-mappable type
5390 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
5391 (omp_check_private): Ignore ORT_TARGET* regions.
5392 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
5393 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
5394 (find_combined_omp_for): New function.
5395 (gimplify_omp_for): Handle gimplification of combined loops.
5396 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
5398 (gimplify_omp_target_update): New function.
5399 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
5400 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
5401 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
5402 (gimplify_body): If fndecl has "omp declare target" attribute, add
5403 implicit ORT_TARGET context around it.
5404 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
5405 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
5406 * tree-nested.c (convert_nonlocal_reference_stmt,
5407 convert_local_reference_stmt, convert_gimple_call): Handle
5408 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5409 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
5410 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5411 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5412 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5413 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5414 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5415 BUILT_IN_GOMP_LOOP_END_CANCEL,
5416 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5417 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5418 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5419 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5420 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5421 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5422 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5423 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5424 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5425 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5426 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5427 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5428 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5429 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5430 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5431 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5432 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5433 gimple_build_omp_teams): New functions.
5434 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5435 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5436 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5437 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5438 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5439 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5440 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5441 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5442 GF_OMP_ATOMIC_SEQ_CST): New.
5443 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5444 gimple_build_omp_teams): New prototypes.
5445 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5446 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5447 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5448 GIMPLE_OMP_SINGLE as end of range.
5449 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5450 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5451 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5452 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5453 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5454 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5455 gimple_omp_target_kind, gimple_omp_target_set_kind,
5456 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5457 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5458 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5459 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5460 gimple_omp_teams_set_clauses): New inlines.
5461 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5462 and GIMPLE_OMP_TASKGROUP.
5463 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5464 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5465 enum omp_clause_proc_bind_kind): New.
5466 (union omp_clause_subcode): Add depend_kind, map_kind and
5467 proc_bind_kind fields.
5468 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5469 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5470 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5471 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5472 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5474 2013-10-10 Teresa Johnson <tejohnson@google.com>
5476 * predict.c (tree_estimate_probability): Add new parameter
5477 for estimate_bb_frequencies.
5478 (estimate_bb_frequencies): Add new parameter to force estimation.
5479 (rebuild_frequencies): When max frequency in function is small,
5480 recompute counts from frequencies.
5481 * predict.h (estimate_bb_frequencies): New parameter.
5483 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5485 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5486 optimizations are disabled.
5488 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5490 * coverage.c (coverage_finish): Fix leak of da_file_name.
5492 2013-10-10 Jan Hubicka <jh@suse.cz>
5494 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5495 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5498 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5501 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5502 if any labels are in FALLTHRU_BB, use a special label emitted
5503 immediately after the asm goto insn rather than label_rtx
5505 (expand_asm_stmt): Adjust caller.
5506 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5507 edge if the last insn in predecessor is a jump with single successor,
5508 but it isn't simplejump_p.
5510 2013-10-10 Richard Biener <rguenther@suse.de>
5512 PR tree-optimization/58656
5513 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5515 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5517 * gimplify.c: Include expr.h and tm_p.h for targets with special
5518 va-arg padding requirements.
5520 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5522 * tree-flow.h: Move some prototypes to gimple.h.
5523 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5524 * gimple.h: Relocate some prototypes from tree-flow.h
5525 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5527 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5528 (build_va_arg_indirect_ref): Relocate and make static.
5529 (std_gimplify_va_arg_expr): Relocate here.
5530 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5531 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5533 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5535 * doc/md.texi: Document the mnemonic attribute.
5537 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5540 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5541 (set_attr_alternative x ...) when searching for user defined
5544 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5546 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5548 2013-10-09 Easwaran Raman <eraman@google.com>
5550 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5551 * cfgexpand.c (defer_stack_allocation): ...use here
5552 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5554 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
5556 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
5557 (operand_equal_for_value_replacement): New function, extracted from
5558 value_replacement and enhanced to catch more cases.
5559 (value_replacement): Use operand_equal_for_value_replacement.
5561 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5563 * loop-doloop.c (doloop_modify, doloop_optimize): Use
5564 get_max_loop_iterations.
5566 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5568 * config/arm/aarch-common.c (arm_early_load_addr_dep):
5569 Place comment above function.
5571 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5573 * tree-flow.h: Remove all remaining prototypes, enums and structs that
5574 are not related to tree-cfg.c.
5575 * tree-ssa-address.h: New file. Relocate prototypes.
5576 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
5577 (addr_for_mem_ref): New. Combine call to get_address_description and
5578 return addr_for_mem_ref.
5579 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
5580 * tree-ssa-live.h: Adjust prototypes.
5581 * passes.c: Include tree-ssa-live.h.
5582 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
5583 * graphite.c (graphite_transform_loops): Make static.
5584 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5585 make_pass_graphite, pass_data_graphite_transforms,
5586 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
5587 * ipa-pure-const.c (warn_function_noreturn): Make static.
5588 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5589 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5590 Relocate from tree-cfg.c
5591 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
5593 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5594 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5595 Move to ipa-pure-const.c.
5596 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
5597 Relocate from tree-optimize.c.
5598 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
5599 make_pass_fixup_cfg): Move to tree-cfg.c.
5600 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
5601 Relocate some prototypes.
5602 * tree-data-ref.h (tree_check_data_deps) Add prototype.
5603 * tree-dump.c (dump_function_to_file): Remove prototype.
5604 Add tree-flow.h to the include file.
5605 * tree-dump.h: Remove prototype.
5606 * tree-parloops.h: New File. Add prototypes.
5607 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5608 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5609 from tree-ssa-loop.c.
5610 * tree-predcom.c (run_tree_predictive_commoning,
5611 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5612 Relocate here from tree-ssa-loop.c.
5613 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5614 ssa_name_values.release ().
5615 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5616 (ssa_name_values): Relocate from tree-flow.h.
5617 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5618 * tree-ssa-loop.c (run_tree_predictive_commoning,
5619 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5620 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5621 make_pass_graphite, pass_data_graphite_transforms,
5622 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5623 tree_parallelize_loops, pass_data_parallelize_loops,
5624 make_pass_parallelize_loops): Move to other files.
5625 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5627 * tree.h: Remove prototypes from tree-address.c.
5629 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5631 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5632 (struct int_tree_map): Move to tree-hasher.h
5633 (SCALE, LABEL, PERCENT): Move to gimple.h
5634 * tree-flow-inline.h: Delete. Move functions to other files.
5635 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5636 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5637 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5638 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5639 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5640 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5641 * tree-hasher.h: Don't include tree-flow.h.
5642 (struct int_tree_map): Relocate from tree-flow.h.
5643 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5644 tree-flow-inline.h and make static.
5645 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5647 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5648 tree-flow-inline.h and make static.
5649 * tree.h (is_global_var, may_be_aliased): Relocate from
5651 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5652 * value-prof.c: No longer include tree-flow-inline.h.
5653 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5655 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5657 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5658 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5659 (enum move_pos): Move to tree-ssa-loop-im.h
5660 * cfgloop.h: Move some prototypes.
5661 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5662 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5663 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5664 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5665 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5666 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5667 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5668 (enum move_pos): Relocate here.
5669 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5671 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5672 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5673 make_pass_lim): Relocate here from tree-ssa-loop.c.
5674 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5676 (loop_edge_to_cancel, unloop_loops): Make static.
5677 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5678 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5679 (tree_complete_unroll, gate_tree_complete_unroll,
5680 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5681 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5682 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5683 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5684 (stmt_invariant_in_loop_p): Remove unused function.
5685 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5686 * tree-ssa-loop-manip.h: New file. Add prototypes.
5687 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5688 (gcov_type_to_double_int): Move to cfgloop.h.
5689 (double_int_cmp, bound_index,
5690 estimate_numbers_of_iterations_loop): Make static.
5691 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5692 (max_loop_iterations): Factor out get_max_loop_iterations.
5693 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5695 * tree-ssa-loop-niter.h: New file. Add prototypes.
5696 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5697 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5698 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5699 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5700 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5701 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5702 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5703 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5704 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5705 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5706 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5707 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5708 pass_data_complete_unroll, make_pass_complete_unroll,
5709 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5710 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5711 tree-ssa-loop-ivcanon.c.
5712 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5713 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5714 tree-ssa-loop-prefetch.c.
5715 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5717 (get_lsm_tmp_name): Relocate and add suffix parameter.
5718 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5719 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5720 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5721 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5722 (get_estimated_loop_iterations): Factor out accessor from
5723 estimated_loop_iterations in tree-ssa-loop-niter.c.
5724 (get_max_loop_iterations): Factor out accessor from
5725 _max_loop_iterations in tree-ssa-niter.c.
5726 * loop-unroll.c (decide_unroll_constant_iterations,
5727 decide_unroll_runtime_iterations, decide_peel_simple,
5728 decide_unroll_stupid): Use new get_* accessors.
5730 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5732 PR tree-optimization/20318
5733 * doc/extend.texi (returns_nonnull): New function attribute.
5734 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5736 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5737 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5739 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5742 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5743 false if both components are bitfields.
5745 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5747 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5748 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5749 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5750 * config/aarch64/aarch64.h
5751 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5752 * config/aarch64/aarch64-simd-builtins.def
5753 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5755 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5757 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5759 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5761 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5763 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5765 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5766 (vneg_f64): New intrinsic.
5767 (vneg_s8): Asm replaced with C.
5768 (vneg_s16): Likewise.
5769 (vneg_s32): Likewise.
5770 (vneg_s64): New intrinsic.
5771 (vnegq_f32): Asm replaced with C.
5772 (vnegq_f64): Likewise.
5773 (vnegq_s8): Likewise.
5774 (vnegq_s16): Likewise.
5775 (vnegq_s32): Likewise.
5776 (vnegq_s64): Likewise.
5778 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5780 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5782 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5784 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5785 packed stack special handling.
5786 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5787 back to fixed stack slots for FPRs saved due to stdarg.
5789 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5791 * config/s390/s390.c (s390_frame_info): Restructure function.
5793 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5795 * config/s390/s390.c (struct s390_frame_layout): New field
5797 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5798 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5799 regs_ever_clobbered to char*.
5800 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5801 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5802 (s390_register_info): Enable FPR save slots. Move/Copy some
5803 functionality into ...
5804 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5805 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5807 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5808 not imply a stack frame.
5809 (s390_init_frame_layout): Remove variable clobbered_regs.
5810 (s390_update_register_info): Remove function.
5811 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5812 cannot be used for register renaming.
5813 (s390_hard_regno_scratch_ok): New function.
5814 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5815 (s390_initial_elimination_offset): Change offset calculation of
5816 the return address pointer.
5817 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5818 (restore_gprs): Set frame related flag.
5819 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5820 (s390_emit_prologue): Call s390_register_info instead of
5821 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5822 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5823 (s390_optimize_prologue): Call s390_optimize_register_info.
5824 Try to remove also FPR slot save/restore INSNs. Remove frame
5825 related flags from restore INSNs.
5827 2013-10-08 DJ Delorie <dj@redhat.com>
5829 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5832 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5833 avoid conflict with the MI use of %c.
5834 * config/rl78/rl78-real.md: change %c to %C throughout.
5835 * config/rl78/rl78-virt.md: Likewise.
5837 2013-10-08 Jan Hubicka <jh@suse.cz>
5839 * config/i386/i386.c (ix86_option_override_internal): Switch
5840 to SSE math for -ffast-math when target ISA supports SSE2.
5842 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5844 * tree-flow.h: Remove some prototypes.
5845 * tree.h: Remove some protypes, add a couple.
5846 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5847 using_eh_for_cleanups_p): Add interface routines for front ends.
5848 * tree-eh.h: New file. Add protoptyes.
5849 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5850 (add_stmt_to_eh_lp_fn): Make static.
5851 (lower_try_finally): Use new using_eh_for_cleanups_p.
5852 * emit-rtl.c: Include tree-eh.h.
5853 * gimple.h: Include tree-eh.h.
5855 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5857 PR tree-optimization/58480
5858 * tree-vrp.c (infer_nonnull_range): New function.
5859 (infer_value_range): Call infer_nonnull_range.
5861 2013-10-08 Dehao Chen <dehao@google.com>
5863 PR tree-optimization/58619
5864 * tree-inline.c (copy_phis_for_bb): Combine location data
5867 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5870 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5871 RTX_FRAME_RELATED_P on INSN.
5873 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5875 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5876 (altivec_expand_vec_perm_const): Call it.
5878 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5880 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5881 sequences for LE VSX loads and stores at expand time.
5882 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5884 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5885 (rs6000_gen_le_vsx_permute): New.
5886 (rs6000_gen_le_vsx_load): New.
5887 (rs6000_gen_le_vsx_store): New.
5888 (rs6000_gen_le_vsx_move): New.
5889 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5890 (*vsx_le_perm_load_v4si): New.
5891 (*vsx_le_perm_load_v8hi): New.
5892 (*vsx_le_perm_load_v16qi): New.
5893 (*vsx_le_perm_store_v2di): New.
5894 (*vsx_le_perm_store_v4si): New.
5895 (*vsx_le_perm_store_v8hi): New.
5896 (*vsx_le_perm_store_v16qi): New.
5897 (*vsx_xxpermdi2_le_<mode>): New.
5898 (*vsx_xxpermdi4_le_<mode>): New.
5899 (*vsx_xxpermdi8_le_V8HI): New.
5900 (*vsx_xxpermdi16_le_V16QI): New.
5901 (*vsx_lxvd2x2_le_<mode>): New.
5902 (*vsx_lxvd2x4_le_<mode>): New.
5903 (*vsx_lxvd2x8_le_V8HI): New.
5904 (*vsx_lxvd2x16_le_V16QI): New.
5905 (*vsx_stxvd2x2_le_<mode>): New.
5906 (*vsx_stxvd2x4_le_<mode>): New.
5907 (*vsx_stxvd2x8_le_V8HI): New.
5908 (*vsx_stxvd2x16_le_V16QI): New.
5910 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5912 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5914 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5916 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5917 loop to work also for 31bit ABI.
5918 Save the stack pointer for frame_size > 0.
5920 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5922 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5923 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5924 constraint letters from expanders.
5925 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5926 retry count to general_operand.
5927 ("tabort"): Give operand 0 a mode.
5928 ("tabort_1"): Add mode and constraint letter for operand 0.
5929 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5931 2013-10-04 Jeff Law <law@redhat.com>
5933 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5935 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5937 (thread_across_edge): Here. Call it.
5939 2013-10-04 Cary Coutant <ccoutant@google.com>
5941 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5942 discarding a location list expression (or a piece of one).
5944 2013-10-03 Jan Hubicka <jh@suse.cz>
5946 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5947 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5948 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5950 2013-10-03 Jan Hubicka <jh@suse.cz>
5952 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5953 accumulate-outgoing-args when producing unwind info.
5955 2013-10-03 Wei Mi <wmi@google.com>
5957 * lra-constraints.c (insert_move_for_subreg): New function
5958 extracted from simplify_operand_subreg.
5959 (simplify_operand_subreg): Add reload for paradoxical subreg.
5961 2013-10-03 Rong Xu <xur@google.com>
5963 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5964 the candidate of builtin_expect such that we should fix the
5965 size/time estimation.
5966 (estimate_function_body_sizes): Do the acutally size/time fix-up
5969 2013-10-03 Rong Xu <xur@google.com>
5971 * predict.c (tree_predict_by_opcode): Get the probability
5972 for builtin_expect from param builtin_expect_probability.
5973 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5974 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5975 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5977 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5980 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5981 * calls.c (alloca_call_p): Use get_callee_fndecl.
5982 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5983 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5985 (vrp_visit_stmt): Remove duplicated code.
5987 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5989 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5990 ceildf2, btruncdf2, instead of vsx_* name.
5992 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5993 iterators to only do V2DF and V4SF here. Move the DF code to
5994 rs6000.md where it is combined with SF mode. Replace <VSv> with
5995 just 'v' since only vector operations are handled with these insns
5996 after moving the DF support to rs6000.md.
5997 (vsx_sub<mode>3): Likewise.
5998 (vsx_mul<mode>3): Likewise.
5999 (vsx_div<mode>3): Likewise.
6000 (vsx_fre<mode>2): Likewise.
6001 (vsx_neg<mode>2): Likewise.
6002 (vsx_abs<mode>2): Likewise.
6003 (vsx_nabs<mode>2): Likewise.
6004 (vsx_smax<mode>3): Likewise.
6005 (vsx_smin<mode>3): Likewise.
6006 (vsx_sqrt<mode>2): Likewise.
6007 (vsx_rsqrte<mode>2): Likewise.
6008 (vsx_fms<mode>4): Likewise.
6009 (vsx_nfma<mode>4): Likewise.
6010 (vsx_copysign<mode>3): Likewise.
6011 (vsx_btrunc<mode>2): Likewise.
6012 (vsx_floor<mode>2): Likewise.
6013 (vsx_ceil<mode>2): Likewise.
6014 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
6015 (vsx_sminsf3): Likewise.
6016 (vsx_fmadf4): Likewise.
6017 (vsx_fmsdf4): Likewise.
6018 (vsx_nfmadf4): Likewise.
6019 (vsx_nfmsdf4): Likewise.
6020 (vsx_cmpdf_internal1): Likewise.
6022 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
6023 simpler to select whether a target has SPE or traditional floating
6024 point support in iterators.
6025 (TARGET_DF_SPE): Likewise.
6026 (TARGET_SF_FPR): Likewise.
6027 (TARGET_DF_FPR): Likewise.
6028 (TARGET_SF_INSN): Macros to say whether floating point support
6029 exists for a given operation for expanders.
6030 (TARGET_DF_INSN): Likewise.
6032 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
6033 combining of SF/DF mode operations, using both traditional and VSX
6041 (abs<mode>2): Combine SF/DF modes using traditional floating point
6042 instructions. Add support for using the upper DF registers with
6043 VSX support, and SF registers with power8-vector support. Update
6044 expanders for operations supported by both the SPE and traditional
6045 floating point units.
6046 (abs<mode>2_fpr): Likewise.
6047 (nabs<mode>2): Likewise.
6048 (nabs<mode>2_fpr): Likewise.
6049 (neg<mode>2): Likewise.
6050 (neg<mode>2_fpr): Likewise.
6051 (add<mode>3): Likewise.
6052 (add<mode>3_fpr): Likewise.
6053 (sub<mode>3): Likewise.
6054 (sub<mode>3_fpr): Likewise.
6055 (mul<mode>3): Likewise.
6056 (mul<mode>3_fpr): Likewise.
6057 (div<mode>3): Likewise.
6058 (div<mode>3_fpr): Likewise.
6059 (sqrt<mode>3): Likewise.
6060 (sqrt<mode>3_fpr): Likewise.
6061 (fre<Fs>): Likewise.
6062 (rsqrt<mode>2): Likewise.
6063 (cmp<mode>_fpr): Likewise.
6064 (smax<mode>3): Likewise.
6065 (smin<mode>3): Likewise.
6066 (smax<mode>3_vsx): Likewise.
6067 (smin<mode>3_vsx): Likewise.
6068 (negsf2): Delete SF operations that are merged with DF.
6075 (fmasf4_fpr): Likewise.
6076 (fmssf4_fpr): Likewise.
6077 (nfmasf4_fpr): Likewise.
6078 (nfmssf4_fpr): Likewise.
6079 (sqrtsf2): Likewise.
6080 (rsqrtsf_internal1): Likewise.
6081 (smaxsf3): Likewise.
6082 (sminsf3): Likewise.
6083 (cmpsf_internal1): Likewise.
6084 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
6085 (negdf2): Delete DF operations that are merged with SF.
6087 (nabsdf2): Likewise.
6093 (rsqrtdf_internal1): Likewise.
6094 (fmadf4_fpr): Likewise.
6095 (fmsdf4_fpr): Likewise.
6096 (nfmadf4_fpr): Likewise.
6097 (nfmsdf4_fpr): Likewise.
6098 (sqrtdf2): Likewise.
6099 (smaxdf3): Likewise.
6100 (smindf3): Likewise.
6101 (cmpdf_internal1): Likewise.
6102 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
6103 (btrunc<mode>2): Delete separate expander, and combine with the
6104 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
6105 (btrunc<mode>2_fpr): Likewise.
6106 (ceil<mode>2): Likewise.
6107 (ceil<mode>2_fpr): Likewise.
6108 (floor<mode>2): Likewise.
6109 (floor<mode>2_fpr): Likewise.
6110 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
6111 Add support for using the upper registers with VSX and
6112 power8-vector. Move insns to be closer to the define_expands. On
6113 VSX systems, prefer the traditional form of FMA over the VSX
6114 version, since the traditional form allows the target not to
6115 overlap with the inputs.
6116 (fms<mode>4_fpr): Likewise.
6117 (nfma<mode>4_fpr): Likewise.
6118 (nfms<mode>4_fpr): Likewise.
6120 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6121 Richard Earnshaw <richard.earnshaw@arm.com>
6123 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
6124 (struct mult_cost_table): Likewise.
6125 (struct mem_cost_table): Likewise.
6126 (struct fp_cost_table): Likewise.
6127 (struct vector_cost_table): Likewise.
6128 (cpu_cost_table): Likewise.
6129 * config/arm/arm.opt (mold-rts-costs): New option.
6130 (mnew-generic-costs): Likewise.
6131 * config/arm/arm.c (generic_extra_costs): New table.
6132 (cortexa15_extra_costs): Likewise.
6133 (arm_slowmul_tune): Use NULL as new costs.
6134 (arm_fastmul_tune): Likewise.
6135 (arm_strongarm_tune): Likewise.
6136 (arm_xscale_tune): Likewise.
6137 (arm_9e_tune): Likewise.
6138 (arm_v6t2_tune): Likewise.
6139 (arm_cortex_a5_tune): Likewise.
6140 (arm_cortex_a9_tune): Likewise.
6141 (arm_v6m_tune): Likewise.
6142 (arm_fa726te_tune): Likewise.
6143 (arm_cortex_a15_tune): Use cortex15_extra_costs.
6144 (arm_cortex_tune): Use generict_extra_costs.
6145 (shifter_op_p): New function.
6146 (arm_unspec_cost): Likewise.
6147 (LIBCALL_COST): Define.
6148 (arm_new_rtx_costs): New function.
6149 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
6150 table is available. Use old costs otherwise unless mnew-generic-costs
6152 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
6153 (cpu_cost_table): Declare.
6155 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
6158 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
6159 (*subs_mul_imm_<mode>)
6160 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
6161 (*sub_<shift>_<mode>)
6162 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
6163 Remove k constraint.
6165 2013-10-03 Ian Bolton <ian.bolton@arm.com>
6167 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
6169 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
6171 2013-10-02 Teresa Johnson <tejohnson@google.com>
6173 * predict.c (probably_never_executed): New function.
6174 (probably_never_executed_bb_p): Invoke probably_never_executed.
6175 (probably_never_executed_edge_p): Ditto.
6176 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
6177 Treat profile insanities conservatively.
6179 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
6181 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
6183 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
6185 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
6186 use it. Use smaller increase for scratch. Don't increase reject
6187 for early clobber scratch.
6188 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
6189 setting eliminated regs except setting fp from hfp.
6190 (lra_eliminate): Check lra_insn_recog_data on NULL.
6192 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6195 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
6196 setting -mvsx-timode by default until the underlying problem is fixed.
6197 (RS6000_CPU, power7 defaults): Likewise.
6199 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6201 * config/x-linux (host-linux.o): Remove header dependencies.
6202 Use $(COMPILE) and $(POSTCOMPILE).
6203 * config/t-linux-android (linux-android.o): Ditto.
6205 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6207 * Makefile.in (expmed.o-warn): Remove.
6209 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6211 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
6212 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
6214 2013-10-02 Teresa Johnson <tejohnson@google.com>
6216 * dojump.c (do_jump_1): Divide probability between
6217 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
6219 2013-10-02 Tom Tromey <tromey@redhat.com>
6221 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
6223 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6225 * tree-flow.h: Remove some prototypes.
6226 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
6227 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
6228 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
6229 mark_virtual_phi_result_for_renaming): Relocate here.
6230 * tree-into-ssa.h: Add prototypes.
6231 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
6232 single_pred_before_succ_order.
6233 (blocks_in_phiopt_order): Rename and move to cfganal.c.
6234 (nonfreeing_call_p) Move to gimple.c.
6235 * cfganal.c (single_pred_before_succ_order): Move and renamed from
6237 * basic-block.h (single_pred_before_succ_order): Add prototype.
6238 * gimple.c (nonfreeing_call_p): Relocate here.
6239 * gimple.h: Add prototype.
6240 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
6241 * tree-ssa-dom.h: New file. Relocate prototypes here.
6242 * tree-ssa.h: Include tree-ssa-dom.h.
6244 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6246 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
6247 Use $(COMPILE) and $(POSTCOMPILE).
6249 * config/alpha/x-alpha (driver-alpha.o): Ditto.
6251 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6253 * tree-flow.h: Remove some prototypes.
6254 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
6255 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6256 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
6257 propagate_tree_value*): Move from here to...
6258 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
6259 propagate_tree_value*): Relocate here.
6260 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6261 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
6262 * gimple-fold.c: Remove gimple-fold.h from include list.
6263 * tree-vrp.c: Remove gimple-fold.h from include list.
6264 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
6265 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
6266 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
6267 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
6268 * sese.c: Add tree-ssa-propagate.h to include list.
6270 2013-10-02 Richard Biener <rguenther@suse.de>
6272 * tree-loop-distribution.c: Include tree-vectorizer.h for
6274 (enum partition_kind): Remove PKIND_REDUCTION.
6275 (struct partition_s): Remove has_writes member, add reduction_p member.
6276 (partition_alloc): Adjust.
6277 (partition_builtin_p): Likewise.
6278 (partition_has_writes): Remove.
6279 (partition_reduction_p): New function.
6280 (partition_merge_into): Likewise.
6281 (generate_code_for_partition): Commonize builtin partition
6283 (rdg_cannot_recompute_vertex_p): Remove.
6284 (already_processed_vertex_p): Likewise.
6285 (rdg_flag_vertex): Do not set has_writes.
6286 (classify_partition): Adjust.
6287 (rdg_build_partitions): Do not set has_writes, treat all
6288 partitions as useful.
6289 (distribute_loop): Record number of library calls generated. Adjust.
6290 (tree_loop_distribution): Report number of loops and library
6291 calls generated as opt-info.
6293 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6295 * tree-flow.h: Include new .h files. Move prototypes.
6296 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
6297 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
6298 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
6299 * tree-pretty-print.h: Add prototypes from tree-flow.h.
6300 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
6301 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
6302 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
6303 * tree.h (get_ref_base_and_extent): Move prototype out.
6304 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
6306 * gimple-low.h: New File. Add prototypes from tree-flow.h.
6307 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
6308 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
6309 * tree-scalar-evolution.c: Include tree.h.
6310 * sese.c: Include tree.h.
6311 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
6312 * dwarf2out.c: Include tree-dfa.h.
6313 * tree-chrec.c: Include tree.h.
6314 * tree-data-ref.c: Include tree.h.
6316 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
6318 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
6321 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6323 * config/t-sol2 (sol2-c.o): Remove header dependencies.
6324 Use $(COMPILE) and $(POSTCOMPILE).
6325 (sol2-cxx.o): Likewise.
6326 (sol2-stubs.o): Likewise.
6328 * config/x-solaris (host-solaris.o): Likewise.
6330 * config/sparc/t-sparc (sparc.o): Remove.
6331 (sparc-c.o): Remove header dependencies.
6332 Use $(COMPILE) and $(POSTCOMPILE).
6333 * config/sparc/x-sparc: Likewise.
6335 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
6337 * config/arc/arc-opts.h: Add 2013 to Copyright years.
6338 * config/arc/arc700.md: Likewise.
6339 * config/arc/arc-modes.def: Likewise.
6340 * config/arc/arc-simd.h: Likewise.
6341 * config/arc/t-arc-uClibc: Likewise.
6342 * config/arc/t-arc-newlib: Likewise.
6344 2013-10-02 Renlin Li <renlin.li@arm.com>
6346 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6348 (aarch64_expand_epilogue): Likewise.
6350 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6351 Yufeng Zhang <yufeng.zhang@arm.com>
6353 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
6355 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
6356 'base_in' represent a conversion and legal_cast_p_1 holds; set
6357 'base_in' with the returned value from get_unwidened.
6359 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6361 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
6362 plus_constant is not used.
6364 2013-10-01 Wei Mi <wmi@google.com>
6366 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
6367 * config/i386/i386.md: Add define_peephole2 to
6368 break partial reg stall for cvtss2sd/cvtsd2ss.
6370 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6372 * config/arc/arc.c (pass_arc_ifcvt::clone):
6373 Update for ctxt_ -> m_ctxt change.
6375 2013-10-01 Jeff Law <law@redhat.com>
6377 * tree-ssa-threadupdate.c (struct redirection_data): Delete
6378 outgoing_edge and intermediate_edge fields. Instead store the path.
6379 (redirection_data::hash): Hash on the last edge's destination index.
6380 (redirection_data::equal): Check the entire thread path.
6381 (lookup_redirectio_data): Corresponding changes.
6382 (create_edge_and_update_destination_phis): Likewise.
6383 (thread_single_edge): Likewise.
6385 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6386 Diego Novillo <dnovillo@google.com>
6388 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
6389 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
6390 (vld32wh_insn, vld32wl_insn): Delete commented-out old
6391 versions of these patterns.
6393 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
6394 (__builtin_arc_aligned): Likewise.
6396 * config/arc/arc.md: Expand adc_0 comment stating the intended
6397 purpose and why it isn't ready.
6398 Replace commented out call_value_via_label_mixed with a
6399 plain comment about bl_s.
6401 * config/arc/arc.c (stdio.h): Don't include directly.
6402 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
6403 Remove if (1) condition.
6404 (arc_encode_section_info): Fix comment.
6406 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6408 * config/arc/arc.c (arc_conditional_register_usage):
6409 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
6410 Also set reg_alloc_order for DMA config regs.
6412 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6413 Jeremy Bennett <jeremy.bennett@embecosm.com>
6415 * doc/install.texi (--with-cpu): Mention ARC.
6416 (arc-*-elf32): New paragraph.
6417 (arc-linux-uclibc): Likewise.
6418 * doc/md.texi (Machine Constraints): Add ARC part.
6419 * doc/invoke.texi: (menu): Add ARC Options.
6420 (Machine Dependent Options) <ARC Options>: Add synopsis.
6421 (node ARC Options): Add.
6422 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6423 (ARC Built-in Functions): New section defining
6424 generic ARC built-in functions.
6425 (ARC SIMD Built-in Functions): New section defining SIMD specific
6427 (Declaring Attributes of Functions): Extended
6428 description of short_call and long_call attributes for ARC and
6429 added index entries.
6431 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6432 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6433 Joern Rennecke <joern.rennecke@embecosm.com>
6434 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6435 Brendan Kehoe <brendan@zen.org>
6436 Michael Eager <eager@eagercon.com>
6437 Simon Cook <simon.cook@embecosm.com>
6438 Jeremy Bennett <jeremy.bennett@embecosm.com>
6440 * config/arc, common/config/arc: New directories.
6442 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6443 Brendan Kehoe <brendan@zen.org>
6444 Simon Cook <simon.cook@embecosm.com>
6446 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6448 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6450 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6451 * tree-ssa-coalesce.h: New. Move prototype to here.
6452 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6453 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6454 (gimple_can_coalesce_p): Move to...
6455 * gimple.c (gimple_can_coalesce_p): Here.
6457 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6459 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6460 (dump_decl_set): Move to gimple.c.
6461 * gimple.h: Don't include tree-ssa-operands.h.
6462 (dump_decl_set): Add prototype.
6463 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6464 Move to gimple-ssa.h.
6465 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6466 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6467 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6468 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6469 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6471 (dump_decl_set): Relocate here.
6472 * gimple-ssa.h: New file.
6473 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6474 Relocate from gimple.h.
6475 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6476 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6477 swap_tree_operands and remove non-ssa path.
6478 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6479 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6481 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6482 vect_is_simple_reduction_1): Use swap_ssa_operands.
6483 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6484 (enum need_phi_state): Move to tree-into-ssa.c.
6485 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6486 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6487 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6488 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6489 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6490 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6491 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6492 num_imm_uses): Move to ssa-iterators.h.
6493 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6494 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6496 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6497 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6498 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6499 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6500 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6501 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6502 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6503 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6504 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6505 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6506 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6507 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6508 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6509 (set_phi_nodes): Move to tree-phinodes.h.
6510 * tree-ssa-operands.h (enum ssa_op_iter_type,
6511 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6512 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6513 (dump_decl_set): Remove prototype.
6514 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6515 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6516 (set_phi_nodes): Relocate from tree-flow-inline.h.
6517 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6519 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6520 include list. Temporarily add gimple.h to include list.
6521 * ssa-iterators.h: New file.
6522 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6523 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6524 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6525 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6526 Relocate from tree-ssa-operands.h.
6527 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6528 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6529 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6530 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6531 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6532 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6533 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6534 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6535 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6536 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6537 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6538 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6539 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6540 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6541 Relocate from tree-flow-inline.h.
6542 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6544 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6547 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6548 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6549 Insert '\t' to output template.
6550 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6551 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6552 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6553 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
6555 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
6557 * doc/install.texi (Host/target specific installation notes for GCC):
6558 Put @anchor before @heading.
6559 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
6560 Use @email for email addresses.
6562 2013-10-01 Jeff Law <law@redhat.com>
6564 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
6565 a vec. Only delete the path if we create one without successfully
6566 registering a jump thread.
6567 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
6569 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
6570 (paths): New vector of jump threading paths.
6571 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
6572 (THREAD_PATH): New accessor macro for the entire thread path.
6573 (lookup_redirection_data): Get intermediate and final outgoing edge
6574 from the thread path.
6575 (create_edge_and_update_destination_phis): Copy the threading path.
6576 (ssa_fix_duplicate_block_edges): Get edges and block types from the
6577 jump threading path.
6578 (ssa_redirect_edges): Get edges and block types from the jump threading
6579 path. Free the path vector.
6580 (thread_block): Get edges from the jump threading path. Look at the
6581 entire path to see if we thread to a loop exit. If we cancel a jump
6582 thread request, then free the path vector.
6583 (thread_single_edge): Get edges and block types from the jump threading
6584 path. Free the path vector.
6585 (thread_through_loop_header): Get edges and block types from the jump
6586 threading path. Free the path vector.
6587 (mark_threaded_blocks): Iterate over the vector of paths and store
6588 the path on the appropriate edge. Get edges and block types from the
6589 jump threading path.
6590 (mark_threaded_blocks): Get edges and block types from the jump
6591 threading path. Free the path vector.
6592 (thread_through_all_blocks): Use the vector of paths rather than
6593 a vector of 3-edge sets.
6594 (register_jump_thread): Accept pointer to a path vector rather
6595 than the path vector itself. Store the path vector for later use.
6598 2013-10-01 Jakub Jelinek <jakub@redhat.com>
6599 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6602 * config/s390/s390.c (s390_split_branches): Modify check for table
6604 (s390_chunkify_start): Rearrange table jump insn check in order to
6605 deal with compare and branch insns correctly.
6607 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6611 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6612 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6613 define_insn_and_split.
6614 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6615 (shiftsi3_compare): New pattern.
6617 * config/arm/unspecs.md (UNSPEC_RRX): New.
6619 2013-10-01 Alan Modra <amodra@gmail.com>
6621 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6622 casing inout operands.
6624 2013-10-01 Richard Biener <rguenther@suse.de>
6626 PR tree-optimization/58553
6627 * tree-loop-distribution.c (struct partition_s): Add niter member.
6628 (classify_partition): Populate niter member for the partition
6629 and properly identify whether the relevant store happens before
6630 or after the loop exit.
6631 (generate_memset_builtin): Use niter member from the partition.
6632 (generate_memcpy_builtin): Likewise.
6634 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6636 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6637 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6639 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6641 * basic-block.h (edge_list): Prefix member names with "m_".
6642 * context.h (context): Likewise.
6643 * domwalk.h (dom_walker): Likewise.
6644 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6645 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6646 * hash-table.h (hash_table): Likewise.
6647 * machmode.h (bit_field_mode_iterator): Likewise.
6648 * pass_manager.h (pass_list): Likewise.
6649 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6650 * tree-pass.h (pass_data): Likewise.
6651 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6652 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6653 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6654 * asan.c (pass_data_asan): Update accordingly.
6655 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6656 (control_dependences::control_dependences): Likewise.
6657 (control_dependences::~control_dependences): Likewise.
6658 (control_dependences::~control_dependences): Likewise.
6659 (control_dependences::get_edges_dependent_on): Likewise.
6660 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6661 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6662 * context.c (gcc::context::context): Likewise.
6663 * cprop.c (pass_rtl_cprop::clone): Likewise.
6664 * domwalk.c (dom_walker::walk): Likewise.
6665 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6666 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6667 * mode-switching.c (pass_mode_switching::clone): Likewise.
6668 * passes.c (opt_pass::opt_pass): Likewise.
6669 (pass_manager::pass_manager): Likewise.
6670 * predict.c (pass_strip_predict_hints::clone): Likewise.
6671 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6672 (pass_split_all_insns::clone): Likewise.
6673 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6675 (bit_field_mode_iterator::next_mode): Likewise.
6676 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6677 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6678 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6679 * tree-complex.c (pass_lower_complex::clone): Likewise.
6680 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6681 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6682 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6683 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6684 (pass_fold_builtins::clone): Likewise.
6685 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6686 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6687 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6688 (pass_cd_dce::clone): Likewise.
6689 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6690 (pass_phi_only_cprop::clone): Likewise.
6691 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6692 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6693 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6694 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6695 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6696 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6697 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6698 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6699 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6700 * tree-vrp.c (pass_vrp::clone): Likewise.
6701 * tsan.c (pass_tsan::clone): Likewise.
6703 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6706 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6707 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6710 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6711 optimization, punt if sign_bit_p looked through any zero extension.
6713 2013-09-30 Teresa Johnson <tejohnson@google.com>
6715 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6716 Update redirected out edge count in joiner case.
6717 (ssa_redirect_edges): Common the joiner and non-joiner cases
6718 so that joiner case gets profile updates.
6720 2013-09-30 Richard Biener <rguenther@suse.de>
6722 PR tree-optimization/58554
6723 * tree-loop-distribution.c (classify_partition): Require
6724 unconditionally executed stores for memcpy and memset recognition.
6725 (tree_loop_distribution): Calculate dominance info.
6727 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6729 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6730 (NO_PROFILE_COUNTERS): Likewise.
6731 (PROFILE_HOOK): Likewise.
6732 (FUNCTION_PROFILER): Likewise.
6733 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6735 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6737 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6738 calls and defines in TARGET_MACHO conditional.
6739 (load_macho_picbase_di): Likewise.
6740 (reload_macho_picbase): Likewise.
6741 (reload_macho_picbase_si): Likewise.
6742 (reload_macho_picbase_di): Likewise.
6743 (nonlocal_goto_receiver): Likewise.
6745 2013-09-30 Nick Clifton <nickc@redhat.com>
6747 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6748 that use the MSP430X ISA.
6749 (msp430_option_override): Scan -mmcu command line option for any
6750 MCU name that supports the MSP430X ISA.
6751 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6752 -mmcu options which enable the MSP430X ISA.
6754 2013-09-30 Richard Biener <rguenther@suse.de>
6757 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6758 before looking for setjmp-like calls.
6760 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6763 * config/darwin-protos.h (machopic_get_function_picbase): New.
6764 * config/darwin.c (machopic_get_function_picbase): New.
6765 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6766 label for a new func. (load_macho_picbase_di): Likewise.
6767 (reload_macho_picbase): New expand.
6768 (reload_macho_picbase_si): New insn.
6769 (reload_macho_picbase_di): New insn.
6770 (nonlocal_goto_receiver): New define and split.
6771 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6772 (unspecv enum): Add UNSPECV_NLGR.
6774 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6776 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6777 that altivec registers are correctly sized on Darwin.
6779 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6781 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6782 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6783 * config/x-darwin (host-darwin.o): Likewise.
6784 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6785 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6786 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6788 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6790 * doc/invoke.texi: Fix usage of @tie{} command.
6792 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6794 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6796 2013-09-28 Jan Hubicka <jh@suse.cz>
6798 * config/i386/x86-tune.def: Add documentation for each of the options;
6801 2013-09-28 Jan Hubicka <jh@suse.cz>
6803 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6805 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6806 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6807 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6808 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6809 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6811 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6813 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6814 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6815 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6816 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6817 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6818 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6819 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6820 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6821 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6822 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6823 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6824 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6825 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6826 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6827 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6828 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6829 gimple.h, godump.c, graphite-clast-to-gimple.c,
6830 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6831 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6832 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6833 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6834 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6835 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6836 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6837 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6838 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6839 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6840 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6841 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6842 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6843 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6844 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6845 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6846 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6847 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6848 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6849 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6850 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6851 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6852 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6853 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6854 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6855 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6856 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6857 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6858 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6859 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6860 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6861 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6862 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6863 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6864 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6865 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6866 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6867 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6868 whitespace before "(".
6870 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6872 * expr.h (extract_bit_field): Remove packedp parameter.
6873 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6874 from forward declaration.
6875 (store_split_bit_field): Remove packedp arg from calls to
6876 extract_fixed_bit_field.
6877 (extract_bit_field_1): Remove packedp parameter and packedp
6878 argument from recursive calls and calls to extract_fixed_bit_field.
6879 (extract_bit_field): Remove packedp parameter and corresponding
6880 arg to extract_bit_field_1.
6881 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6883 (extract_split_bit_field): Remove packedp arg from call to
6884 extract_fixed_bit_field.
6885 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6886 (copy_blkmode_from_reg): Likewise.
6887 (copy_blkmode_to_reg): Likewise.
6888 (read_complex_part): Likewise.
6889 (store_field): Likewise.
6890 (expand_expr_real_1): Likewise.
6891 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6892 to extract_bit_field.
6893 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6894 call to extract_bit_field.
6895 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6896 call to extract_bit_field.
6897 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6898 and special packedp behavior from -fstrict-volatile-bitfields
6901 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6903 * lra-eliminations.c (init_elim_table): Guard value_p.
6905 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6907 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6908 DFmode, DImode, and SFmode in the upper VSX registers based on the
6909 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6910 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6911 of the VSX register handling.
6913 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6915 (zero_extendsidi2_lfiwzx): Add support for loading into the
6916 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6917 only do VSX memory options on Altivec registers.
6918 (extendsidi2_lfiwax): Likewise.
6919 (extendsfdf2_fpr): Likewise.
6920 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6921 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6922 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6923 (movdi_internal64): Likewise.
6925 2013-09-27 Xinliang David Li <davidxl@google.com>
6927 * opts.c (finish_options): Adjust parameters
6928 according to vect cost model.
6929 (common_handle_option): Set dynamic vect cost
6931 targhooks.c (default_add_stmt_cost): Compute stmt cost
6933 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6934 Use helper function.
6935 * tree-vectorizer.h (unlimited_cost_model): New function.
6936 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6937 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6939 (vect_enhance_data_refs_alignment): Ditto.
6940 * flag-types.h: New enum.
6941 * common/config/i386/i386-common.c (ix86_option_init_struct):
6942 No need to initialize vect_cost_model flag.
6943 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6946 2013-09-27 Diego Novillo <dnovillo@google.com>
6948 * gimple.h (enum ssa_mode): Remove.
6950 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6952 * cfgloop.h (number_of_loops): Fix typo in check for null.
6954 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6957 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6958 are children of outermost saved_cfun's loop, and set it up to
6959 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6960 if loop != loop0 and SESE region contains bbs that belong to loop0.
6962 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6964 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6965 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6966 (set_address_segment, set_address_base, set_address_index)
6967 (set_address_disp): Accept the argument unconditionally.
6968 (baseness): Remove must_be_base_p and must_be_index_p checks.
6969 (decompose_normal_address): Classify as much as possible in the
6972 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6974 * cse.c (count_reg_usage): Handle INT_LIST.
6975 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6976 * reginfo.c (reg_scan_mark_refs): Likewise.
6977 * reload1.c (eliminate_regs_1): Likewise.
6979 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6982 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6985 2013-09-27 Richard Biener <rguenther@suse.de>
6987 PR tree-optimization/58459
6988 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6989 restriction not propagating into loops.
6991 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6993 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6994 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6995 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6997 2013-09-26 Richard Biener <rguenther@suse.de>
6999 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
7001 2013-09-26 Richard Biener <rguenther@suse.de>
7003 * alias.h (component_uses_parent_alias_set): Rename to ...
7004 (component_uses_parent_alias_set_from): ... this.
7005 * alias.c (component_uses_parent_alias_set): Rename to ...
7006 (component_uses_parent_alias_set_from): ... this and return
7008 (reference_alias_ptr_type_1): Use the result from
7009 component_uses_parent_alias_set_from instead of stripping
7010 components one at a time.
7011 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
7013 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7015 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
7016 Move prototypes to...
7017 * tree-ssa-ter.h: New File. Move prototypes here.
7018 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
7019 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
7020 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
7021 from is_replaceable_p.
7022 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
7023 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
7024 (process_replaceable): Use ter_is_replaceable_p.
7025 (find_replaceable_in_bb): Use ter_is_replaceable_p.
7026 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
7027 newly refactored ssa_is_replaceable_p.
7028 * cfgexpand.c: Include tree-outof-ssa.h.
7029 * ssaexpand.h: Delete.
7031 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7033 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
7034 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
7036 (create_gimple_tmp): Delete.
7037 (get_expr_type, build_assign, build_type_cast): Move to...
7038 * gimple-builder.c: New File.
7039 (get_expr_type): Relocate from gimple.c.
7040 (build_assign, build_type_cast): Change to only create ssanames.
7041 * gimple.h: Move prototypes to...
7042 * gimple-builder.h: New File. Here.
7043 * tree-ssa.h: And here.
7044 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
7045 count_uses_and_derefs): Relocate from gimple.c.
7046 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
7047 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
7048 * tree-ssa-math-opts (execute_cse_reciprocals): Use
7049 gimple_replace_ssa_lhs.
7050 * asan.c: Include gimple-builder.h.
7051 * Makefile.in: Add gimple-builder.o.
7053 2013-09-26 Richard Biener <rguenther@suse.de>
7055 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
7056 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
7057 (loe_visit_block): Use gcc_checking_assert.
7058 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
7059 gimple_assign_ssa_name_copy_p.
7060 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
7063 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
7065 * config/rs6000/t-rs6000 (rs6000.o): Remove.
7066 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
7068 2013-09-26 Richard Biener <rguenther@suse.de>
7070 PR tree-optimization/58539
7071 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
7072 the fact that debug statements are not taking part in loop-closed
7075 2013-09-26 Nick Clifton <nickc@redhat.com>
7077 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
7078 time warning message.
7079 (msp430_print_operand_raw): Delete unused letter parameter.
7080 (TARGET_PRINT_OPERAND_ADDRESS): Define.
7081 (msp430_print_operand_address): New function.
7082 (msp430_print_operand): Move address printing code from here to
7084 * config/msp430/msp430.md (movsipsi2): Add comment in generated
7086 (zero_extendpsisi2): Likewise.
7087 (extendpsisi2): New pattern.
7088 (andneghi3): New pattern.
7090 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
7092 * config/aarch64/aarch64.opt (mlra): New option.
7093 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
7094 (TARGET_LRA_P): Define.
7096 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
7098 * expr.c (expand_assignment): Remove obsolete comment.
7100 2013-09-25 Jeff Law <law@redhat.com>
7102 * tree-flow.h (thread_through_all_blocks): Prototype moved into
7103 tree-ssa-threadupdate.h.
7104 (register_jump_thread): Similarly.
7105 * tree-ssa-threadupdate.h: New header file.
7106 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
7107 * tree-vrp.c: Likewise.
7108 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
7109 (thread_around_empty_blocks): Change type of path vector argument to
7110 an edge,type pair from just an edge. Initialize both elements when
7111 appending to a jump threading path. Tweak references to elements
7113 (thread_across_edge): Similarly. Release memory for the elements
7115 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
7116 (dump_jump_thread_path): New function broken out from
7117 register_jump_thread.
7118 (register_jump_thread): Use dump_jump_thread_path. Change type of
7119 path vector entries. Search the path for NULL edges and dump
7120 the path if one is found. Tweak the conversion of path to 3-edge
7121 form to use the block copy type information embedded in the path.
7123 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7125 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
7127 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7128 Vladimir Makarov <vmakarov@redhat.com>
7130 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
7131 from the least significant bit.
7132 (strip_address_mutations): Add bitfield operations handling.
7133 (must_be_index_p): Add shifting and rotate operations handling.
7134 (set_address_base): Use must_be_base_p predicate.
7135 (set_address_index): Use must_be_index_p predicate.
7137 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
7138 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7139 Sergey Lega <sergey.s.lega@intel.com>
7140 Anna Tikhonova <anna.tikhonova@intel.com>
7141 Ilya Tocar <ilya.tocar@intel.com>
7142 Andrey Turetskiy <andrey.turetskiy@intel.com>
7143 Ilya Verbin <ilya.verbin@intel.com>
7144 Kirill Yukhin <kirill.yukhin@intel.com>
7145 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7147 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
7149 (ix86_expand_vector_move_misalign): Support new unaligned load and
7150 stores and use new names.
7151 (CODE_FOR_sse2_storedqu): Rename to ...
7152 (CODE_FOR_sse2_storedquv16qi): ... this.
7153 (CODE_FOR_sse2_loaddqu): Rename to ...
7154 (CODE_FOR_sse2_loaddquv16qi): ... this.
7155 (CODE_FOR_avx_loaddqu256): Rename to ...
7156 (CODE_FOR_avx_loaddquv32qi): ... this.
7157 (CODE_FOR_avx_storedqu256): Rename to ...
7158 (CODE_FOR_avx_storedquv32qi): ... this.
7159 * config/i386/i386.md (fpint_logic): New.
7160 * config/i386/sse.md (VMOVE): Extend for AVX512.
7164 (VI_UNALIGNED_LOADSTORE): Ditto.
7165 (sse2_avx_avx512f): Ditto.
7166 (sse2_avx2): Extend for AVX512.
7167 (sse4_1_avx2): Ditto.
7168 (avx2_avx512f): New.
7169 (sse): Extend for AVX512.
7172 (avxsizesuffix): Ditto.
7173 (sseintvecmode): Ditto.
7175 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
7176 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
7177 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7178 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
7179 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7180 (<sse2_avx_avx512f>_storedqu<mode): ... this.
7181 (<sse>_movnt<mode>): Replace constraint "x" with "v".
7182 (STORENT_MODE): Extend for AVX512.
7183 (*absneg<mode>2): Replace constraint "x" with "v".
7184 (*mul<mode>3): Ditto.
7185 (*ieee_smin<mode>3): Ditto.
7186 (*ieee_smax<mode>3): Ditto.
7187 (avx_cmp<mode>3): Replace VF with VF_128_256.
7188 (*<sse>_maskcmp<mode>3_comm): Ditto.
7189 (<sse>_maskcmp<mode>3): Ditto.
7190 (<sse>_andnot<mode>3): Extend for AVX512.
7191 (<code><mode>3, anylogic): Replace VF with VF_128_256.
7192 (<code><mode>3, fpint_logic): New.
7193 (*<code><mode>3): Extend for AVX512.
7194 (avx512flogicsuff): New.
7195 (avx512f_<logic><mode>): Ditto.
7196 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
7198 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
7199 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
7200 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
7201 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
7202 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
7203 (xop_vpermil2<mode>3): Ditto.
7204 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
7205 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
7206 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
7207 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
7209 2013-09-25 Tom Tromey <tromey@redhat.com>
7211 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
7212 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
7213 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
7214 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
7215 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
7216 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
7217 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
7218 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
7219 (GRAPHITE_HTAB_H): Remove.
7221 2013-09-25 Tom Tromey <tromey@redhat.com>
7223 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
7225 2013-09-25 Tom Tromey <tromey@redhat.com>
7227 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
7229 2013-09-25 Tom Tromey <tromey@redhat.com>
7231 * config/i386/t-i386 (i386.o): Remove.
7232 (i386-c.o): Use COMPILE and POSTCOMPILE.
7234 2013-09-25 Tom Tromey <tromey@redhat.com>
7236 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
7238 2013-09-25 Tom Tromey <tromey@redhat.com>
7240 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
7241 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
7242 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
7243 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
7244 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
7245 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
7246 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
7247 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
7248 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
7249 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
7250 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
7251 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
7252 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
7253 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
7254 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
7255 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
7256 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
7257 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
7258 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
7259 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
7260 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
7261 (tree-ssa-pre.o, tree-ssa-sccvn.o)
7262 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
7263 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
7264 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
7265 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
7266 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
7267 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
7268 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
7269 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
7270 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
7271 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
7272 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
7273 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
7274 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
7275 (graphite-clast-to-gimple.o, graphite-dependences.o)
7276 (graphite-interchange.o, graphite-poly.o)
7277 (graphite-scop-detection.o, graphite-sese-to-poly.o)
7278 (graphite-optimize-isl.o, tree-vect-loop.o)
7279 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
7280 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
7281 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
7282 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
7283 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
7284 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
7285 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
7286 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
7287 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
7288 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
7289 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
7290 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
7291 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
7292 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
7293 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
7294 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
7295 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
7296 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
7297 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
7298 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
7299 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
7300 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
7301 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
7302 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
7303 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
7304 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
7305 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
7306 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
7307 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
7308 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
7309 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
7310 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
7311 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
7312 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
7313 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
7314 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
7315 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
7316 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
7317 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
7318 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
7319 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
7320 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
7321 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
7322 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
7323 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
7324 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
7325 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
7326 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
7327 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
7328 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
7329 (gcov-dump.o): Remove.
7330 (default-c.o): Use COMPILE and POSTCOMPILE.
7331 (CFLAGS-gcc.o): New variable.
7332 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
7334 2013-09-25 Tom Tromey <tromey@redhat.com>
7336 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
7337 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
7338 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
7339 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
7340 (c-family/c-pragma.o, c-family/c-pretty-print.o)
7341 (c-family/c-semantics.o, c-family/c-ada-spec.o)
7342 (c-family/array-notation-common.o, c-family/stub-objc.o)
7343 (c-family/c-ubsan.o): Remove.
7345 2013-09-25 Tom Tromey <tromey@redhat.com>
7347 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
7349 2013-09-25 Tom Tromey <tromey@redhat.com>
7351 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
7352 to add -DENABLE_SHARED_LIBGCC.
7353 (gcc.o): Don't use subshell.
7355 2013-09-25 Tom Tromey <tromey@redhat.com>
7357 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
7358 * configure.ac: Don't invoke AM_PROG_CC_C_O.
7359 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
7360 * configure, config.in: Rebuild.
7362 2013-09-25 Tom Tromey <tromey@redhat.com>
7364 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
7365 (COMPILE, POSTCOMPILE): New variables.
7366 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
7367 (DEPFILES): New variable.
7368 Include ".Po" files.
7369 * configure.ac: Add checks for dependency checking.
7370 * configure, aclocal.m4: Regenerate.
7372 2013-09-25 Tom Tromey <tromey@redhat.com>
7374 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7375 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
7377 2013-09-25 Tom Tromey <tromey@redhat.com>
7379 * Makefile.in (generated_files): Add options.h,
7380 target-hooks-def.h, insn-opinit.h,
7381 common/common-target-hooks-def.h, pass-instances.def,
7382 c-family/c-target-hooks-def.h.
7384 2013-09-25 Jeff Law <law@redhat.com>
7386 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
7387 than foo[foo.length () - 1] to access last member in a vec.
7388 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
7390 2013-09-25 Richard Biener <rguenther@suse.de>
7393 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
7395 2013-09-25 Jan Hubicka <jh@suse.cz>
7397 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
7400 2013-09-25 Marek Polacek <polacek@redhat.com>
7403 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
7404 when determining the type name.
7406 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
7408 * config/sh/sh.md: Fix formatting.
7410 2013-09-24 Xinliang David Li <davidxl@google.com>
7412 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7413 max peel iterations parameter.
7414 * param.def: New parameter.
7415 * doc/invoke.texi: Document New parameter.
7417 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7419 * gimple-pretty-print.c: Various whitespace tweaks.
7420 * tree-core.h: Likewise.
7421 * tree-pretty-print.c: Likewise.
7422 * tree-ssa-alias.c: Likewise.
7423 * tree-ssa-copy.c: Likewise.
7424 * tree-ssanames.c: Likewise.
7425 * tree-ssanames.h: Likewise.
7426 * tree-vrp.c: Likewise.
7428 2013-09-24 Alan Modra <amodra@gmail.com>
7432 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7433 for output operands that disallow regs. Don't use EXPAND_WRITE on
7436 2013-09-24 Richard Biener <rguenther@suse.de>
7439 * tree.c (reference_alias_ptr_type): Move ...
7440 * alias.c (reference_alias_ptr_type): ... here and implement
7441 in terms of the new reference_alias_ptr_type_1.
7442 (ref_all_alias_ptr_type_p): New helper.
7443 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7444 use ref_all_alias_ptr_type_p.
7445 (get_deref_alias_set): Add flag_strict_aliasing check here.
7446 (reference_alias_ptr_type_1): New function, split out from ...
7447 (get_alias_set): ... here.
7448 (alias_ptr_types_compatible_p): New function.
7449 * alias.h (reference_alias_ptr_type): Declare.
7450 (alias_ptr_types_compatible_p): Likewise.
7451 * tree.h (reference_alias_ptr_type): Remove.
7452 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7453 to compare MEM_REF alias types.
7455 2013-09-24 Richard Biener <rguenther@suse.de>
7457 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7459 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7461 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7462 reload helper function arrays into a single array reg_addr.
7463 (reload_fpr_gpr): Likewise.
7464 (reload_gpr_vsx): Likewise.
7465 (reload_vsx_gpr): Likewise.
7466 (struct rs6000_reg_addr): Likewise.
7467 (reg_addr): Likewise.
7468 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7469 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7470 (rs6000_init_hard_regno_mode_ok): Likewise.
7471 (rs6000_secondary_reload_direct_move): Likewise.
7472 (rs6000_secondary_reload): Likewise.
7474 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7475 constraints: wu, ww, and wy. Repurpose wv constraint added during
7476 power8 changes. Put wg constraint in alphabetical order.
7478 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7479 for future work to add ISA 2.07 VSX single precision support.
7480 (-mvsx-scalar-double): Change default from -1 to 1, update
7481 documentation comment.
7482 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7483 (-mupper-regs-df): New debug switch to control whether DF values
7484 can go in the traditional Altivec registers.
7485 (-mupper-regs-sf): New debug switch to control whether SF values
7486 can go in the traditional Altivec registers.
7488 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7490 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7491 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7492 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7493 (rs6000_debug_legitimate_address_p): Print if we are running
7494 before, during, or after reload.
7495 (rs6000_secondary_reload): Add a comment.
7496 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7498 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7499 constraints. Update documentation string.
7500 (wd constraint): Likewise.
7501 (wf constraint): Likewise.
7502 (wg constraint): Likewise.
7503 (wn constraint): Likewise.
7504 (ws constraint): Likewise.
7505 (wt constraint): Likewise.
7506 (wx constraint): Likewise.
7507 (wz constraint): Likewise.
7508 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7510 (ww constraint): Likewise.
7511 (wy constraint): Likewise.
7512 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7513 in the previous submissions.
7514 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7516 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7518 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7519 (REG_BR_PROB): Say that the probability is stored in an int_list.
7520 * reg-notes.def: Update commentary to mention INT_LIST.
7521 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7522 (INT_LIST): New rtx.
7523 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7524 * rtlanal.c (int_reg_note_p): New function.
7525 (alloc_reg_note): Assert that the note does not have an int argument.
7526 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7527 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7528 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7529 rather than an INSN_LIST. Handle INT_LIST.
7530 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7531 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7532 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7533 Manipulate them as ints rather than rtxes.
7534 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7535 * regmove.c (copy_src_to_dest): Likewise.
7536 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7538 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7539 into the cases that need it.
7540 * config/arm/arm.c (arm_unwind_emit): Likewise.
7542 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7543 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7544 * loop-doloop.c (add_test, doloop_modify): Likewise.
7545 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7546 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7547 * predict.c (combine_predictions_for_insn): Likewise.
7548 * print-rtl.c (print_rtx): Handle INT_LIST.
7549 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7550 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7551 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7552 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7553 (ix86_split_fp_branch, predict_jump): Likewise.
7554 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
7555 * config/sh/sh.c (expand_cbranchsi4): Likewise.
7556 * config/spu/spu.c (ea_load_store_inline): Likewise.
7558 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
7559 value of a REG_BR_PROB note.
7560 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
7561 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
7562 * emit-rtl.c (try_split): Likewise.
7563 * predict.c (br_prob_note_reliable_p): Likewise.
7564 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
7565 * reorg.c (mostly_true_jump): Likewise.
7566 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
7567 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
7568 * config/i386/i386.c (ix86_print_operand): Likewise.
7569 * config/ia64/ia64.c (ia64_print_operand): Likewise.
7570 * config/mmix/mmix.c (mmix_print_operand): Likewise.
7571 * config/rs6000/rs6000.c (output_cbranch): Likewise.
7572 * config/s390/s390.c (s390_expand_tbegin): Likewise.
7573 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
7574 * config/sparc/sparc.c (output_cbranch): Likewise.
7575 * config/spu/spu.c (get_branch_target): Likewise.
7576 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
7577 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
7579 2013-09-23 Jan Hubicka <jh@suse.cz>
7581 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
7583 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
7584 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
7586 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
7587 (gimple_fold_call): Dump inconsistent devirtualizations; add
7588 sanity check for type based devirtualizations.
7589 * ipa-prop.c: Include ipa-utils.h
7590 (ipa_intraprocedural_devirtualization): Add sanity check.
7591 (try_make_edge_direct_virtual_call): Likewise.
7593 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
7595 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
7596 assignment statements.
7598 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
7600 * gimple-pretty-print.c (dump_ssaname_info): New function.
7601 (dump_gimple_phi): Call it.
7602 (pp_gimple_stmt_1): Likewise.
7603 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7604 (range_info_def): Declare.
7605 * tree-pretty-print.c (pp_double_int): New function.
7606 (dump_generic_node): Call it.
7607 * tree-pretty-print.h (pp_double_int): Declare.
7608 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7609 * tree-ssanames.h (range_info_def): New structure.
7610 (value_range_type): Move definition here.
7611 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7613 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7615 (set_range_info): New function.
7616 (get_range_info): Likewise.
7617 (duplicate_ssa_name_range_info): Likewise.
7618 (duplicate_ssa_name_fn): Check pointer type and call
7619 duplicate_ssa_name_range_info.
7620 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7621 * tree-vrp.c (value_range_type): Remove definition, now in
7623 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7624 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7625 (SSA_NAME_RANGE_INFO): New macro.
7627 2013-09-23 Richard Biener <rguenther@suse.de>
7629 PR tree-optimization/58464
7630 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7631 (phi_trans_add): Change to add conditionally on being not present.
7632 (phi_translate_1): Remove recursion detection here.
7633 (phi_translate): Pre-seed the cache with NULL to catch
7634 recursion here in a more generic way.
7635 (bitmap_find_leader): Adjust comment.
7636 (get_representative_for): Dump value-numbers.
7637 (create_expression_by_pieces): Likewise.
7638 (insert_into_preds_of_block): Likewise.
7640 2013-09-23 Christian Bruel <christian.bruel@st.com>
7643 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7644 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7646 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7649 2013-09-20 Renlin Li <renlin.li@arm.com>
7651 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7653 (aarch64_expand_epilogue): Likewise.
7654 (aarch64_legitimize_reload_address): Likewise.
7656 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7658 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7659 NULL_TREE before pushing them onto the vector. Likewise for labels.
7661 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7663 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7664 comparison operators when -fno-trapping-math is in effect.
7665 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7666 comparison operators in TFmode and assert that unsupported operators
7668 (ia64_print_operand): Likewise.
7670 2013-09-21 Jan Hubicka <jh@suse.cz>
7672 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7673 (sse_typeless_stores): Enable for core
7674 (sse_load0_by_pxor): Likewise.
7675 (four_jump_limit): Disable for core.
7676 (pad_returns): Likewise.
7677 (avoid_vector_decode): Likewise.
7678 (fuse_cmp_and_branch): Enable for cores.
7679 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7681 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7684 * config/pa/pa.c (pa_option_override): Disable auto increment and
7685 decrement instructions until reload is completed.
7687 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7688 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7689 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7691 2013-09-20 DJ Delorie <dj@redhat.com>
7692 Nick Clifton <nickc@redhat.com>
7694 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7695 (need_to_save): Save bank 0 on interrupts.
7696 (characterize_address): Strip far address wrappers.
7697 (rl78_as_legitimate_address): Likewise.
7698 (transcode_memory_rtx): Likewise.
7699 (rl78_peep_movhi_p): Disable this peephole after devirt.
7700 (rl78_propogate_register_origins): Forget all origins when a
7702 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7703 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7705 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7706 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7707 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7709 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7711 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7713 * config/pa/pa.md: In "scc" insn patterns, change output template to
7714 handle const0_rtx in reg_or_0_operand operands.
7716 2013-09-20 Martin Husemann <martin@NetBSD.org>
7719 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7720 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7722 2013-09-20 Richard Biener <rguenther@suse.de>
7725 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7726 remembering SSA name version and block index.
7727 (new_scev_info_str): Adjust.
7728 (hash_scev_info): Likewise. Also hash the block index.
7729 (eq_scev_info): Adjust.
7730 (find_var_scev_info): Likewise.
7731 (struct instantiate_cache_entry): Remove.
7732 (struct instantiate_cache_type): Use a htab to map name, block
7734 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7735 (get_instantiated_value_entry): Likewise.
7736 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7737 (instantiate_scev_name): Adjust.
7739 2013-09-20 Jeff Law <law@redhat.com>
7741 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7743 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7745 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7746 Call aarch64_simd_expand_args to update op[argc].
7748 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7750 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7753 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7755 * gengtype.c (file_rules): Added rule for *.cc files.
7756 (get_output_file_with_visibility): Give fatal message when no
7759 2013-09-20 Renlin Li <renlin.li@arm.com>
7761 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7762 (aarch64_expand_epilogue): Likewise.
7763 (aarch64_legitimize_reload_address): Likewise.
7765 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7768 * expr.c (expand_assignment): Remove misalignp code path.
7770 2013-09-20 Marek Polacek <polacek@redhat.com>
7773 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7774 TYPE_PRECISION. Add asserts.
7776 2013-09-20 Richard Biener <rguenther@suse.de>
7778 PR tree-optimization/58453
7779 * tree-loop-distribution.c (distribute_loop): Apply the cost
7780 model for -ftree-loop-distribute-patterns, too.
7782 2013-09-20 Richard Biener <rguenther@suse.de>
7785 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7786 make type comparison less strict.
7788 2013-09-20 Alan Modra <amodra@gmail.com>
7790 * configure: Regenerate.
7791 * aclocal.m4: Regenerate.
7793 2013-09-20 Marek Polacek <polacek@redhat.com>
7796 * doc/extend.texi: Document that attribute used is meant to be used
7797 on variables with static storage duration.
7799 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7801 PR tree-optimization/58472
7802 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7803 simd_lane_access set inv_p = false.
7804 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7805 the simduid magic VAR_DECL.
7807 2013-09-19 Jan Hubicka <jh@suse.cz>
7809 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7811 2013-09-17 Jeff Law <law@redhat.com>
7813 * tree-ssa-dom.c (record_temporary_equivalences): New function
7814 split out of dom_opt_dom_walker::after_dom_children.
7815 (dom_opt_dom_walker::thread_across_edge): Move common code
7816 in here from dom_opt_dom_walker::after_dom_children.
7817 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7819 2013-09-19 Jan Hubicka <jh@suse.cz>
7821 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7822 (TARGET_GENERIC): Use PROCESOR_GENERIC
7823 (enum processor_type): Unify generic32 and 64.
7824 * i386.md (cpu): Likewise.
7825 * x86-tune.def (use_leave): Enable for generic32.
7826 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7827 * athlon.md: Change generic64 to generic in all occurences.
7828 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7829 (ix86_target_macros_internal): Likewise.
7830 * driver-i386.c (host_detect_local_cpu): Likewise.
7831 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7833 (generic_memcpy, generic_memset, generic_cost): This one.
7834 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7835 (m_GENERIC32, m_GENERIC64): Remove.
7836 (m_GENERIC): Turn into one flag.
7837 (processor_target): Unify generic tunnings.
7838 (ix86_option_override_internal): Replace generic32/64 by generic.
7839 (ix86_issue_rate): Likewise.
7840 (ix86_adjust_cost): Likewise.
7842 2013-09-19 Jan Hubicka <jh@suse.cz>
7844 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7845 of speculative flag.
7847 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7849 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7850 GOMP_sections_start, even if !exit_reachable.
7852 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7854 * lra-constraints.c (need_for_all_save_p): Use macro
7855 HARD_REGNO_CALL_PART_CLOBBERED.
7856 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7857 set up pseudo conflict hard regs.
7859 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7862 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7865 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7867 PR rtl-optimization/58438
7868 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7869 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7870 from previous subpasses.
7872 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7874 * arm.c (arm_get_frame_offsets): Validate architecture supports
7875 LDRD/STRD before accepting the tuning preference.
7876 (arm_expand_prologue): Likewise.
7877 (arm_expand_epilogue): Likewise.
7879 2013-09-18 Richard Biener <rguenther@suse.de>
7881 PR tree-optimization/58417
7882 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7883 have chrecs with symbols defined in the loop as operands.
7884 (chrec_fold_multiply): Likewise.
7885 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7886 parameters before folding binary operations.
7887 (struct instantiate_cache_entry_hasher): Remove.
7888 (struct instantiate_cache_type): Use a pointer-map.
7889 (instantiate_cache_type::instantiate_cache_type): New function.
7890 (instantiate_cache_type::get): Likewise.
7891 (instantiate_cache_type::set): Likewise.
7892 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7893 (get_instantiated_value_entry): Likewise.
7894 (global_cache): New global.
7895 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7896 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7897 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7898 (instantiate_scev_name): Adjust.
7899 (instantiate_scev): Construct global instead of local cache.
7900 (resolve_mixers): Likewise.
7902 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7903 Paolo Carlini <paolo.carlini@oracle.com>
7906 * doc/implement-cxx.texi: Fix references to the C++ standards.
7908 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7910 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7911 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7912 simd_lane_access drs, update also DR_ALIGNED_TO.
7914 2013-09-18 Marek Polacek <polacek@redhat.com>
7917 * doc/extend.texi: Document no_sanitize_undefined attribute.
7918 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7919 no_sanitize_undefined attribute.
7921 2013-09-18 Nick Clifton <nickc@redhat.com>
7923 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7924 (ASM_DECLARE_FUNCTION_NAME): Define.
7926 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7928 * compare-elim.c (find_comparison_dom_walker): New class
7929 (find_comparisons_in_bb): Rename to
7930 find_comparison_dom_walker::before_dom_children
7931 (find_comparisons): Adjust
7932 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7934 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7935 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7936 (init_walk_dominator_tree): Remove declaration.
7937 (fini_walk_dominator_tree): Remove declaration.
7938 * fwprop.c (single_def_use_dom_walker): New class
7939 (single_def_use_enter_block): Convert to
7940 single_def_use_dom_walker::before_dom_children.
7941 (single_def_use_leave_block): Convert to
7942 single_def_use_dom_walker::after_dom_children.
7943 (build_single_def_use_links): Adjust.
7944 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7946 (find_candidates_in_block): Convert to
7947 find_candidates_dom_walker::before_dom_children.
7948 (execute_strength_reduction): Adjust.
7949 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7950 (sese_dom_walker): New class.
7951 (sese_dom_walker::sese_dom_walker): New constructor.
7952 (sese_dom_walker::~sese_dom_walker): New destructor.
7953 (build_sese_conditions_before): Convert to
7954 sese_dom_walker::before_dom_children.
7955 (build_sese_conditions_after): Convert to
7956 sese_dom_walker::after_dom_children.
7957 (build_poly_scop): Adjust
7958 * tree-into-ssa.c (rewrite_dom_walker): New class
7959 (rewrite_enter_block): Convert to
7960 rewrite_dom_walker::before_dom_children.
7961 (rewrite_leave_block): Convert to
7962 rewrite_dom_walker::after_dom_children.
7963 (rewrite_update_dom_walker): New class.
7964 (rewrite_update_enter_block): Convert to
7965 rewrite_update_dom_walker::before_dom_children.
7966 (rewrite_update_leave_block): Convert to
7967 rewrite_update_dom_walker::after_dom_children.
7968 (rewrite_blocks, rewrite_into_ssa): Adjust.
7969 (mark_def_dom_walker): New class.
7970 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7971 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7972 (mark_def_sites_blocks): Convert to
7973 mark_def_dom_walker::before_dom_children.
7974 (mark_def_site_blocks): Remove.
7975 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7976 (tree_ssa_dominator_optimize): Adjust.
7977 (dom_thread_across_edge): Convert to method
7978 dom_opt_dom_walker::thread_across_edge.
7979 (dom_opt_enter_block): Convert to member function
7980 dom_opt_dom_walker::before_dom_children.
7981 (dom_opt_leave_block): Convert to member function
7982 dom_opt_dom_walker::after_dom_children.
7983 * tree-ssa-dse.c (dse_dom_walker): New class.
7984 (dse_enter_block): Convert to member function
7985 dse_dom_walker::before_dom_children.
7986 (tree_ssa_dse): Adjust.
7987 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7988 (determine_invariantness_stmt): Convert to method
7989 invariantness_dom_walker::before_dom_children.
7990 (determine_invariantness): Remove
7991 (move_computations_dom_walker): New class.
7992 (move_computations_stmt): Convert to method
7993 move_computations_dom_walker::before_dom_children.
7994 (move_computations, tree_ssa_lim): Adjust.
7995 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7996 (nt_init_block): Convert to method
7997 notrappping_dom_walker::before_dom_children.
7998 (nt_fini_block): Convert to method
7999 method nontrapping_dom_walker::after_dom_children.
8000 (get_non_trapping): Adjust.
8001 * tree-ssa-pre.c (eliminate_dom_walker): New class.
8002 (eliminate_bb): Convert to method
8003 eliminate_dom_walker::before_dom_children.
8004 (eliminate_leave_block): Convert to method
8005 eliminate_dom_walker::after_dom_children.
8006 (eliminate): Adjust.
8007 * tree-ssa-strlen.c (strlen_dom_walker): New class.
8008 (strlen_enter_block): Convert to method
8009 strlen_dom_walker::before_dom_children.
8010 (strlen_leave_block): Convert to method
8011 method strlen_dom_walker::after_dom_children.
8012 (tree_ssa_strlen): Adjust.
8013 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
8014 (tree_ssa_uncprop): Adjust.
8015 (uncprop_leave_block): Convert to method
8016 uncprop_dom_walker::after_dom_children.
8017 (uncprop_leave_block): Convert to method
8018 uncprop_dom_walker::before_dom_children.
8020 2013-09-18 Bin Cheng <bin.cheng@arm.com>
8022 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
8023 branch in same basic block. Check both src and dest of the move insn.
8025 2013-09-17 Nick Clifton <nickc@redhat.com>
8027 * config/rl78/rl78-real.md (bf): New pattern.
8029 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
8030 (rl78_print_operand): Do not put a # before a %B.
8031 * config/rl78/rl78.opt: Tweak doc strings.
8033 2013-09-17 DJ Delorie <dj@redhat.com>
8035 * config/rl78/constraints.md (Wcv): Allow up to $r31.
8036 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
8037 (rl78_option_override): Likewise, if -mallregs.
8038 (is_virtual_register): Likewise.
8039 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
8040 (REGNO_OK_FOR_BASE_P): Likewise.
8041 * config/rl78/rl78.opt (-mallregs): New.
8043 2013-09-17 Nick Clifton <nickc@redhat.com>
8045 * config/rl78/rl78.c (need_to_save): Change return type to bool.
8046 For interrupt functions: save all call clobbered registers if the
8047 interrupt handler is not a leaf function.
8048 (rl78_expand_prologue): Always recompute the frame information.
8049 For interrupt functions: only select bank 0 if one of the bank 0
8050 registers is going to be psuhed.
8052 2013-09-17 DJ Delorie <dj@redhat.com>
8054 * config/rl78/constraints.md: For each W* constraint, rename to C*
8055 and create a W* constraint that checks for an optional ES: prefix
8057 * config/rl78/rl78.md (UNS_ES_ADDR): New.
8058 (es_addr): New. Used to wrap far addresses.
8059 * config/rl78/rl78-protos.h (rl78_es_addr): New.
8060 (rl78_es_base): New.
8061 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
8062 wrapped far addresses.
8063 (rl78_print_operand_1): Unwrap far addresses before processing.
8064 (rl78_lo16): Wrap far addresses in unspecs.
8065 (rl78_es_addr): New.
8066 (rl78_es_base): New.
8067 (insn_ok_now): Check for not-yet-wrapped far addresses.
8068 (transcode_memory_rtx): Properly re-wrap far addresses.
8070 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
8072 * config/sparc/t-rtems: Add leon3 multilibs.
8074 2013-09-17 Cong Hou <congh@google.com>
8076 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
8077 when checking the dot production pattern. The type of rhs operand
8078 of multiply is now checked correctly.
8080 2013-09-17 Jeff Law <law@redhat.com>
8082 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8083 edge implied equivalences into successor phis.
8084 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
8085 here from tree-ssa-threadedge.c.
8086 (mark_threaded_blocks): When threading through a joiner, if both
8087 successors of the joiner's clone reach the same block, verify the
8088 PHI arguments are equal. If not, cancel the jump threading request.
8089 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
8090 tree-ssa-threadupdate.c
8091 (thread_across_edge): Don't check PHI argument equality when
8092 threading through joiner block here.
8094 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
8096 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
8097 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
8098 (warn_uninit, warn_uninitialized_vars,
8099 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8100 Move to tree-ssa-uninit.c.
8101 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
8102 (has_undefined_value_p): New. Pass dependant parts of
8103 ssa_undefined_value_p.
8104 (uninit_undefined_value_p): Use has_undefined_value_p.
8105 (warn_uninit, warn_uninitialized_vars,
8106 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8107 Move from tree-ssa.c.
8108 * tree-ssa.h: Adjust prototypes.
8110 2013-09-17 Jan Hubicka <jh@suse.cz>
8113 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
8114 * ipa-inline.c (can_inline_edge_p): Do not downgrade
8115 FUNCTION_NOT_OPTIMIZED.
8116 * ipa-inline-analysis.c (compute_inline_parameters): Function
8117 not optimized is not inlinable unless it is alwaysinline.
8118 (inline_analyze_function): Force calls in not optimized
8119 function not inlinable.
8121 2013-09-17 Jan Hubicka <jh@suse.cz>
8124 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
8126 * ipa.c (function_and_variable_visibility): Likewise.
8127 * ipa-profile.c (ipa_profile): Likewise.
8129 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
8132 * cgraph.c (cgraph_function_body_availability): Check for ifunc
8133 attribute, and don't inline the resolver in this case.
8135 2013-09-17 Teresa Johnson <tejohnson@google.com>
8137 * coverage.c (get_coverage_counts): Add missing newline.
8139 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8141 PR tree-optimization/58088
8142 * fold-const.c (mask_with_trailing_zeros): New function.
8143 (fold_binary_loc): Make sure we don't recurse infinitely
8144 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
8145 Use mask_with_trailing_zeros where appropriate.
8147 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
8149 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
8150 of 'prev' var to get better distance estimation.
8152 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
8154 * tree-inline.h (struct copy_body_data): Add transform_parameter.
8155 * tree-inline.c (is_parameter_of): New predicate.
8156 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
8157 a parameter has been remapped.
8158 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
8159 (optimize_inline_calls): Initialize transform_parameter.
8160 (copy_gimple_seq_and_replace_locals): Likewise.
8161 (tree_function_versioning): Likewise.
8162 (maybe_inline_call_in_expr): Likewise.
8164 2013-09-17 Nick Clifton <nickc@redhat.com>
8166 * config/msp430/msp430-protos.h: Add prototypes for new functions.
8167 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
8169 (is_attr_func): New function.
8170 (msp430_is_interrupt_func): New function.
8171 (is_naked_func): New function.
8172 (is_reentrant_func): New function.
8173 (is_critical_func): New function.
8174 (msp430_start_function): Add annotations for function attributes.
8175 (msp430_attr): New function.
8176 (msp430_attribute_table): New.
8177 (msp430_function_section): New function.
8178 (TARGET_ASM_FUNCTION_SECTION): Define.
8179 (msp430_builtin): New enum.
8180 (msp430_init_builtins): New function.
8181 (msp430_builtin_devl): New function.
8182 (msp430_expand_builtin): New function.
8183 (TARGET_INIT_BUILTINS): Define.
8184 (TARGET_EXPAND_BUILTINS): Define.
8185 (TARGET_BUILTIN_DECL): Define.
8186 (msp430_expand_prologue): Add support for naked, interrupt,
8187 critical and reentrant functions.
8188 (msp430_expand_epilogue): Likewise.
8189 (msp430_print_operand): Handle 'O' character.
8190 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
8192 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
8193 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
8194 (pushm): Use a 'n' rather than an 'i' constraint.
8195 (msp_return): Add generation of the interrupt return instruction.
8196 (disable_interrupts): New pattern.
8197 (enable_interrupts): New pattern.
8198 (push_intr_state): New pattern.
8199 (pop_intr_state): New pattern.
8200 (bic_SR): New pattern.
8201 (bis_SR): New pattern.
8202 * doc/extend.texi: Document MSP430 function attributes and builtin
8205 2013-09-17 Richard Biener <rguenther@suse.de>
8207 PR tree-optimization/58432
8208 * tree-loop-distribution.c (tree_loop_distribution): Also
8209 scan PHIs for outside loop uses and seed a partition from them.
8211 2013-09-17 Bin Cheng <bin.cheng@arm.com>
8213 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
8214 (restructure_reference): Call backtrace_base_for_ref.
8216 2013-09-17 Alan Modra <amodra@gmail.com>
8219 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
8222 2013-09-16 DJ Delorie <dj@redhat.com>
8224 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
8225 vregs location for RL78/G10.
8226 (rl78_expand_prologue): Avoid SEL on G10.
8227 (rl78_expand_epilogue): Likewise.
8228 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
8229 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
8230 __RL78_G10__ when appropriate.
8231 (ASM_SPEC): Pass -mg10 along to the assembler.
8232 * config/rl78/rl78.md (sel_rb): Disable for G10.
8233 * config/rl78/rl78.opt: Add -mg10 option.
8234 * config/rl78/t-rl78: Add -mg10 multilib.
8236 2013-09-16 Xinliang David Li <davidxl@google.com>
8238 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
8239 * omp-low.c (omp_max_vf): Ditto.
8240 (expand_omp_simd): Ditto.
8241 * tree-vectorizer.c (vectorize_loops): Ditto.
8242 (gate_vect_slp): Ditto.
8243 (gate_increase_alignment): Ditto.
8244 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
8245 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
8246 (gate_tree_vectorize): Name change.
8247 (tree_vectorize): Ditto.
8248 (pass_vectorize::gate): Call new function.
8249 (pass_vectorize::execute): Ditto.
8250 * opts.c: O3 default setting change.
8251 (finish_options): Check new flag.
8252 * doc/invoke.texi: Document new flags.
8253 * common.opt: New flags.
8255 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
8257 * doc/tm.texi.in (Cond Exec Macros): Remove node.
8258 (Condition Code): Don't reference it.
8259 * doc/tm.texi: Regenerate.
8261 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
8264 * lra-constraints.c (undo_optional_reloads): Consider all optional
8265 reload even if it did not get a hard reg.
8267 2013-09-16 Teresa Johnson <tejohnson@google.com>
8269 * dumpfile.c (dump_loc): Remove newline emission.
8270 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
8271 emission to dump_printf_loc calls where missing.
8272 (vect_mark_for_runtime_alias_test): Ditto.
8273 (vect_analyze_data_ref_dependence): Ditto.
8274 (vect_analyze_data_ref_dependences): Ditto.
8275 (vect_slp_analyze_data_ref_dependence): Ditto.
8276 (vect_slp_analyze_data_ref_dependences): Ditto.
8277 (vect_compute_data_ref_alignment): Ditto.
8278 (vect_update_misalignment_for_peel): Ditto.
8279 (vect_verify_datarefs_alignment): Ditto.
8280 (vector_alignment_reachable_p): Ditto.
8281 (vect_get_data_access_cost): Ditto.
8282 (vect_enhance_data_refs_alignment): Ditto.
8283 (vect_find_same_alignment_drs): Ditto.
8284 (vect_analyze_data_refs_alignment): Ditto.
8285 (vect_analyze_group_access): Ditto.
8286 (vect_analyze_data_ref_access): Ditto.
8287 (vect_analyze_data_ref_accesses): Ditto.
8288 (vect_prune_runtime_alias_test_list): Ditto.
8289 (vect_analyze_data_refs): Ditto.
8290 (vect_create_addr_base_for_vector_ref): Ditto.
8291 (vect_create_data_ref_ptr): Ditto.
8292 (vect_grouped_store_supported): Ditto.
8293 (vect_grouped_load_supported): Ditto.
8294 * value-prof.c (check_counter): Ditto.
8295 (check_ic_target): Ditto.
8296 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
8297 (vect_recog_widen_mult_pattern): Ditto.
8298 (vect_recog_widen_sum_pattern): Ditto.
8299 (vect_recog_over_widening_pattern): Ditto.
8300 (vect_recog_widen_shift_pattern): Ditto.
8301 (vect_recog_rotate_pattern): Ditto.
8302 (vect_recog_vector_vector_shift_pattern): Ditto.
8303 (vect_recog_divmod_pattern): Ditto.
8304 (vect_recog_mixed_size_cond_pattern): Ditto.
8305 (vect_recog_bool_pattern): Ditto.
8306 (vect_pattern_recog_1): Ditto.
8307 (vect_pattern_recog): Ditto.
8308 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
8309 (vect_is_simple_iv_evolution): Ditto.
8310 (vect_analyze_scalar_cycles_1): Ditto.
8311 (vect_get_loop_niters): Ditto.
8312 (vect_analyze_loop_1): Ditto.
8313 (vect_analyze_loop_form): Ditto.
8314 (vect_analyze_loop_operations): Ditto.
8315 (vect_analyze_loop_2): Ditto.
8316 (vect_analyze_loop): Ditto.
8317 (report_vect_op): Ditto.
8318 (vect_is_slp_reduction): Ditto.
8319 (vect_is_simple_reduction_1): Ditto.
8320 (vect_get_known_peeling_cost): Ditto.
8321 (vect_estimate_min_profitable_iters): Ditto.
8322 (vect_model_reduction_cost): Ditto.
8323 (vect_model_induction_cost): Ditto.
8324 (get_initial_def_for_induction): Ditto.
8325 (vect_create_epilog_for_reduction): Ditto.
8326 (vectorizable_reduction): Ditto.
8327 (vectorizable_induction): Ditto.
8328 (vectorizable_live_operation): Ditto.
8329 (vect_loop_kill_debug_uses): Ditto.
8330 (vect_transform_loop): Ditto.
8331 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
8332 (vect_stmt_relevant_p): Ditto.
8333 (process_use): Ditto.
8334 (vect_mark_stmts_to_be_vectorized): Ditto.
8335 (vect_model_simple_cost): Ditto.
8336 (vect_model_promotion_demotion_cost): Ditto.
8337 (vect_model_store_cost): Ditto.
8338 (vect_get_store_cost): Ditto.
8339 (vect_model_load_cost): Ditto.
8340 (vect_get_load_cost): Ditto.
8341 (vect_init_vector_1): Ditto.
8342 (vect_get_vec_def_for_operand): Ditto.
8343 (vect_finish_stmt_generation): Ditto.
8344 (vectorizable_call): Ditto.
8345 (vectorizable_conversion): Ditto.
8346 (vectorizable_assignment): Ditto.
8347 (vectorizable_shift): Ditto.
8348 (vectorizable_operation): Ditto.
8349 (vectorizable_store): Ditto.
8350 (vectorizable_load): Ditto.
8351 (vectorizable_condition): Ditto.
8352 (vect_analyze_stmt): Ditto.
8353 (vect_transform_stmt): Ditto.
8354 (vect_is_simple_use): Ditto.
8355 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
8356 (vect_can_advance_ivs_p): Ditto.
8357 (vect_update_ivs_after_vectorizer): Ditto.
8358 (vect_do_peeling_for_loop_bound): Ditto.
8359 (vect_gen_niters_for_prolog_loop): Ditto.
8360 (vect_update_inits_of_drs): Ditto.
8361 (vect_create_cond_for_alias_checks): Ditto.
8362 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
8363 (vect_build_slp_tree_1): Ditto.
8364 (vect_supported_load_permutation_p): Ditto.
8365 (vect_analyze_slp_instance): Ditto.
8366 (vect_analyze_slp): Ditto.
8367 (vect_make_slp_decision): Ditto.
8368 (vect_detect_hybrid_slp): Ditto.
8369 (vect_bb_vectorization_profitable_p): Ditto.
8370 (vect_slp_analyze_bb_1): Ditto.
8371 (vect_update_slp_costs_according_to_vf): Ditto.
8372 (vect_get_mask_element): Ditto.
8373 (vect_transform_slp_perm_load): Ditto.
8374 (vect_schedule_slp_instance): Ditto.
8375 (vect_schedule_slp): Ditto.
8376 (vect_slp_transform_bb): Ditto.
8377 * profile.c (read_profile_edge_counts): Ditto.
8378 (compute_branch_probabilities): Ditto.
8379 * coverage.c (get_coverage_counts): Ditto.
8381 2013-09-16 Diego Novillo <dnovillo@google.com>
8383 * tree-core.h: Add missing comment lines from refactoring of tree.h.
8385 2013-09-16 Jan Hubicka <jh@suse.cz>
8387 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
8388 abstract functions; for static functions check the presence of body.
8390 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8392 * config/aarch64/aarch64-simd-builtins.def (fma): New.
8393 * config/aarch64/aarch64-simd.md
8394 (aarch64_mla_elt<mode>): New.
8395 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8396 (aarch64_mls_elt<mode>): Likewise.
8397 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8398 (aarch64_fma4_elt<mode>): Likewise.
8399 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8400 (aarch64_fma4_elt_to_128v2df): Likewise.
8401 (aarch64_fma4_elt_to_64df): Likewise.
8402 (fnma<mode>4): Likewise.
8403 (aarch64_fnma4_elt<mode>): Likewise.
8404 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8405 (aarch64_fnma4_elt_to_128v2df): Likewise.
8406 (aarch64_fnma4_elt_to_64df): Likewise.
8407 * config/aarch64/iterators.md (VDQSF): New.
8408 * config/aarch64/arm_neon.h
8409 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
8410 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8412 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8414 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8415 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8416 (aarch64_mul3_elt_to_128df): Likewise.
8417 (aarch64_mul3_elt_to_64v2df): Likewise.
8418 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8420 (VMUL_CHANGE_NLANES) Likewise.
8423 * config/aarch64/arm_neon.h
8424 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8426 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8428 * config/aarch64/arm_neon.h
8429 (vcvtx_high_f32_f64): Fix parameters.
8431 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8432 Uros Bizjak <ubizjak@gmail.com>
8434 * config/alpha.c: Include tree-ssanames.h.
8436 2013-09-16 Richard Biener <rguenther@suse.de>
8438 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8439 (dot_rdg_1): Handle control_dd.
8440 (create_edge_for_control_dependence): New function.
8441 (create_rdg_edges): Add control dependences if asked for.
8442 (build_rdg): Likewise.
8443 (generate_loops_for_partition): If there are not necessary
8444 control stmts remove all their dependencies.
8445 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8446 (distribute_loop): Pass on control dependences.
8447 (tree_loop_distribution): Compute control dependences and remove
8448 restriction on number of loop nodes.
8450 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8452 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8455 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8457 * cse.c (try_const_anchors): Punt on CC modes.
8459 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8461 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8463 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8466 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8469 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8472 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8474 (darwin_objc1_section): Likewise.
8475 (darwin_file_end): Emit Image Info section when required.
8477 2013-09-14 Jan Hubicka <jh@suse.cz>
8479 * tree-into-ssa.c (gate_into_ssa): New.
8480 (pass_data_build_ssa): Use it.
8481 * cgraph.h (expand_thunk): Update prototype.
8482 * cgraphunit.c (analyze_function): Expand thunks early.
8483 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8484 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8485 set lowered flag; do not add new function.
8486 (assemble_thunks_and_aliases): Update.
8487 * tree-ssa.c (gate_init_datastructures): New gate.
8488 (pass_data_init_datastructures): Use it.
8490 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8493 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8494 xmm register set described in the psABI.
8496 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8498 * dwarf2out.c (should_emit_struct_debug): Add check
8499 for type_decl variable is not NULL.
8501 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8503 * config.gcc: Use new winnt-c.c target hooks
8504 * config/t-winnt: New file
8505 * config/winnt-c.c: New file
8506 * doc/tm.texi.in: Document new hook
8507 * doc/tm.texi: Regenerated
8509 2013-09-13 Jan Hubicka <jh@suse.cz>
8512 * ipa-inline.c (check_callers): New function.
8513 (check_caller_edge): Remove.
8514 (want_inline_function_to_all_callers_p): Also permit alises that are
8516 (inline_to_all_callers): Terminate the walk when devirtualization
8517 introduce new calls.
8519 2013-09-13 Jan Hubicka <jh@suse.cz>
8521 * ipa-inline-analysis.c (struct growth_data): Add node.
8522 (do_estimate_growth_1): Fix detection of recursion.
8524 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8526 PR tree-optimization/58392
8527 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8528 to avoid shadowing of outer loop variable. If
8529 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8530 replace_by_duplicate_decl simduid of loops that have it set and
8531 set dest_cfun->has_simduid_loops and/or
8532 dest_cfun->has_force_vect_loops.
8533 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8534 instead of maybe_lookup_decl.
8535 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8536 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8537 call. Copy over force_vect and copy and remap simduid. Set
8538 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8539 (copy_cfg_body): Remove blocks_to_copy argument. Use
8540 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8541 caller. Don't set cfun->has_simduid_loops and/or
8542 cfun->has_force_vect_loops here.
8543 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8545 (expand_call_inline, tree_function_versioning): Adjust copy_body
8548 2013-09-13 Martin Jambor <mjambor@suse.cz>
8551 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8552 the assert if the edge was a speculative one.
8554 2013-09-13 Richard Biener <rguenther@suse.de>
8556 * tree-data-ref.h (known_dependences_p): Move here ...
8557 * tree-loop-distribution.c (known_dependences_p): ... from here.
8558 (dump_rdg_component, debug_rdg_component): Remove.
8560 (generate_loops_for_partition): Use gimple_uid instead of
8561 relying on matching stmt visit order.
8562 (rdg_build_partitions): Take starting stmt vector.
8563 (ldist_gen): Merge into ...
8564 (distribute_loop): ... this function. Do not compute starting
8566 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
8568 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8570 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
8571 Set type attribute correctly. Set predicable_short_it attribute.
8572 (cmpsi_shiftsi): Remove %? from output template.
8574 2013-09-13 Richard Biener <rguenther@suse.de>
8576 * tree-loop-distribution.c (struct rdg_component,
8577 rdg_defs_used_in_other_loops_p, free_rdg_components,
8578 rdg_build_components): Remove.
8579 (stmts_from_loop): Do not record virtual PHIs.
8580 (generate_loops_for_partition): Skip virtual PHIs.
8581 (build_rdg_partition_for_component): Rename to ...
8582 (build_rdg_partition_for_vertex): ... this and adjust.
8583 (rdg_build_partitions): Take a vector of starting vertices
8584 instead of components. Remove unnecessary leftover handling.
8585 (ldist_gen): Do not build components or record other stores.
8586 (distribute_loop): Do not distribute loops containing stmts
8589 2013-09-13 Christian Bruel <christian.bruel@st.com>
8592 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
8594 2013-09-13 Kai Tietz <ktietz@redhat.com>
8596 * config.gcc: Separate cases for mingw and cygwin targets,
8597 and add 64-bit cygwin target case.
8599 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
8600 dll-export inline-functions.
8601 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
8603 2013-09-13 Jeff Law <law@redhat.com>
8607 2013-09-06 Jeff Law <law@redhat.com>
8609 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8610 edge implied equivalences into successor phis.
8612 2013-09-12 DJ Delorie <dj@redhat.com>
8614 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8617 2013-09-12 Brooks Moses <bmoses@google.com>
8620 * Makefile.in: Do not install driver binaries in $(target)/bin.
8622 2013-09-12 DJ Delorie <dj@redhat.com>
8624 * config/rl78/rl78.opt (mrelax): New.
8625 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8626 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8628 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8630 (rl78_expand_epilogue): Likewise.
8632 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8635 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8636 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8637 pointer elimination with using remove_reg_equal_offset_note.
8639 2013-09-12 DJ Delorie <dj@redhat.com>
8641 * config/msp430/: New port.
8642 * config.gcc (msp430): Added.
8643 * doc/invoke.texi: Document MSP430 options.
8644 * doc/install.texi: Document msp430-elf
8645 * doc/md.texi: Document msp430-elf
8646 * doc/contrib.texi: Document msp430-elf
8648 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8651 2013-09-12 Martin Jambor <mjambor@suse.cz>
8654 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8655 reference descriptor is NULL.
8656 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8659 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8661 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8662 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8663 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8664 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8665 make_temp_ssa_name): move to tree-ssanames.h
8666 * tree-ssa-alias.h: Move prototype.
8667 * tree-ssa.h: Include tree-ssanames.h.
8668 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8669 * tree-ssanames.h: New. Move items from tree-flow*.h
8670 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8672 2013-09-12 Richard Biener <rguenther@suse.de>
8674 PR tree-optimization/58404
8675 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8676 propagate non-invariant addresses into dereferences wrapped
8677 in component references.
8679 2013-09-12 Richard Biener <rguenther@suse.de>
8681 PR tree-optimization/58402
8682 * passes.def: Move pass_late_warn_uninitialized later.
8684 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8686 * tree-ssa.h: New. Move content from tree-flow.h and
8688 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8689 Move prototypes belonging to tree-ssa.c.
8690 * tree-flow-inline.h (redirect_edge_var_map_def,
8691 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8693 * gimple.h: Adjust prototypes.
8694 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8696 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8697 * tree.h: Move prototype to tree-ssa.h.
8698 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8699 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8700 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8701 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8702 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8703 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8704 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8705 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8706 graphite-blocking.c, graphite-clast-to-gimple.c,
8707 graphite-dependences.c, graphite-interchange.c,
8708 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8709 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8710 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8711 ipa-reference.c, ipa-split.c, ipa-utils.c,
8712 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8713 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8714 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8715 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8716 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8717 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8718 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8719 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8720 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8721 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8722 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8723 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8724 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8725 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8726 value-prof.c, var-tracking.c,
8727 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8729 2013-09-12 Richard Biener <rguenther@suse.de>
8731 PR tree-optimization/58396
8732 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8733 (build_rdg): Take a loop-nest parameter, fix memleaks.
8734 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8736 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8738 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8741 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8743 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8744 The FMA instruction names should have a 'v' prefix.
8746 2013-09-12 Richard Biener <rguenther@suse.de>
8748 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8749 (dot_rdg): Use popen instead of system in optional code.
8750 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8751 (already_processed_vertex_p): Adjust.
8752 (has_anti_or_output_dependence, predecessor_has_mem_write,
8753 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8754 rdg_flag_uses): Remove.
8755 (rdg_flag_vertex): Simplify.
8756 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8758 (build_rdg_partition_for_component): Process the first vertex
8759 of a component only.
8760 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8762 2013-09-12 Alan Modra <amodra@gmail.com>
8764 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8766 2013-09-11 DJ Delorie <dj@redhat.com>
8767 Nick Clifton <nickc@redhat.com>
8769 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8770 (rl78_stack_based_mem): New.
8771 * config/rl78/constraints.md (Iv08): New.
8780 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8782 (movsi): Change from expand to insn-and-split.
8783 (ashrsi3): Clobber AX.
8787 * config/rl78/rl78.md (CC_REG): Fix.
8788 (addsi3): Allow memory and immediate operands.
8789 (addsi3_internal): Split into...
8790 (addsi3_internal_virt): ...new, and ...
8791 (addsi3_internal_real): ...new.
8793 (subsi3_internal_virt): New.
8794 (subsi3_internal_real): New.
8795 (mulsi3): Add memory operand.
8796 (mulsi3_rl78): Likewise.
8797 (mulsi3_g13): Likewise.
8798 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8799 (cbranchqi4_real): Add more constraint options.
8800 (cbranchhi4_real): Expand pattern.
8801 (cbranchhi4_real_signed): New.
8802 (cbranchhi4_real_inverted): New.
8803 (cbranchsi4_real_lt): New.
8804 (cbranchsi4_real_ge): New.
8805 (cbranchsi4_real_signed): New.
8806 (cbranchsi4_real): New.
8808 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8810 (lshrsi3_virt): Likewise.
8811 (ashlsi3_virt): Likewise.
8812 (cbranchqi4_virt_signed): New.
8813 (cbranchhi4_virt_signed): New.
8814 (cbranchsi4_virt): New.
8815 * config/rl78/rl78.c: Whitespace fixes throughout.
8816 (move_elim_pass): New.
8817 (pass_data_rl78_move_elim): New.
8818 (pass_rl78_move_elim): New.
8819 (make_pass_rl78_move_elim): New.
8820 (rl78_devirt_info): Run devirt earlier.
8821 (rl78_move_elim_info): New.
8822 (rl78_asm_file_start): Register it.
8823 (rl78_split_movsi): New.
8824 (rl78_as_legitimate_address): Allow virtual base registers when
8826 (rl78_addr_space_convert): Remove spurious debug stuff.
8827 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8828 (rl78_print_operand): More cases for not printing '#'.
8829 (rl78_expand_compare): Remove most of the logic.
8830 (content_memory): New.
8831 (clear_content_memory): New.
8832 (get_content_index): New.
8833 (get_content_name): New.
8834 (display_content_memory): New.
8835 (update_content): New.
8836 (record_content): New.
8837 (already_contains): New.
8838 (insn_ok_now): Re-recog insns with virtual registers.
8839 (add_postponed_content_update): New.
8840 (process_postponed_content_update): New.
8841 (gen_and_emit_move): New.
8842 (transcode_memory_rtx): Record new location content.
8843 Use gen_and_emit_move.
8844 (force_into_acc): New.
8845 (move_to_acc): Use gen_and_emit_move.
8846 (move_from_acc): Likewise.
8847 (move_acc_to_reg): Likewise.
8848 (move_to_x): Likewise.
8849 (move_to_hl): Likewise.
8850 (move_to_de): Likewise.
8851 (rl78_alloc_physical_registers_op1): Record location content.
8852 (has_constraint): New.
8853 (rl78_alloc_physical_registers_op2): Record location content.
8855 (rl78_alloc_physical_registers_ro1): Likewise.
8856 (rl78_alloc_physical_registers_cmp): Likewise.
8857 (rl78_alloc_physical_registers_umul): Likewise.
8858 (rl78_alloc_address_registers_macax): New.
8859 (rl78_alloc_physical_registers): Initialize and set location
8860 content memory as needed.
8861 (rl78_reorg): Make sure split2 is called.
8862 (rl78_rtx_costs): New.
8864 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8866 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8867 for (not (neg ...)) and (neg (not ...)) cases.
8869 2013-09-11 Richard Biener <rguenther@suse.de>
8872 * passes.def: Split critical edges before late uninit warning passes.
8873 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8875 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8877 PR tree-optimization/58385
8878 * fold-const.c (build_range_check): If both low and high are NULL,
8879 use omit_one_operand_loc to preserve exp side-effects.
8881 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8883 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8885 2013-09-11 Richard Biener <rguenther@suse.de>
8887 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8888 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8889 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8890 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8891 stmts_from_loop, known_dependences_p, build_empty_rdg,
8892 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8893 * tree-loop-distribution.c: ... here.
8894 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8895 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8896 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8897 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8898 * tree-loop-distribution.c: ... here.
8899 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8900 (struct partition_s): Add loops member.
8901 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8902 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8903 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8905 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8906 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8907 Sergey Lega <sergey.s.lega@intel.com>
8908 Anna Tikhonova <anna.tikhonova@intel.com>
8909 Ilya Tocar <ilya.tocar@intel.com>
8910 Andrey Turetskiy <andrey.turetskiy@intel.com>
8911 Ilya Verbin <ilya.verbin@intel.com>
8912 Kirill Yukhin <kirill.yukhin@intel.com>
8913 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8915 * config/i386/constraints.md (k): New.
8917 * config/i386/i386.c (const regclass_map): Add new mask registers.
8918 (dbx_register_map): Ditto.
8919 (dbx64_register_map): Ditto.
8920 (svr4_dbx_register_map): Ditto.
8921 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8923 (ix86_preferred_reload_class): Disable constants for mask registers.
8924 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8925 (ix86_hard_regno_mode_ok): Support new mask registers.
8926 (x86_order_regs_for_local_alloc): Ditto.
8927 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8928 (FIXED_REGISTERS): Add new mask registers.
8929 (CALL_USED_REGISTERS): Ditto.
8930 (REG_ALLOC_ORDER): Ditto.
8931 (VALID_MASK_REG_MODE): New.
8932 (FIRST_MASK_REG): Ditto.
8933 (LAST_MASK_REG): Ditto.
8934 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8935 (MAYBE_MASK_CLASS_P): New.
8936 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8937 (REG_CLASS_CONTENTS): Ditto.
8938 (MASK_REGNO_P): New.
8939 (ANY_MASK_REG_P): Ditto.
8940 (HI_REGISTER_NAMES): Add new mask registers.
8941 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8942 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8944 (attribute "type"): Add mskmov, msklog.
8945 (attribute "length_immediate"): Support them.
8946 (attribute "memory"): Ditto.
8947 (attribute "prefix_0f"): Ditto.
8948 (*movhi_internal): Support new mask registers.
8949 (*movqi_internal): Ditto.
8950 (define_split): Split out clobber pattern is a logic
8951 insn on mask registers.
8952 (*k<logic><mode>): New.
8953 (*andhi_1): Extend to support mask regs.
8954 (*andqi_1): Extend to support mask regs.
8956 (define_split): Split and-not to and and not if operands
8958 (*<code><mode>_1): Separate HI mode to new pattern...
8959 (*<code>hi_1): This.
8960 (*<code>qi_1): Extend to support mask regs.
8962 (kortestzhi): Ditto.
8963 (kortestchi): Ditto.
8965 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8966 (*one_cmplhi2_1): ...Here, now with mask registers support.
8967 (*one_cmplqi2_1): Support new mask registers.
8968 (HI/QImode arithmetics splitter): Don't split if mask registers
8970 (HI/QImode not splitter): Ditto.
8971 * config/i386/predicated.md (mask_reg_operand): New.
8972 (general_reg_operand): Ditto.
8974 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8976 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8977 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8979 2013-09-10 Jeff Law <law@redhat.com>
8981 PR tree-optimization/58380
8982 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8983 of threading through a buried loop header.
8985 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8986 return value for single successor case.
8988 2013-09-10 Jan Hubicka <jh@suse.cz>
8990 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8992 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8995 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8996 support conditional execution.
8997 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8999 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
9001 * lra.c (lra): Clear lra_optional_reload_pseudos before every
9003 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
9004 Check destination too to check move insn.
9005 (undo_optional_reloads): Add check that the original peudo did not
9006 changed its allocation and the optional reload was inherited on last
9007 inheritance pass. Break loop after deciding to keep optional reload.
9008 (lra_undo_inheritance): Add check that inherited pseudo still in
9011 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
9013 * config/aarch64/aarch64.md (generic_sched): New.
9014 * config/aarch64/aarch64-generic.md (load): Make conditional
9015 on generic_sched attribute.
9016 (nonload): Likewise.
9018 2013-09-10 Jan Hubicka <jh@suse.cz>
9020 * lto-cgraph.c: Include ipa-utils.h.
9021 (compute_ltrans_boundary): Also add possible targets into the boundary.
9023 2013-09-10 Jan Hubicka <jh@suse.cz>
9025 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
9026 VAR_DECL of vtable rather than full expression.
9028 2013-09-10 Jan Hubicka <jh@suse.cz>
9029 Paolo Carlini <paolo.carlini@oracle.com>
9031 * cgraphunit.c (analyze_functions): Save input_location, set it
9032 to UNKNOWN_LOCATION and restore it at the end.
9034 2013-09-10 Martin Jambor <mjambor@suse.cz>
9036 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
9037 represented by a thunk.
9039 2013-09-10 Jeff Law <law@redhat.com>
9041 PR tree-optimization/58343
9042 * tree-ssa-threadupdate.c (thread_block): Identify and disable
9043 jump threading requests through loop headers buried in the middle
9044 of a jump threading path.
9046 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
9047 in return value/type.
9049 2013-09-10 Jakub Jelinek <jakub@redhat.com>
9051 PR rtl-optimization/58365
9052 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
9053 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
9056 2013-09-10 Richard Biener <rguenther@suse.de>
9058 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
9059 * tree-data-ref.c (create_rdg_vertices): Collect all data
9060 references, signal failure to the caller, use data-ref API.
9061 (build_rdg): Compute data references only once. Maintain lifetime
9062 of data references and data dependences from within RDG.
9063 (free_rdg): Free dependence relations.
9064 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
9065 inventing extra dependences.
9066 (distribute_loop): Update for RDG API changes.
9068 2013-09-10 Kai Tietz <ktietz@redhat.com>
9070 * doc/invoke.texi (fms-extensions): Document changed
9071 behavior for ms-abi targets.
9072 * config/i386/i386.c (ix86_option_override_internal):
9073 Set default value of option -fms-extension for ms-abi targets.
9075 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9077 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
9079 2013-09-10 Alan Modra <amodra@gmail.com>
9082 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
9084 2013-09-10 Alan Modra <amodra@gmail.com>
9086 * config/rs6000/predicates.md (add_cint_operand): New.
9087 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
9088 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
9089 using add_cint_operand.
9090 (largetoc_high_plus_aix): Likewise.
9092 2013-09-09 Jakub Jelinek <jakub@redhat.com>
9094 PR tree-optimization/58364
9095 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
9096 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
9097 the current range can't be an unconditional true or false.
9099 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
9101 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
9103 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
9105 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
9107 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
9110 * doc/invoke.texi (-Wdelete-incomplete): Document it.
9112 2013-09-09 Ian Bolton <ian.bolton@arm.com>
9114 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
9115 NO_REGS for immediate that can't be moved directly into FP_REGS.
9117 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9119 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
9120 comparison with negated operand.
9121 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
9124 2013-09-09 Richard Biener <rguenther@suse.de>
9127 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
9128 of a subloop record all its block as affecting loop-closed SSA form.
9130 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9132 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
9133 of an rtx/bitpos pair.
9134 (store_fixed_bit_field): Update accordingly.
9136 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9138 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
9140 * builtins.c (expand_errno_check): Likewise.
9141 * dwarf2cfi.c (init_return_column_size): Likewise.
9142 * except.c (sjlj_mark_call_sites): Likewise.
9143 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
9144 * lra-constraints.c (emit_inc): Likewise.
9145 * ree.c (combine_set_extension): Likewise.
9146 * regmove.c (fixup_match_2): Likewise.
9147 * reload1.c (inc_for_reload): Likewise.
9149 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9151 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
9152 (force_to_mode, simplify_shift_const_1, simplify_comparison):
9153 Use gen_int_mode with the mode of the associated simplify_* call.
9154 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
9155 * expmed.c (expand_shift_1): Likewise.
9156 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9157 * loop-iv.c (iv_number_of_iterations): Likewise.
9158 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9159 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
9161 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9163 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
9164 of the associated expand_* call.
9165 (asan_emit_stack_protection): Likewise.
9166 * builtins.c (round_trampoline_addr): Likewise.
9167 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
9168 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
9169 (emit_store_flag): Likewise.
9170 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
9171 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
9173 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9174 * ifcvt.c (noce_try_store_flag_constants): Likewise.
9175 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9176 * modulo-sched.c (generate_prolog_epilog): Likewise.
9177 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
9178 (expand_ctz, expand_ffs, expand_unop): Likewise.
9180 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9182 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
9183 of the associated gen_rtx_* call.
9184 * caller-save.c (init_caller_save): Likewise.
9185 * combine.c (find_split_point, make_extraction): Likewise.
9186 (make_compound_operation): Likewise.
9187 * dwarf2out.c (mem_loc_descriptor): Likewise.
9188 * explow.c (plus_constant, probe_stack_range): Likewise.
9189 * expmed.c (expand_mult_const): Likewise.
9190 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
9191 * reload1.c (init_reload): Likewise.
9192 * valtrack.c (cleanup_auto_inc_dec): Likewise.
9193 * var-tracking.c (adjust_mems): Likewise.
9194 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
9195 rather than gen_rtx_fmt_ee.
9197 2013-09-09 Jan Hubicka <jh@suse.cz>
9200 * value-prof.c (gimple_ic): Copy also abnormal edges.
9202 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9204 * asan.c (asan_shadow_cst): Use gen_int_mode.
9206 2013-09-08 Jan Hubicka <jh@suse.cz>
9208 * ipa-profile.c: Add toplevel comment.
9209 (ipa_propagate_frequency_1): Be more conservative when profile is read.
9210 (contains_hot_call_p): New function.
9211 (ipa_propagate_frequency): Set frequencies based on counts when
9213 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
9214 profile; do not tamper with profile after inlining if it is read.
9216 2013-09-08 Jan Hubicka <jh@suse.cz>
9218 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
9221 2013-09-08 Jan Hubicka <jh@suse.cz>
9223 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
9226 2013-09-08 Jeff Law <law@redhat.com>
9229 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
9232 2013-09-08 Andi Kleen <ak@linux.intel.com>
9234 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
9236 2013-09-08 Jan Hubicka <jh@suse.cz>
9238 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
9240 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
9242 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
9243 for non-debug insns.
9244 * lra.c (new_insn_reg): Take the containing insn as a parameter.
9245 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
9246 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
9249 2013-09-08 Jan Hubicka <jh@suse.cz>
9251 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
9252 targets and devirtualize to BUILT_IN_UNREACHABLE.
9253 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
9254 * ipa.c (walk_polymorphic_call_targets): New function.
9255 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
9256 functions; use the new timevar.
9257 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
9258 was removed from the program.
9259 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
9260 not consider it in the walk when its vtable is dead.
9261 (possible_polymorphic_call_targets_1): Pass anonymous flag to
9263 (devirt_variable_node_removal_hook): New function.
9264 (possible_polymorphic_call_targets): Also register
9265 devirt_variable_node_removal_hook.
9266 (ipa_devirt): Do not do non-speculative devirtualization.
9267 (gate_ipa_devirt): One execute if devirtualizing speculatively.
9269 2013-09-08 Jan Hubicka <jh@suse.cz>
9271 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
9272 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
9273 varpool_remove_variable_insertion_hook): Declare.
9274 * varpool.c (varpool_node_hook_list): New structure.
9275 (first_varpool_node_removal_hook,
9276 first_varpool_variable_insertion_hook): New variables.
9277 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
9278 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
9279 varpool_remove_variable_insertion_hook,
9280 varpool_call_variable_insertion_hooks): New functions.
9281 (varpool_remove_node): Use it.
9283 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
9286 * diagnostic.c (diagnostic_build_prefix): When s.file is
9287 "<built-in>" don't output line and column numbers.
9289 2013-09-06 Jan Hubicka <jh@suse.cz>
9291 * cgraphunit.c (expand_thunk): Get body before touching arguments.
9292 * lto-streamer-out.c: Stream thunks, too.
9293 * lto-streamer-in.c (input_function): Pop cfun here
9294 (lto_read_body): Instead of here.
9296 2013-09-06 Caroline Tice <cmtice@google.com>
9298 * doc/install.texi: Add documentation for the --enable-vtable-verify
9299 and the --disable-libvtv configure options.
9301 2013-09-06 Jeff Law <law@redhat.com>
9303 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9304 edge implied equivalences into successor phis.
9306 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
9308 * resource.c (mark_referenced_resources): Handle COND_EXEC.
9310 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
9312 * resource.c (mark_target_live_regs): Compute resources taking
9313 into account if a call is predicated or not.
9315 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9317 * toplev.c (output_stack_usage): Be prepared for suffixes created by
9318 the compiler in the function names.
9320 2013-09-06 Jan Hubicka <jh@suse.cz>
9323 * ipa-inline.c (has_caller_p): New function.
9324 (want_inline_function_to_all_callers_p): Use it.
9325 (sum_callers, inline_to_all_callers): Break out from ...
9326 (ipa_inline): ... here.
9328 2013-09-06 Jan Hubicka <jh@suse.cz>
9330 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
9331 only when AVX is enabled.
9333 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9335 * config/aarch64/aarch64.md
9336 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
9337 is fpsimd_<load/store>2.
9338 (load_pair<mode>): Likewise.
9339 (store_pair<mode>): Likewise.
9341 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9343 * config/arm/types.md (type): Add "mrs" type.
9344 * config/aarch64/aarch64.md
9345 (aarch64_load_tp_hard): Make type "mrs".
9347 (load_tp_hard): Make type "mrs".
9348 * config/arm/cortex-a15.md: Update with new attributes.
9349 * config/arm/cortex-a5.md: Update with new attributes.
9350 * config/arm/cortex-a53.md: Update with new attributes.
9351 * config/arm/cortex-a7.md: Update with new attributes.
9352 * config/arm/cortex-a8.md: Update with new attributes.
9353 * config/arm/cortex-a9.md: Update with new attributes.
9354 * config/arm/cortex-m4.md: Update with new attributes.
9355 * config/arm/cortex-r4.md: Update with new attributes.
9356 * config/arm/fa526.md: Update with new attributes.
9357 * config/arm/fa606te.md: Update with new attributes.
9358 * config/arm/fa626te.md: Update with new attributes.
9359 * config/arm/fa726te.md: Update with new attributes.
9361 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9363 * config/aarch64/aarch64.md
9364 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
9365 (*movtf_aarch64): Likewise.
9367 (thumb1_movdi_insn): Use "multiple" for type where more than one
9368 instruction is used for a move.
9369 (*arm32_movhf): Likewise.
9370 (*thumb_movdf_insn): Likewise.
9372 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9374 * config/arm/types.md (type): Rename fcpys to fmov.
9376 (*arm_movsi_vfp): Rename type fcpys as fmov.
9377 (*thumb2_movsi_vfp): Likewise
9378 (*movhf_vfp_neon): Likewise
9379 (*movhf_vfp): Likewise
9380 (*movsf_vfp): Likewise
9381 (*thumb2_movsf_vfp): Likewise
9382 (*movsfcc_vfp): Likewise
9383 (*thumb2_movsfcc_vfp): Likewise
9384 * config/aarch64/aarch64-simd.md
9385 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
9386 * config/aarch64/aarch64.md
9387 (*movsi_aarch64): Replace type mov_reg with fmovs.
9388 (*movdi_aarch64): Likewise
9389 (*movsf_aarch64): Likewise
9390 (*movdf_aarch64): Likewise
9392 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
9393 * config/arm/iwmmxt.md
9394 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
9395 * config/arm/arm1020e.md: Update with new attributes.
9396 * config/arm/cortex-a15-neon.md: Update with new attributes.
9397 * config/arm/cortex-a5.md: Update with new attributes.
9398 * config/arm/cortex-a53.md: Update with new attributes.
9399 * config/arm/cortex-a7.md: Update with new attributes.
9400 * config/arm/cortex-a8-neon.md: Update with new attributes.
9401 * config/arm/cortex-a9.md: Update with new attributes.
9402 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9403 * config/arm/cortex-r4f.md: Update with new attributes.
9404 * config/arm/marvell-pj4.md: Update with new attributes.
9405 * config/arm/vfp11.md: Update with new attributes.
9407 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9409 * config/aarch64/aarch64.md
9410 (*madd<mode>): Fix type attribute.
9411 (*maddsi_uxtw): Likewise.
9412 (*msub<mode>): Likewise.
9413 (*msubsi_uxtw): Likewise.
9414 (<su_optab>maddsidi4): Likewise.
9415 (<su_optab>msubsidi4): Likewise.
9417 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9419 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9420 * config/arm/arm.md (core_cycles): Remove fdiv.
9421 * config/arm/vfp.md:
9422 (*sqrtsf2_vfp): Update for attribute changes.
9423 (*sqrtdf2_vfp): Likewise.
9424 * config/aarch64/aarch64.md:
9425 (sqrt<mode>2): Update for attribute changes.
9426 * config/arm/arm1020e.md: Update with new attributes.
9427 * config/arm/cortex-a15-neon.md: Update with new attributes.
9428 * config/arm/cortex-a5.md: Update with new attributes.
9429 * config/arm/cortex-a53.md: Update with new attributes.
9430 * config/arm/cortex-a7.md: Update with new attributes.
9431 * config/arm/cortex-a8-neon.md: Update with new attributes.
9432 * config/arm/cortex-a9.md: Update with new attributes.
9433 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9434 * config/arm/cortex-r4f.md: Update with new attributes.
9435 * config/arm/marvell-pj4.md: Update with new attributes.
9436 * config/arm/vfp11.md: Update with new attributes.
9438 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9440 * config/arm/types.md
9441 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9442 * config/aarch64/aarch64.md
9443 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9445 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9446 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9447 (float<GPI:mode><GPF:mode>2): Likewise.
9449 (*truncsisf2_vfp): Update with new attributes.
9450 (*truncsidf2_vfp): Likewise.
9451 (fixuns_truncsfsi2): Likewise.
9452 (fixuns_truncdfsi2): Likewise.
9453 (*floatsisf2_vfp): Likewise.
9454 (*floatsidf2_vfp): Likewise.
9455 (floatunssisf2): Likewise.
9456 (floatunssidf2): Likewise.
9457 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9458 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9459 * config/arm/arm1020e.md: Update with new attributes.
9460 * config/arm/cortex-a15-neon.md: Update with new attributes.
9461 * config/arm/cortex-a5.md: Update with new attributes.
9462 * config/arm/cortex-a53.md: Update with new attributes.
9463 * config/arm/cortex-a7.md: Update with new attributes.
9464 * config/arm/cortex-a8-neon.md: Update with new attributes.
9465 * config/arm/cortex-a9.md: Update with new attributes.
9466 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9467 * config/arm/cortex-r4f.md: Update with new attributes.
9468 * config/arm/marvell-pj4.md: Update with new attributes.
9469 * config/arm/vfp11.md: Update with new attributes.
9471 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9473 * config/aarch64/arm_neon.h
9474 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9475 (vqtbx<1,2,3,4><q>_s8): Likewise.
9477 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9479 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9480 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9481 (add<mode>3): Add type attribute.
9482 (add<mode>3): Likewise.
9483 (usadd<mode>3): Likewise.
9484 (ssadd<mode>3): Likewise.
9485 (sub<mode>3): Likewise.
9486 (sub<mode>3): Likewise.
9487 (ussub<mode>3): Likewise.
9488 (sssub<mode>3): Likewise.
9489 (ssmulsa3): Likewise.
9490 (usmulusa3): Likewise.
9491 (arm_usatsihi): Likewise.
9493 (*movdi_vfp): Add types for all instructions.
9494 (*movdi_vfp_cortexa8): Likewise.
9495 (*movhf_vfp_neon): Likewise.
9496 (*movhf_vfp): Likewise.
9497 (*movdf_vfp): Likewise.
9498 (*thumb2_movdf_vfp): Likewise.
9499 (*thumb2_movdfcc_vfp): Likewise.
9500 * config/arm/arm.md: Add type attribute to all insn patterns.
9501 (*thumb1_adddi3): Add type attribute.
9502 (*arm_adddi3): Likewise.
9503 (*adddi_sesidi_di): Likewise.
9504 (*adddi_zesidi_di): Likewise.
9505 (*thumb1_addsi3): Likewise.
9506 (addsi3_compare0): Likewise.
9507 (*addsi3_compare0_scratch): Likewise.
9508 (*compare_negsi_si): Likewise.
9509 (cmpsi2_addneg): Likewise.
9510 (*addsi3_carryin_<optab>): Likewise.
9511 (*addsi3_carryin_alt2_<optab>): Likewise.
9512 (*addsi3_carryin_clobercc_<optab>): Likewise.
9513 (*subsi3_carryin): Likewise.
9514 (*subsi3_carryin_const): Likewise.
9515 (*subsi3_carryin_compare): Likewise.
9516 (*subsi3_carryin_compare_const): Likewise.
9517 (*arm_subdi3): Likewise.
9518 (*thumb_subdi3): Likewise.
9519 (*subdi_di_zesidi): Likewise.
9520 (*subdi_di_sesidi): Likewise.
9521 (*subdi_zesidi_di): Likewise.
9522 (*subdi_sesidi_di): Likewise.
9523 (*subdi_zesidi_ze): Likewise.
9524 (thumb1_subsi3_insn): Likewise.
9525 (*arm_subsi3_insn): Likewise.
9526 (*anddi3_insn): Likewise.
9527 (*anddi_zesidi_di): Likewise.
9528 (*anddi_sesdi_di): Likewise.
9529 (*ne_zeroextracts): Likewise.
9530 (*ne_zeroextracts): Likewise.
9531 (*ite_ne_zeroextr): Likewise.
9532 (*ite_ne_zeroextr): Likewise.
9533 (*anddi_notdi_di): Likewise.
9534 (*anddi_notzesidi): Likewise.
9535 (*anddi_notsesidi): Likewise.
9536 (andsi_notsi_si): Likewise.
9537 (thumb1_bicsi3): Likewise.
9538 (*iordi3_insn): Likewise.
9539 (*iordi_zesidi_di): Likewise.
9540 (*iordi_sesidi_di): Likewise.
9541 (*thumb1_iorsi3_insn): Likewise.
9542 (*xordi3_insn): Likewise.
9543 (*xordi_zesidi_di): Likewise.
9544 (*xordi_sesidi_di): Likewise.
9545 (*arm_xorsi3): Likewise.
9546 (*andsi_iorsi3_no): Likewise.
9547 (*smax_0): Likewise.
9548 (*smax_m1): Likewise.
9549 (*arm_smax_insn): Likewise.
9550 (*smin_0): Likewise.
9551 (*arm_smin_insn): Likewise.
9552 (*arm_umaxsi3): Likewise.
9553 (*arm_uminsi3): Likewise.
9554 (*minmax_arithsi): Likewise.
9555 (*minmax_arithsi_): Likewise.
9556 (*satsi_<SAT:code>): Likewise.
9557 (arm_ashldi3_1bit): Likewise.
9558 (arm_ashrdi3_1bit): Likewise.
9559 (arm_lshrdi3_1bit): Likewise.
9560 (*arm_negdi2): Likewise.
9561 (*thumb1_negdi2): Likewise.
9562 (*arm_negsi2): Likewise.
9563 (*thumb1_negsi2): Likewise.
9564 (*negdi_extendsid): Likewise.
9565 (*negdi_zero_extend): Likewise.
9566 (*arm_abssi2): Likewise.
9567 (*thumb1_abssi2): Likewise.
9568 (*arm_neg_abssi2): Likewise.
9569 (*thumb1_neg_abss): Likewise.
9570 (one_cmpldi2): Likewise.
9571 (extend<mode>di2): Likewise.
9572 (*compareqi_eq0): Likewise.
9573 (*arm_extendhisi2addsi): Likewise.
9574 (*arm_movdi): Likewise.
9575 (*thumb1_movdi_insn): Likewise.
9576 (*arm_movt): Likewise.
9577 (*thumb1_movsi_insn): Likewise.
9578 (pic_add_dot_plus_four): Likewise.
9579 (pic_add_dot_plus_eight): Likewise.
9580 (tls_load_dot_plus_eight): Likewise.
9581 (*thumb1_movhi_insn): Likewise.
9582 (*thumb1_movsf_insn): Likewise.
9583 (*movdf_soft_insn): Likewise.
9584 (*thumb_movdf_insn): Likewise.
9585 (cbranchsi4_insn): Likewise.
9586 (cbranchsi4_scratch): Likewise.
9587 (*negated_cbranchsi4): Likewise.
9588 (*tbit_cbranch): Likewise.
9589 (*tlobits_cbranch): Likewise.
9590 (*tstsi3_cbranch): Likewise.
9591 (*cbranchne_decr1): Likewise.
9592 (*addsi3_cbranch): Likewise.
9593 (*addsi3_cbranch_scratch): Likewise.
9594 (*arm_cmpdi_insn): Likewise.
9595 (*arm_cmpdi_unsig): Likewise.
9596 (*arm_cmpdi_zero): Likewise.
9597 (*thumb_cmpdi_zero): Likewise.
9598 (*deleted_compare): Likewise.
9599 (*mov_scc): Likewise.
9600 (*mov_negscc): Likewise.
9601 (*mov_notscc): Likewise.
9602 (*cstoresi_eq0_thumb1_insn): Likewise.
9603 (cstoresi_nltu_thumb1): Likewise.
9604 (cstoresi_ltu_thu): Likewise.
9605 (thumb1_addsi3_addgeu): Likewise.
9606 (*arm_jump): Likewise.
9607 (*thumb_jump): Likewise.
9608 (*check_arch2): Likewise.
9609 (arm_casesi_internal): Likewise.
9610 (thumb1_casesi_dispatch): Likewise.
9611 (*arm_indirect_jump): Likewise.
9612 (*thumb1_indirect_jump): Likewise.
9614 (*and_scc): Likewise.
9615 (*ior_scc): Likewise.
9616 (*compare_scc): Likewise.
9617 (*cond_move): Likewise.
9618 (*cond_arith): Likewise.
9619 (*cond_sub): Likewise.
9620 (*cmp_ite0): Likewise.
9621 (*cmp_ite1): Likewise.
9622 (*cmp_and): Likewise.
9623 (*cmp_ior): Likewise.
9624 (*ior_scc_scc): Likewise.
9625 (*ior_scc_scc_cmp): Likewise.
9626 (*and_scc_scc): Likewise.
9627 (*and_scc_scc_cmp): Likewise.
9628 (*and_scc_scc_nod): Likewise.
9629 (*negscc): Likewise.
9630 (movcond_addsi): Likewise.
9631 (movcond): Likewise.
9632 (*ifcompare_plus_move): Likewise.
9633 (*if_plus_move): Likewise.
9634 (*ifcompare_move_plus): Likewise.
9635 (*if_move_plus): Likewise.
9636 (*ifcompare_arith_arith): Likewise.
9637 (*if_arith_arith): Likewise.
9638 (*ifcompare_arith_move): Likewise.
9639 (*if_arith_move): Likewise.
9640 (*ifcompare_move_arith): Likewise.
9641 (*if_move_arith): Likewise.
9642 (*ifcompare_move_not): Likewise.
9643 (*if_move_not): Likewise.
9644 (*ifcompare_not_move): Likewise.
9645 (*if_not_move): Likewise.
9646 (*ifcompare_shift_move): Likewise.
9647 (*if_shift_move): Likewise.
9648 (*ifcompare_move_shift): Likewise.
9649 (*if_move_shift): Likewise.
9650 (*ifcompare_shift_shift): Likewise.
9651 (*ifcompare_not_arith): Likewise.
9652 (*ifcompare_arith_not): Likewise.
9653 (*if_arith_not): Likewise.
9654 (*ifcompare_neg_move): Likewise.
9655 (*if_neg_move): Likewise.
9656 (*ifcompare_move_neg): Likewise.
9657 (*if_move_neg): Likewise.
9658 (prologue_thumb1_interwork): Likewise.
9659 (*cond_move_not): Likewise.
9660 (*sign_extract_onebit): Likewise.
9661 (*not_signextract_onebit): Likewise.
9662 (stack_tie): Likewise.
9663 (align_4): Likewise.
9664 (align_8): Likewise.
9665 (consttable_end): Likewise.
9666 (consttable_1): Likewise.
9667 (consttable_2): Likewise.
9668 (consttable_4): Likewise.
9669 (consttable_8): Likewise.
9670 (consttable_16): Likewise.
9671 (*thumb1_tablejump): Likewise.
9672 (prefetch): Likewise.
9673 (force_register_use): Likewise.
9674 (thumb_eh_return): Likewise.
9675 (load_tp_hard): Likewise.
9676 (load_tp_soft): Likewise.
9677 (tlscall): Likewise.
9678 (*arm_movtas_ze): Likewise.
9679 (*arm_rev): Likewise.
9680 (*arm_revsh): Likewise.
9681 (*arm_rev16): Likewise.
9682 * config/arm/thumb2.md
9683 (*thumb2_smaxsi3): Likewise.
9684 (*thumb2_sminsi3): Likewise.
9685 (*thumb32_umaxsi3): Likewise.
9686 (*thumb2_uminsi3): Likewise.
9687 (*thumb2_negdi2): Likewise.
9688 (*thumb2_abssi2): Likewise.
9689 (*thumb2_neg_abss): Likewise.
9690 (*thumb2_movsi_insn): Likewise.
9691 (tls_load_dot_plus_four): Likewise.
9692 (*thumb2_movhi_insn): Likewise.
9693 (*thumb2_mov_scc): Likewise.
9694 (*thumb2_mov_negs): Likewise.
9695 (*thumb2_mov_negs): Likewise.
9696 (*thumb2_mov_nots): Likewise.
9697 (*thumb2_mov_nots): Likewise.
9698 (*thumb2_movsicc_): Likewise.
9699 (*thumb2_movsfcc_soft_insn): Likewise.
9700 (*thumb2_indirect_jump): Likewise.
9701 (*thumb2_and_scc): Likewise.
9702 (*thumb2_ior_scc): Likewise.
9703 (*thumb2_ior_scc_strict_it): Likewise.
9704 (*thumb2_cond_move): Likewise.
9705 (*thumb2_cond_arith): Likewise.
9706 (*thumb2_cond_ari): Likewise.
9707 (*thumb2_cond_sub): Likewise.
9708 (*thumb2_negscc): Likewise.
9709 (*thumb2_movcond): Likewise.
9710 (thumb2_casesi_internal): Likewise.
9711 (thumb2_casesi_internal_pic): Likewise.
9712 (*thumb2_alusi3_short): Likewise.
9713 (*thumb2_mov<mode>_shortim): Likewise.
9714 (*thumb2_addsi_short): Likewise.
9715 (*thumb2_subsi_short): Likewise.
9716 (thumb2_addsi3_compare0): Likewise.
9717 (*thumb2_cbz): Likewise.
9718 (*thumb2_cbnz): Likewise.
9719 (*thumb2_one_cmplsi2_short): Likewise.
9720 (*thumb2_negsi2_short): Likewise.
9721 (*orsi_notsi_si): Likewise.
9722 * config/arm/arm1020e.md: Update with new attributes.
9723 * config/arm/arm1026ejs.md: Update with new attributes.
9724 * config/arm/arm1136jfs.md: Update with new attributes.
9725 * config/arm/arm926ejs.md: Update with new attributes.
9726 * config/arm/cortex-a15.md: Update with new attributes.
9727 * config/arm/cortex-a5.md: Update with new attributes.
9728 * config/arm/cortex-a53.md: Update with new attributes.
9729 * config/arm/cortex-a7.md: Update with new attributes.
9730 * config/arm/cortex-a8.md: Update with new attributes.
9731 * config/arm/cortex-a9.md: Update with new attributes.
9732 * config/arm/cortex-m4.md: Update with new attributes.
9733 * config/arm/cortex-r4.md: Update with new attributes.
9734 * config/arm/fa526.md: Update with new attributes.
9735 * config/arm/fa606te.md: Update with new attributes.
9736 * config/arm/fa626te.md: Update with new attributes.
9737 * config/arm/fa726te.md: Update with new attributes.
9739 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9741 * config/aarch64/aarch64-simd.md
9742 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9743 <vwx> iterator to ensure correct register choice.
9744 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9745 (aarch64_sqdmull_n<mode>): Likewise.
9746 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9747 * config/aarch64/arm_neon.h
9748 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9749 (vml<as><q>_n_<su>16): Likewise.
9750 (vml<as>l_high_lane<q>_<su>16): Likewise.
9751 (vml<as>l_high_n_<su>16): Likewise.
9752 (vml<as>l_lane<q>_<su>16): Likewise.
9753 (vml<as>l_n_<su>16): Likewise.
9754 (vmul<q>_lane<q>_<su>16): Likewise.
9755 (vmul<q>_n_<su>16): Likewise.
9756 (vmull_lane<q>_<su>16): Likewise.
9757 (vmull_n_<su>16): Likewise.
9758 (vmull_high_lane<q>_<su>16): Likewise.
9759 (vmull_high_n_<su>16): Likewise.
9760 (vqrdmulh<q>_n_s16): Likewise.
9762 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9764 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9765 have the correct lane parameter.
9767 2013-09-06 Richard Biener <rguenther@suse.de>
9769 * cfganal.c (control_dependences::~control_dependences):
9770 Properly free all of the vector.
9772 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9775 * config/i386/i386.c (ix86_conditional_register_usage):
9776 Proper initialize extended SSE registers.
9778 2013-09-06 Jan Hubicka <jh@suse.cz>
9780 PR tree-optimization/58311
9781 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9783 2013-09-06 Jan Hubicka <jh@suse.cz>
9785 * Makefile.in (tree-sra.o): Update dependencies.
9786 * tree-sra.c: Include ipa-utils.h
9787 (scan_function): Use recursive_call_p.
9788 (has_caller_p): New function.
9789 (cgraph_for_node_and_aliases): Count also callers of aliases.
9791 2013-09-06 Jan Hubicka <jh@suse.cz>
9794 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9795 * tree-tailcall.c: Include ipa-utils.h.
9796 (find_tail_calls): Use it.
9797 * ipa-pure-const.c (check_call): Likewise.
9798 * ipa-utils.c (recursive_call_p): New function.
9799 * ipa-utils.h (recursive_call_p): Dclare.
9800 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9801 (symtab_semantically_equivalent_p): New function.
9802 * Makefile.in (tree-tailcall.o): Update dependencies.
9804 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9806 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9809 2013-09-06 Richard Biener <rguenther@suse.de>
9811 * lto-streamer.h (lto_global_var_decls): Remove.
9812 * Makefile.in (OBJS): Remove lto-symtab.o.
9813 (lto-symtab.o): Remove.
9814 (GTFILES): Remove lto-symtab.c
9815 * lto-symtab.c: Move to lto/
9817 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9819 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9820 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9821 (UNSPEC_FPINT_RINT): New constant definitions.
9822 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9823 definition with 2 attributes.
9824 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9825 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9828 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9830 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9831 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9832 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9833 attribute to "z196_alone".
9834 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9837 2013-09-06 Richard Biener <rguenther@suse.de>
9839 * basic-block.h (class control_dependences): New.
9840 * tree-ssa-dce.c (control_dependence_map): Remove.
9842 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9843 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9844 find_pdom, find_control_dependence, find_all_control_dependences):
9846 (mark_control_dependent_edges_necessary,
9847 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9848 tree_dce_done, perform_tree_ssa_dce): Adjust.
9849 * cfganal.c (set_control_dependence_map_bit,
9850 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9851 find_all_control_dependences): Move from tree-ssa-dce.c and
9852 implement as methods of control_dependences class.
9853 (control_dependences::control_dependences): New.
9854 (control_dependences::~control_dependences): Likewise.
9855 (control_dependences::get_edges_dependent_on): Likewise.
9856 (control_dependences::get_edge): Likewise.
9858 2013-09-04 Jan Hubicka <jh@suse.cz>
9860 * tree.c (types_same_for_odr): Drop overactive check.
9861 * ipa-devirt.c (hash_type_name): Likewise.
9863 2013-09-04 Jan Hubicka <jh@suse.cz>
9865 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9866 (analyze_functions): ... here.
9868 2013-09-04 Jan Hubicka <jh@suse.cz>
9871 * cgraphunit.c (analyze_functions): Clear AUX fields
9872 after processing; initialize assembler name has.
9874 2013-09-05 Jeff Law <law@redhat.com>
9876 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9877 from thread_around_empty_block. Record threading path into PATH.
9878 Recurse if threading through the initial block is successful.
9879 (thread_across_edge): Corresponding changes to slightly simplify.
9881 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9883 * config/aarch64/aarch64.md
9884 (type): Remove frecpe, frecps, frecpx.
9885 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9886 fix to be a TARGET_SIMD instruction.
9887 (aarch64_frecps): Remove.
9888 * config/aarch64/aarch64-simd.md
9889 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9890 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9892 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9893 Sofiane Naci <sofiane.naci@arm.com>
9895 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9896 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9897 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9898 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9899 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9900 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9901 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9902 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9903 into "clz, "rbit". Rename "shift" to "shift_imm".
9904 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9905 changes. Update for attribute changes all occurrences of arlo_* and
9907 * config/arm/arm-fixed.md: Update for attribute changes
9908 all occurrences of arlo_* types.
9909 * config/arm/thumb2.md: Update for attribute changes all occurrences
9911 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9912 (cortexa7_older_only): Likewise.
9913 (cortexa7_younger): Likewise.
9914 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9915 (1020alu_shift_op): Likewise.
9916 (1020alu_shift_reg_op): Likewise.
9917 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9918 (alu_shift_op): Likewise.
9919 (alu_shift_reg_op): Likewise.
9920 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9921 (11_alu_shift_op): Likewise.
9922 (11_alu_shift_reg_op): Likewise.
9923 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9924 (9_alu_shift_reg_op): Likewise.
9925 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9927 (cortex_a15_alu_shift): Likewise.
9928 (cortex_a15_alu_shift_reg): Likewise.
9929 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9931 (cortex_a5_alu_shift): Likewise.
9932 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9934 (cortex_a53_alu_shift): Likewise.
9935 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9937 (cortex_a7_alu_reg): Likewise.
9938 (cortex_a7_alu_shift): Likewise.
9939 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9941 (cortex_a8_alu_shift): Likewise.
9942 (cortex_a8_alu_shift_reg): Likewise.
9943 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9944 (cortex_a9_dp_shift): Likewise.
9945 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9947 * config/arm/cortex-r4.md
9948 (cortex_r4_alu): Update for attribute changes.
9949 (cortex_r4_mov): Likewise.
9950 (cortex_r4_alu_shift_reg): Likewise.
9951 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9952 (526_alu_shift_op): Likewise.
9953 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9954 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9955 (626te_alu_shift_op): Likewise.
9956 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9957 (726te_alu_shift_op): Likewise.
9958 (726te_alu_shift_reg_op): Likewise.
9959 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9960 (mp626_alu_shift_op): Likewise.
9961 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9962 (pj4_alu_conds): Likewise.
9963 (pj4_shift): Likewise.
9964 (pj4_shift_conds): Likewise.
9965 (pj4_alu_shift): Likewise.
9966 (pj4_alu_shift_conds): Likewise.
9967 * config/aarch64/aarch64.md: Update for attribute change
9968 all occurrences of arlo_* and shift* types.
9970 2013-09-05 Mike Stump <mikestump@comcast.net>
9972 * tree.h: Move documentation for tree_function_decl to tree-core.h
9973 with the declaration.
9975 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9978 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9979 looking for widest mode.
9981 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9983 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9985 2013-09-05 Richard Biener <rguenther@suse.de>
9987 PR tree-optimization/58137
9988 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9989 Do not create vectors of pointers.
9990 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9991 types for the components of the vector initializer.
9992 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9993 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9995 2013-09-05 Martin Jambor <mjambor@suse.cz>
9997 * ipa-prop.c (remove_described_reference): Accept missing references,
9998 return false if that hppens, otherwise return true.
9999 (cgraph_node_for_jfunc): New function.
10000 (try_decrement_rdesc_refcount): Likewise.
10001 (try_make_edge_direct_simple_call): Use them.
10002 (ipa_edge_removal_hook): Remove references from rdescs.
10003 (ipa_edge_duplication_hook): Clone rdescs and their references
10004 when the new edge has the same caller as the old one.
10005 * cgraph.c (cgraph_resolve_speculation): Remove speculative
10006 reference before removing any edges.
10008 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
10010 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
10012 * thumb2.md (thumb2_storewb_parisi): New pattern.
10014 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
10016 * config/aarch64/aarch64-option-extensions.def: Add
10017 AARCH64_OPT_EXTENSION of 'crc'.
10018 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
10019 (AARCH64_ISA_CRC): Ditto.
10020 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
10021 description of the CRC extension.
10023 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
10025 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
10026 * config/rs6000/linux.h: Ditto.
10027 * alpha/linux.h: Ditto.
10028 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
10029 no_c99_libc_has_function.
10030 * config/c6x/uclinux-elf.h: Ditto.
10031 * config/lm32/uclinux-elf.h: Ditto.
10032 * config/m68k/uclinux.h: Ditto.
10033 * config/moxie/uclinux.h: Ditto.
10034 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
10035 (crisv32-*-linux*, cris-*-linux*): Ditto.
10036 * config/bfin/bfin.c: Include "tm_p.h".
10038 2013-09-05 Richard Biener <rguenther@suse.de>
10040 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
10041 check for a definition without a basic-block.
10043 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10044 Sofiane Naci <sofiane.naci@arm.com>
10046 * config/aarch64/aarch64.md
10047 (*movti_aarch64): Rename r_2_f and f_2_r.
10048 (*movsf_aarch64): Likewise.
10049 (*movdf_aarch64): Likewise.
10050 (*movtf_aarch64): Likewise.
10051 (aarch64_movdi_<mode>low): Likewise.
10052 (aarch64_movdi_<mode>high): Likewise.
10053 (aarch64_mov<mode>high_di): Likewise.
10054 (aarch64_mov<mode>low_di): Likewise.
10055 (aarch64_movtilow_tilow): Likewise.
10056 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
10057 values to config/arm/types.md
10058 (attribute "conds"): Update for attribute change.
10059 (anddi3_insn): Likewise.
10060 (iordi3_insn): Likewise.
10061 (xordi3_insn): Likewise.
10062 (one_cmpldi2): Likewise.
10063 * config/arm/types.md (type): Add Neon types.
10064 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
10065 use "type" attribute.
10066 (movmisalign<mode>_neon_store): Likewise.
10067 (movmisalign<mode>_neon_load): Likewise.
10068 (vec_set<mode>_internal): Likewise.
10069 (vec_setv2di_internal): Likewise.
10070 (vec_extract<mode>): Likewise.
10071 (vec_extractv2di): Likewise.
10072 (add<mode>3_neon): Likewise.
10073 (adddi3_neon): Likewise.
10074 (sub<mode>3_neon): Likewise.
10075 (subdi3_neon): Likewise.
10076 (mul<mode>3_neon): Likewise.
10077 (mul<mode>3add<mode>_neon): Likewise.
10078 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
10079 (fma<VCVTF:mode>4)): Likewise.
10080 (fma<VCVTF:mode>4_intrinsic): Likewise.
10081 (fmsub<VCVTF:mode>4)): Likewise.
10082 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
10083 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
10084 (ior<mode>3): Likewise.
10085 (and<mode>3): Likewise.
10086 (anddi3_neon): Likewise.
10087 (orn<mode>3_neon): Likewise.
10088 (orndi3_neon): Likewise.
10089 (bic<mode>3_neon): Likewise.
10090 (bicdi3_neon): Likewise.
10091 (xor<mode>3): Likewise.
10092 (one_cmpl<mode>2): Likewise.
10093 (abs<mode>2): Likewise.
10094 (neg<mode>2): Likewise.
10095 (umin<mode>3_neon): Likewise.
10096 (umax<mode>3_neon): Likewise.
10097 (smin<mode>3_neon): Likewise.
10098 (smax<mode>3_neon): Likewise.
10099 (vashl<mode>3): Likewise.
10100 (vashr<mode>3_imm): Likewise.
10101 (vlshr<mode>3_imm): Likewise.
10102 (ashl<mode>3_signed): Likewise.
10103 (ashl<mode>3_unsigned): Likewise.
10104 (neon_load_count): Likewise.
10105 (ashldi3_neon_noclobber): Likewise.
10106 (signed_shift_di3_neon): Likewise.
10107 (unsigned_shift_di3_neon): Likewise.
10108 (ashrdi3_neon_imm_noclobber): Likewise.
10109 (lshrdi3_neon_imm_noclobber): Likewise.
10110 (widen_ssum<mode>3): Likewise.
10111 (widen_usum<mode>3): Likewise.
10112 (quad_halves_<code>v4si): Likewise.
10113 (quad_halves_<code>v4sf): Likewise.
10114 (quad_halves_<code>v8hi): Likewise.
10115 (quad_halves_<code>v16qi): Likewise.
10116 (reduc_splus_v2di): Likewise.
10117 (neon_vpadd_internal<mode>): Likewise.
10118 (neon_vpsmin<mode>): Likewise.
10119 (neon_vpsmax<mode>): Likewise.
10120 (neon_vpumin<mode>): Likewise.
10121 (neon_vpumax<mode>): Likewise.
10122 (ss_add<mode>_neon): Likewise.
10123 (us_add<mode>_neon): Likewise.
10124 (ss_sub<mode>_neon): Likewise.
10125 (us_sub<mode>_neon): Likewise.
10126 (neon_vadd<mode>_unspec): Likewise.
10127 (neon_vaddl<mode>): Likewise.
10128 (neon_vaddw<mode>): Likewise.
10129 (neon_vhadd<mode>): Likewise.
10130 (neon_vqadd<mode>): Likewise.
10131 (neon_vaddhn<mode>): Likewise.
10132 (neon_vmul<mode>): Likewise.
10133 (neon_vmla<mode>): Likewise.
10134 (neon_vmlal<mode>): Likewise.
10135 (neon_vmls<mode>): Likewise.
10136 (neon_vmlsl<mode>): Likewise.
10137 (neon_vqdmulh<mode>): Likewise.
10138 (neon_vqdmlal<mode>): Likewise.
10139 (neon_vqdmlsl<mode>): Likewise.
10140 (neon_vmull<mode>): Likewise.
10141 (neon_vqdmull<mode>): Likewise.
10142 (neon_vsub<mode>_unspec): Likewise.
10143 (neon_vsubl<mode>): Likewise.
10144 (neon_vsubw<mode>): Likewise.
10145 (neon_vqsub<mode>): Likewise.
10146 (neon_vhsub<mode>): Likewise.
10147 (neon_vsubhn<mode>): Likewise.
10148 (neon_vceq<mode>): Likewise.
10149 (neon_vcge<mode>): Likewise.
10150 (neon_vcgeu<mode>): Likewise.
10151 (neon_vcgt<mode>): Likewise.
10152 (neon_vcgtu<mode>): Likewise.
10153 (neon_vcle<mode>): Likewise.
10154 (neon_vclt<mode>): Likewise.
10155 (neon_vcage<mode>): Likewise.
10156 (neon_vcagt<mode>): Likewise.
10157 (neon_vtst<mode>): Likewise.
10158 (neon_vabd<mode>): Likewise.
10159 (neon_vabdl<mode>): Likewise.
10160 (neon_vaba<mode>): Likewise.
10161 (neon_vabal<mode>): Likewise.
10162 (neon_vmax<mode>): Likewise.
10163 (neon_vmin<mode>): Likewise.
10164 (neon_vpaddl<mode>): Likewise.
10165 (neon_vpadal<mode>): Likewise.
10166 (neon_vpmax<mode>): Likewise.
10167 (neon_vpmin<mode>): Likewise.
10168 (neon_vrecps<mode>): Likewise.
10169 (neon_vrsqrts<mode>): Likewise.
10170 (neon_vqabs<mode>): Likewise.
10171 (neon_vqneg<mode>): Likewise.
10172 (neon_vcls<mode>): Likewise.
10173 (clz<mode>2): Likewise.
10174 (popcount<mode>2): Likewise.
10175 (neon_vrecpe): Likewise.
10176 (neon_vrsqrte): Likewise.
10177 (neon_vget_lane<mode>_sext_internal): Likewise.
10178 (neon_vget_lane<mode>_zext_internal): Likewise.
10179 (neon_vdup_n<mode>): Likewise.
10180 (neon_vdup_nv2di): Likewise.
10181 (neon_vdpu_lane<mode>_internal): Likewise.
10182 (neon_vswp<mode>): Likewise.
10183 (float<mode><V_cvtto>2): Likewise.
10184 (floatuns<mode><V_cvtto>2): Likewise.
10185 (fix_trunc<mode><V_cvtto>)2): Likewise
10186 (fixuns_trunc<mode><V_cvtto)2): Likewise.
10187 (neon_vcvt<mode>): Likewise.
10188 (neon_vcvtv4sfv4hf): Likewise.
10189 (neon_vcvtv4hfv4sf): Likewise.
10190 (neon_vcvt_n<mode>): Likewise.
10191 (neon_vmovn<mode>): Likewise.
10192 (neon_vqmovn<mode>): Likewise.
10193 (neon_vqmovun<mode>): Likewise.
10194 (neon_vmovl<mode>): Likewise.
10195 (neon_vmul_lane<mode>): Likewise.
10196 (neon_vmull_lane<mode>): Likewise.
10197 (neon_vqdmull_lane<mode>): Likewise.
10198 (neon_vqdmulh_lane<mode>): Likewise.
10199 (neon_vmla_lane<mode>): Likewise.
10200 (neon_vmlal_lane<mode>): Likewise.
10201 (neon_vqdmlal_lane<mode>): Likewise.
10202 (neon_vmls_lane<mode>): Likewise.
10203 (neon_vmlsl_lane<mode>): Likewise.
10204 (neon_vqdmlsl_lane<mode>): Likewise.
10205 (neon_vext<mode>): Likewise.
10206 (neon_vrev64<mode>): Likewise.
10207 (neon_vrev32<mode>): Likewise.
10208 (neon_vrev16<mode>): Likewise.
10209 (neon_vbsl<mode>_internal): Likewise.
10210 (neon_vshl<mode>): Likewise.
10211 (neon_vqshl<mode>): Likewise.
10212 (neon_vshr_n<mode>): Likewise.
10213 (neon_vshrn_n<mode>): Likewise.
10214 (neon_vqshrn_n<mode>): Likewise.
10215 (neon_vqshrun_n<mode>): Likewise.
10216 (neon_vshl_n<mode>): Likewise.
10217 (neon_vqshl_n<mode>): Likewise.
10218 (neon_vqshlu_n<mode>): Likewise.
10219 (neon_vshll_n<mode>): Likewise.
10220 (neon_vsra_n<mode>): Likewise.
10221 (neon_vsri_n<mode>): Likewise.
10222 (neon_vsli_n<mode>): Likewise.
10223 (neon_vtbl1v8qi): Likewise.
10224 (neon_vtbl2v8qi): Likewise.
10225 (neon_vtbl3v8qi): Likewise.
10226 (neon_vtbl4v8qi): Likewise.
10227 (neon_vtbx1v8qi): Likewise.
10228 (neon_vtbx2v8qi): Likewise.
10229 (neon_vtbx3v8qi): Likewise.
10230 (neon_vtbx4v8qi): Likewise.
10231 (neon_vtrn<mode>_internal): Likewise.
10232 (neon_vzip<mode>_internal): Likewise.
10233 (neon_vuzp<mode>_internal): Likewise.
10234 (neon_vld1<mode>): Likewise.
10235 (neon_vld1_lane<mode>): Likewise.
10236 (neon_vld1_dup<mode>): Likewise.
10237 (neon_vld1_dupv2di): Likewise.
10238 (neon_vst1<mode>): Likewise.
10239 (neon_vst1_lane<mode>): Likewise.
10240 (neon_vld2<mode>): Likewise.
10241 (neon_vld2_lane<mode>): Likewise.
10242 (neon_vld2_dup<mode>): Likewise.
10243 (neon_vst2<mode>): Likewise.
10244 (neon_vst2_lane<mode>): Likewise.
10245 (neon_vld3<mode>): Likewise.
10246 (neon_vld3qa<mode>): Likewise.
10247 (neon_vld3qb<mode>): Likewise.
10248 (neon_vld3_lane<mode>): Likewise.
10249 (neon_vld3_dup<mode>): Likewise.
10250 (neon_vst3<mode>): Likewise.
10251 (neon_vst3qa<mode>): Likewise.
10252 (neon_vst3qb<mode>): Likewise.
10253 (neon_vst3_lane<mode>): Likewise.
10254 (neon_vld4<mode>): Likewise.
10255 (neon_vld4qa<mode>): Likewise.
10256 (neon_vld4qb<mode>): Likewise.
10257 (neon_vld4_lane<mode>): Likewise.
10258 (neon_vld4_dup<mode>): Likewise.
10259 (neon_vst4<mode>): Likewise.
10260 (neon_vst4qa<mode>): Likewise.
10261 (neon_vst4qb<mode>): Likewise.
10262 (neon_vst4_lane<mode>): Likewise.
10263 (neon_vec_unpack<US>_lo_<mode>): Likewise.
10264 (neon_vec_unpack<US>_hi_<mode>): Likewise.
10265 (neon_vec_<US>mult_lo_<mode>): Likewise.
10266 (neon_vec_<US>mult_hi_<mode>): Likewise.
10267 (neon_vec_<US>shiftl_<mode>): Likewise.
10268 (neon_unpack<US>_<mode>): Likewise.
10269 (neon_vec_<US>mult_<mode>): Likewise.
10270 (vec_pack_trunc_<mode>): Likewise.
10271 (neon_vec_pack_trunk_<mode>): Likewise.
10272 (neon_vabd<mode>_2): Likewise.
10273 (neon_vabd<mode>_3): Likewise.
10274 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
10275 (thumb2_movsi_vfp): Likewise.
10276 (movdi_vfp): Likewise.
10277 (movdi_vfp_cortexa8): Likewise.
10278 (movhf_vfp_neon): Likewise.
10279 (movhf_vfp): Likewiwse.
10280 (movsf_vfp): Likewiwse.
10281 (thumb2_movsf_vfp): Likewiwse.
10282 (movdf_vfp): Likewise.
10283 (thumb2_movdf_vfp): Likewise.
10284 (movsfcc_vfp): Likewise.
10285 (thumb2_movsfcc_vfp): Likewise.
10286 (movdfcc_vfp): Likewise.
10287 (thumb2_movdfcc_vfp): Likewise.
10288 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
10289 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
10290 (v10_v2c): Likewise.
10291 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
10293 (cortex_a15_neon_int_2): Likewise.
10294 (cortex_a15_neon_int_3): Likewise.
10295 (cortex_a15_neon_int_4): Likewise.
10296 (cortex_a15_neon_int_5): Likewise.
10297 (cortex_a15_neon_vqneg_vqabs): Likewise.
10298 (cortex_a15_neon_vmov): Likewise.
10299 (cortex_a15_neon_vaba): Likewise.
10300 (cortex_a15_neon_vaba_qqq): Likewise.
10301 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10302 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10303 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
10304 scalar_64_32_long_scalar): Likewise.
10305 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10306 (cortex_a15_neon_mla_qqq_8_16): Likewise.
10307 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10308 lotype_qdd_64_32_long): Likewise.
10309 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10310 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10311 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
10312 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10313 (cortex_a15_neon_shift_1): Likewise.
10314 (cortex_a15_neon_shift_2): Likewise.
10315 (cortex_a15_neon_shift_3): Likewise.
10316 (cortex_a15_neon_vshl_ddd): Likewise.
10317 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10318 (cortex_a15_neon_vsra_vrsra): Likewise.
10319 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
10320 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
10321 (cortex_a15_neon_fp_vmul_ddd): Likewise.
10322 (cortex_a15_neon_fp_vmul_qqd): Likewise.
10323 (cortex_a15_neon_fp_vmla_ddd): Likewise.
10324 (cortex_a15_neon_fp_vmla_qqq): Likewise.
10325 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
10326 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
10327 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10328 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10329 (cortex_a15_neon_bp_simple): Likewise.
10330 (cortex_a15_neon_bp_2cycle): Likewise.
10331 (cortex_a15_neon_bp_3cycle): Likewise.
10332 (cortex_a15_neon_vld1_1_2_regs): Likewise.
10333 (cortex_a15_neon_vld1_3_4_regs): Likewise.
10334 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10335 (cortex_a15_neon_vld2_4_regs): Likewise.
10336 (cortex_a15_neon_vld3_vld4): Likewise.
10337 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10338 (cortex_a15_neon_vst1_3_4_regs): Likewise.
10339 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
10340 (cortex_a15_neon_vst3_vst4): Likewise.
10341 (cortex_a15_neon_vld1_vld2_lane): Likewise.
10342 (cortex_a15_neon_vld3_vld4_lane" 10
10343 (cortex_a15_neon_vst1_vst2_lane): Likewise.
10344 (cortex_a15_neon_vst3_vst4_lane): Likewise.
10345 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
10346 (cortex_a15_neon_ldm_2): Likewise.0
10347 (cortex_a15_neon_stm_2): Likewise.
10348 (cortex_a15_neon_mcr): Likewise.
10349 (cortex_a15_neon_mcr_2_mcrr): Likewise.
10350 (cortex_a15_neon_mrc): Likewise.
10351 (cortex_a15_neon_mrrc): Likewise.
10352 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10354 (cortex_a15_alu_shift): Likewise.
10355 (cortex_a15_alu_shift_reg): Likewise.
10356 (cortex_a15_mult32): Likewise.
10357 (cortex_a15_mult64): Likewise.
10358 (cortex_a15_block): Likewise.
10359 (cortex_a15_branch): Likewise.
10360 (cortex_a15_load1): Likewise.
10361 (cortex_a15_load3): Likewise.
10362 (cortex_a15_store1): Likewise.
10363 (cortex_a15_store3): Likewise.
10364 (cortex_a15_call): Likewise.
10365 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
10366 (cortex_a5_f2r): Likewise.
10367 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
10369 (cortex_a53_f2r): Likewise.
10370 * config/arm/cortex-a7.md
10371 (cortex_a7_branch): Update for attribute change.
10372 (cortex_a7_call): Likewise.
10373 (cortex_a7_alu_imm): Likewise.
10374 (cortex_a7_alu_reg): Likewise.
10375 (cortex_a7_alu_shift): Likewise.
10376 (cortex_a7_mul): Likewise.
10377 (cortex_a7_load1): Likewise.
10378 (cortex_a7_store1): Likewise.
10379 (cortex_a7_load2): Likewise.
10380 (cortex_a7_store2): Likewise.
10381 (cortex_a7_load3): Likewise.
10382 (cortex_a7_store3): Likewise.
10383 (cortex_a7_load4): Likewise.
10384 (cortex_a7_store4): Likewise.
10385 (cortex_a7_fpalu): Likewise.
10386 (cortex_a7_fconst): Likewise.
10387 (cortex_a7_fpmuls): Likewise.
10388 (cortex_a7_neon_mul): Likewise.
10389 (cortex_a7_fpmacs): Likewise.
10390 (cortex_a7_neon_mla: Likewise.
10391 (cortex_a7_fpmuld: Likewise.
10392 (cortex_a7_fpmacd: Likewise.
10393 (cortex_a7_fpfmad: Likewise.
10394 (cortex_a7_fdivs: Likewise.
10395 (cortex_a7_fdivd: Likewise.
10396 (cortex_a7_r2f: Likewise.
10397 (cortex_a7_f2r: Likewise.
10398 (cortex_a7_f_flags: Likewise.
10399 (cortex_a7_f_loads: Likewise.
10400 (cortex_a7_f_loadd: Likewise.
10401 (cortex_a7_f_stores: Likewise.
10402 (cortex_a7_f_stored: Likewise.
10403 (cortex_a7_neon): Likewise.
10404 * config/arm/cortex-a8-neon.md
10405 (cortex_a8_neon_mrc): Update for attribute change.
10406 (cortex_a8_neon_mrrc): Likewise.
10407 (cortex_a8_neon_int_1): Likewise.
10408 (cortex_a8_neon_int_2): Likewise.
10409 (cortex_a8_neon_int_3): Likewise.
10410 (cortex_a8_neon_int_4): Likewise.
10411 (cortex_a8_neon_int_5): Likewise.
10412 (cortex_a8_neon_vqneg_vqabs): Likewise.
10413 (cortex_a8_neon_vmov): Likewise.
10414 (cortex_a8_neon_vaba): Likewise.
10415 (cortex_a8_neon_vaba_qqq): Likewise.
10416 (cortex_a8_neon_vsma): Likewise.
10417 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10418 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10419 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10420 long_scalar): Likewise.
10421 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10422 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10423 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10424 long_scalar_qdd_64_32_long): Likewise.
10425 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10426 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10427 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10428 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10429 (cortex_a8_neon_shift_1): Likewise.
10430 (cortex_a8_neon_shift_2): Likewise.
10431 (cortex_a8_neon_shift_3): Likewise.
10432 (cortex_a8_neon_vshl_ddd): Likewise.
10433 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10434 (cortex_a8_neon_vsra_vrsra): Likewise.
10435 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10436 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10437 (cortex_a8_neon_fp_vsum): Likewise.
10438 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10439 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10440 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10441 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10442 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10443 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10444 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10445 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10446 (cortex_a8_neon_bp_simple): Likewise.
10447 (cortex_a8_neon_bp_2cycle): Likewise.
10448 (cortex_a8_neon_bp_3cycle): Likewise.
10449 (cortex_a8_neon_ldr): Likewise.
10450 (cortex_a8_neon_str): Likewise.
10451 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10452 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10453 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10454 (cortex_a8_neon_vld2_4_regs): Likewise.
10455 (cortex_a8_neon_vld3_vld4): Likewise.
10456 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10457 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10458 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10459 (cortex_a8_neon_vst3_vst4): Likewise.
10460 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10461 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10462 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10463 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10464 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10465 (cortex_a8_neon_mcr): Likewise.
10466 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10467 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10468 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10470 (ca9_neon_mrrc): Likewise.
10471 (cortex_a9_neon_int_1): Likewise.
10472 (cortex_a9_neon_int_2): Likewise.
10473 (cortex_a9_neon_int_3): Likewise.
10474 (cortex_a9_neon_int_4): Likewise.
10475 (cortex_a9_neon_int_5): Likewise.
10476 (cortex_a9_neon_vqneg_vqabs): Likewise.
10477 (cortex_a9_neon_vmov): Likewise.
10478 (cortex_a9_neon_vaba): Likewise.
10479 (cortex_a9_neon_vaba_qqq): Likewise.
10480 (cortex_a9_neon_vsma): Likewise.
10481 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10482 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10483 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10484 long_scalar): Likewise.
10485 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10486 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10487 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10488 long_scalar_qdd_64_32_long): Likewise.
10489 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10490 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10491 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10492 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10493 (cortex_a9_neon_shift_1): Likewise.
10494 (cortex_a9_neon_shift_2): Likewise.
10495 (cortex_a9_neon_shift_3): Likewise.
10496 (cortex_a9_neon_vshl_ddd): Likewise.
10497 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10498 (cortex_a9_neon_vsra_vrsra): Likewise.
10499 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10500 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10501 (cortex_a9_neon_fp_vsum): Likewise.
10502 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10503 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10504 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10505 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10506 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10507 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10508 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10509 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10510 (cortex_a9_neon_bp_simple): Likewise.
10511 (cortex_a9_neon_bp_2cycle): Likewise.
10512 (cortex_a9_neon_bp_3cycle): Likewise.
10513 (cortex_a9_neon_ldr): Likewise.
10514 (cortex_a9_neon_str): Likewise.
10515 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10516 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10517 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10518 (cortex_a9_neon_vld2_4_regs): Likewise.
10519 (cortex_a9_neon_vld3_vld4): Likewise.
10520 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10521 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10522 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10523 (cortex_a9_neon_vst3_vst4): Likewise.
10524 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10525 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10526 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10527 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10528 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10529 (cortex_a9_neon_mcr): Likewise.
10530 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10531 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10532 (cortex_a9_fps): Likewise.
10533 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10535 (cortex_m4_fmuls): Likewise.
10536 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10538 (cortex_r4_mrc): Likewise.
10539 * config/arm/iterators.md: Update comment referring to neon_type.
10540 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10541 (iwmmxt_movsi_insn): Likewise.
10542 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10544 (pj4_core_to_vfp): Likewise.
10545 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10547 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10548 (vfp_fstore): Likewise.
10549 * doc/md.texi: Change references to neon_type to refer to type.
10551 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10553 * tree.h (DECL_BUILT_IN): Fix typo in comment.
10555 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
10557 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
10558 lglobl if not weak.
10560 2013-09-04 Easwaran Raman <eraman@google.com>
10562 PR middle-end/57370
10563 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
10564 (build_and_add_sum): Use it.
10565 (appears_later_in_bb): Simplify code.
10567 2013-09-04 Teresa Johnson <tejohnson@google.com>
10569 * dumpfile.c (dump_finish): Don't close stderr/stdout.
10571 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
10573 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
10575 2013-09-04 Jan Hubicka <jh@suse.cz>
10577 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
10578 * ipa-devirt.c: Include diganostic.h
10579 (odr_type_d): Add types and types_set.
10580 (hash_type_name): Work for types with vtables during LTO.
10581 (odr_hasher::remove): Fix comment; destroy types_set.
10582 (add_type_duplicate): New function,
10583 (get_odr_type): Use it.
10584 (dump_type_inheritance_graph): Dump type duplicates.
10585 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
10587 * tree.c (types_same_for_odr): Give exact answers on types with
10590 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10592 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
10593 explaining their differences.
10595 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
10597 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
10599 2013-09-03 Jeff Law <law@redhat.com>
10601 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
10602 when not threading through a joiner block. Pass joiner/no joiner
10603 state to register_jump_thread.
10604 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10605 state from argument rather than implying on path length.
10606 Dump the entire jump thread path into debugging dump.
10607 * tree-flow.h (register_jump_thread): Update prototype.
10609 2013-08-29 Xinliang David Li <davidxl@google.com>
10611 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10612 Remove a trivial gcc_assert.
10614 2013-08-29 Xinliang David Li <davidxl@google.com>
10616 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10617 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10618 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10619 Delay base decl alignment adjustment.
10620 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10621 * tree-vectorizer.h: New data structure.
10622 (set_dr_misalignment): New function.
10623 (dr_misalignment): Ditto.
10624 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10625 (vectorizable_load): Ditto.
10626 (ensure_base_align): New function.
10627 (vectorize_loops): Add dbg_cnt support.
10628 (execute_vect_slp): Ditto.
10629 * dbgcnt.def: New debug counter.
10630 * Makefile: New dependency.
10632 2013-09-03 Meador Inge <meadori@codesourcery.com>
10636 2013-08-30 Meador Inge <meadori@codesourcery.com>
10638 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10640 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10642 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10643 function descriptor.
10645 2013-09-03 Richard Biener <rguenther@suse.de>
10647 * tree-affine.c (add_elt_to_tree): Fix association issue,
10648 avoid useless converts and make sure to always return a
10649 properly typed result.
10651 2013-09-03 Richard Biener <rguenther@suse.de>
10653 PR middle-end/57656
10654 * fold-const.c (negate_expr_p): Fix division case.
10655 (negate_expr): Likewise.
10657 2013-09-03 Richard Biener <rguenther@suse.de>
10660 * tree-streamer-out.c: Include tm.h.
10661 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10663 2013-09-03 Jan Hubicka <jh@suse.cz>
10665 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10667 2013-09-03 Alan Modra <amodra@gmail.com>
10669 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10670 multilibs to big-endian target and vice versa.
10671 * config/rs6000/t-linux64: Use := assignment on all vars.
10672 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10673 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10674 * config/rs6000/t-linux64le: New file.
10675 * config/rs6000/t-linux64bele: New file.
10676 * config/rs6000/t-linux64lebe: New file.
10678 2013-09-02 Jan Hubicka <jh@suse.cz>
10680 * ipa-inline-transform.c (inline_transform): Do not
10681 optimize_inline_calls when not optimizing.
10683 2013-09-02 Jan Hubicka <jh@suse.cz>
10685 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10686 duplicated nodes for assembler names.
10687 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10688 hard registers from assembler name hash.
10690 2013-09-02 Jan Hubicka <jh@suse.cz>
10692 * ipa-split.c (execute_split_functions): Split externally visible
10693 functions called once.
10695 2013-09-02 Martin Jambor <mjambor@suse.cz>
10698 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10699 linked list. When finding the correct duplicate, also consider also
10700 the caller in additon to its inlined_to node.
10702 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10704 * config/aarch64/aarch64-simd-builtins.def
10705 (dup_lane_scalar): Remove.
10706 * config/aarch64/aarch64-simd.md
10707 (aarch64_simd_dup): Add 'w->w' alternative.
10708 (aarch64_dup_lane<mode>): Allow for VALL.
10709 (aarch64_dup_lane_scalar<mode>): Remove.
10710 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10711 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10712 (aarch64_get_lane_unsigned<mode>): Likewise.
10713 (aarch64_get_lane<mode>): Likewise.
10714 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10715 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10716 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10717 (VCON): Change container of V2SF.
10718 (vswap_width_name): Likewise.
10719 * config/aarch64/arm_neon.h
10720 (__aarch64_vdup_lane_any): New.
10721 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10722 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10723 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10725 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10727 PR middle-end/56382
10728 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10729 the source or the destination is a single hard register.
10731 2013-09-02 Richard Biener <rguenther@suse.de>
10733 PR middle-end/57511
10734 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10737 2013-09-02 Richard Biener <rguenther@suse.de>
10739 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10740 arithmetic to sizetype.
10742 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10744 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10745 Find auto-increment use both before and after candidate.
10747 2013-09-02 Marek Polacek <polacek@redhat.com>
10749 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10751 2013-09-01 Jan Hubicka <jh@suse.cz>
10753 * Makefile.in: Add ipa-profile.o
10754 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10755 * cgraph.c (struct cgraph_propagate_frequency_data,
10756 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10757 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10758 * cgraph.h (cgraph_propagate_frequency): Remove.
10759 * ipa-inline-analysis.c: Include ipa-utils.h;
10760 drop duplicated cfgloop.h.
10761 (inline_update_callee_summaries): Update.
10762 * ipa-profile.c: New file.
10763 * ipa-utils.h (ipa_propagate_frequency): Declare.
10764 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10765 data-streamer.h, value-prof.h.
10766 (symtab_remove_unreachable_nodes): Update profile.
10767 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10768 account_time_size, cmp_counts, dump_histogram,
10769 ipa_profile_generate_summary, ipa_profile_write_summary,
10770 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10771 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10772 Move to ipa-profile.c.
10774 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10776 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10778 2013-09-01 Jan Hubicka <jh@suse.cz>
10780 * common.opt (fdevirtualize-speculatively): New function.
10781 * invoke.texi (fdevirtualize-speculatively): Document.
10782 * ipa-devirt.c: Include ipa-inline.h
10783 (likely_target_p): New function.
10784 (ipa_devirt): New function.
10785 (gate_ipa_devirt): New function.
10786 (pass_data_ipa_devirt): New static var.
10787 (pass_ipa_devirt): Likewise.
10788 (make_pass_ipa_devirt): New function.
10789 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10790 (common_handle_option): Disable devirtualization when
10791 value range profiling is available.
10792 * passes.def (pass_ipa_devirt): Add.
10793 * timever.def (TV_IPA_DEVIRT): New timevar.
10794 * tree-pass.h (make_pass_ipa_devirt):
10796 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10798 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10799 include sanitize(undefined).
10801 2013-08-31 Diego Novillo <dnovillo@google.com>
10803 * Makefile.in (TREE_CORE_H): Define.
10805 (GTFILES): Add tree-core.h.
10806 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10808 * tree-core.h: New file.
10809 Move all data structures, enum, typedefs, global
10810 declarations and constants from ...
10811 * tree.h: ... here.
10813 2013-08-31 Jan Hubicka <jh@suse.cz>
10815 * bulitins.c (expand_builtin): Do not early exit for gcov
10816 instrumented functions.
10818 2013-08-31 Marek Polacek <polacek@redhat.com>
10820 * ubsan.c: Include tm_p.h.
10822 2013-08-31 Jan Hubicka <jh@suse.cz>
10824 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10827 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10828 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10829 * tree-cfg.c (verify_gimple_label): ... here.
10830 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10831 (ipa_merge_profiles): New function.
10832 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10833 (lto_input_function_body): Likewise.
10834 * ipa-utils.h (ipa_merge_profiles): Declare.
10835 * lto-streamer.h (lto_input_function_body): Update prototype.
10836 (emit_label_in_global_context_p): Remove.
10837 * lto-symtab.c: Include ipa-utils.h
10838 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10840 2013-08-31 Jan Hubicka <jh@suse.cz>
10842 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10844 2013-08-31 Jan Hubicka <jh@suse.cz>
10846 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10848 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10850 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10851 "cmp" RTX before signed_comparison_operator check to account
10852 for "code" changes.
10854 2013-08-30 Jan Hubicka <jh@suse.cz>
10856 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10857 (detect_type_change_1): Rename to ...
10858 (detect_type_change): ... this one; early return on non-polymorphic
10860 (detect_type_change_ssa): Add comp_type parameter; update
10861 use of detect_type_change.
10862 (compute_complex_assign_jump_func): Add param_type parameter;
10863 update use of detect_type_change_ssa.
10864 (compute_complex_ancestor_jump_func): Likewise.
10865 (ipa_get_callee_param_type): New function.
10866 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10867 update calls to the jump function computation functions.
10869 2013-08-30 Teresa Johnson <tejohnson@google.com>
10870 Steven Bosscher <steven@gcc.gnu.org>
10872 * cfgrtl.c (fixup_new_cold_bb): New routine.
10873 (commit_edge_insertions): Invoke fixup_partitions.
10874 (find_partition_fixes): New routine.
10875 (fixup_partitions): Ditto.
10876 (verify_hot_cold_block_grouping): Update comments.
10877 (rtl_verify_edges): Invoke find_partition_fixes.
10878 (rtl_verify_bb_pointers): Update comments.
10879 (rtl_verify_bb_layout): Ditto.
10880 * basic-block.h (probably_never_executed_edge_p): Declare.
10881 (fixup_partitions): Ditto.
10882 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10883 * bb-reorder.c (sanitize_hot_paths): New function.
10884 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10885 sanitize_hot_paths.
10886 * predict.c (probably_never_executed_edge_p): New routine.
10887 * cfg.c (check_bb_profile): Add partition insanity warnings.
10889 2013-08-30 Meador Inge <meadori@codesourcery.com>
10891 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10893 2013-08-30 Marek Polacek <polacek@redhat.com>
10895 * Makefile.in (ubsan.o): Add.
10896 (c-family/c-ubsan.o): Add.
10897 (builtins.o): Add ubsan.h dependency.
10898 * ubsan.h: New file.
10899 * ubsan.c: New file.
10900 * common.opt: Add -fsanitize=undefined option.
10901 (flag_sanitize): Add variable.
10902 (fsanitize=): Add option. Add Driver.
10903 (fsanitize=thread): Remove option.
10904 (fsanitize=address): Likewise.
10905 (static-libubsan): New option.
10906 * doc/invoke.texi: Document the new flag and -static-libubsan.
10907 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10908 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10909 * builtin-attrs.def (ATTR_COLD): Define.
10910 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10911 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10912 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10913 * flag-types.h (sanitize_code): New enum.
10914 * opts.c (common_handle_option): Parse command line arguments
10915 of -fsanitize=. Add -fsanitize=unreachable option.
10916 * varasm.c (get_variable_section): Adjust.
10917 (assemble_noswitch_variable): Likewise.
10918 (assemble_variable): Likewise.
10919 (output_constant_def_contents): Likewise.
10920 (categorize_decl_for_section): Likewise.
10921 (place_block_symbol): Likewise.
10922 (output_object_block): Likewise.
10923 * builtins.def: Likewise.
10924 * toplev.c (compile_file): Likewise.
10925 (process_options): Likewise.
10926 * cppbuiltin.c: Likewise.
10927 * tsan.c (tsan_pass): Likewise.
10928 (tsan_gate): Likewise.
10929 (tsan_gate_O0): Likewise.
10930 * cfgexpand.c (partition_stack_vars): Likewise.
10931 (expand_stack_vars): Likewise.
10932 (defer_stack_allocation): Likewise.
10933 (expand_used_vars): Likewise.
10934 * cfgcleanup.c (old_insns_match_p): Likewise.
10935 * asan.c (asan_finish_file): Likewise.
10936 (asan_instrument): Likewise.
10937 (gate_asan): Likewise.
10938 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10939 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10940 (asan_global_struct): Use pointer_sized_int_node instead
10941 calling build_nonstandard_integer_type.
10942 (initialize_sanitizer_builtins): Likewise.
10943 (asan_finish_file): Likewise.
10944 * gcc.c: Document %{%:function(args):X}.
10945 (static_spec_functions): Add sanitize.
10946 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10947 store funcval != NULL there.
10948 (do_spec_1): Adjust handle_spec_function caller.
10949 (handle_braces): Allow %:function(args) as condition.
10950 (sanitize_spec_function): New function.
10951 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10952 (LIBUBSAN_SPEC): Likewise.
10953 (LIBUBSAN_EARLY_SPEC): Likewise.
10954 (SANITIZER_SPEC): Handle libubsan.
10955 (SANITIZER_EARLY_SPEC): Likewise.
10956 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10957 instead of fsanitize=address.
10958 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10959 instead of fsanitize=address*.
10960 * builtins.c: Include ubsan.h.
10961 (fold_builtin_0): Instrument __builtin_unreachable.
10962 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10963 instead of flag_asan.
10964 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10965 (pointer_sized_int_node): Define.
10966 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10968 2013-08-30 Mike Stump <mikestump@comcast.net>
10970 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10973 2013-08-29 Jan Hubicka <jh@suse.cz>
10975 * cgraph.c (cgraph_function_body_availability): Handle weakref
10977 * passes.def: Remove pass_fixup_cfg.
10978 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10979 track when we need to remove functions.
10980 (gate_ipa_inline): Execute inlining always; add comment why.
10981 (pass_data_ipa_inline): Remove TODO_remove_functions.
10982 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10983 do not produce summaries.
10984 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10985 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10986 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10987 constant pool and vtable.
10989 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10991 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10992 New arm_neon.h's internal macros to specify 64-bit constants.
10993 Avoid using stdint.h's macros.
10995 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10997 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10998 before applying REG_P.
11000 2013-08-30 Jakub Jelinek <jakub@redhat.com>
11002 PR tree-optimization/58277
11003 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
11004 after seeing too many stmts with vdef in between dombb and current
11005 bb, invalidate everything.
11007 2013-08-30 Richard Biener <rguenther@suse.de>
11009 * fold-const.c (fold_single_bit_test): Fix overflow test.
11011 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
11013 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
11014 and which can live in a register, always retrieve the value on entry.
11015 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
11016 passed by invisible reference specially.
11017 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
11018 (vt_add_function_parameter): Correctly deal with a parameter passed by
11019 invisible reference.
11021 2013-08-30 Jan Hubicka <jh@suse.cz>
11023 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
11025 2013-08-30 Richard Biener <rguenther@suse.de>
11027 PR tree-optimization/58228
11028 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
11029 allow invariant loads in nested loop vectorization.
11031 2013-08-30 Richard Biener <rguenther@suse.de>
11033 PR tree-optimization/58223
11034 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
11035 (has_anti_or_output_dependence): ... this and adjust to also
11036 look for output dependences.
11037 (mark_nodes_having_upstream_mem_writes): Adjust.
11038 (rdg_flag_uses): Likewise.
11040 2013-08-30 Richard Biener <rguenther@suse.de>
11042 PR tree-optimization/58010
11043 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
11044 assert that we have a loop-closed PHI.
11046 2013-08-29 Jan Hubicka <jh@suse.cz>
11048 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
11049 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
11050 * lto-section-in.c (lto_free_function_in_decl_state): New function.
11051 (lto_free_function_in_decl_state_for_node): New function.
11053 2013-08-29 Xinliang David Li <davidxl@google.com>
11055 * loop-unroll.c (report_unroll_peel): Minor message change.
11056 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
11057 Emit alignment peeling message with default -fopt-info.
11058 (vect_loop_versioning): Emit loop version info message.
11059 * tree-vectorizer.c (vectorize_loops): Minor message change.
11060 (execute_vect_slp): Ditto.
11062 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
11064 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
11065 of the clone from the DECL_NAME of the original function.
11067 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
11069 * passes.c (register_pass): Add overload.
11070 * tree-pass.h (register_pass): Forward declare it. Add comment.
11072 2013-08-29 Jan Hubicka <jh@suse.cz>
11074 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
11075 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
11076 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
11077 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11078 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
11079 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
11080 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11081 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
11083 2013-08-29 Teresa Johnson <tejohnson@google.com>
11085 * dumpfile.c (dump_loc): Output column number.
11086 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
11087 * doc/invoke.texi: Document optall -fopt-info flag.
11088 * profile.c (read_profile_edge_counts): Use new dump framework.
11089 (compute_branch_probabilities): Ditto.
11090 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
11091 when pass not in any opt group.
11092 * pass_manager.h (pass_manager::get_pass_profile): New method.
11093 * value-prof.c (check_counter): Use new dump framework.
11094 (check_ic_target): Ditto.
11095 * coverage.c (get_coverage_counts): Ditto.
11096 (coverage_init): Setup new dump framework.
11098 2013-08-29 Richard Biener <rguenther@suse.de>
11100 PR tree-optimization/58246
11101 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
11102 handle the dominance check inside a basic-block.
11104 2013-08-29 Richard Biener <rguenther@suse.de>
11106 PR middle-end/57287
11107 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
11108 of default defs that appear in abnormal PHI nodes.
11110 2013-08-29 Richard Biener <rguenther@suse.de>
11112 PR tree-optimization/57685
11113 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
11114 single-use operands to avoid exponential complexity.
11116 2013-08-28 Dehao Chen <dehao@google.com>
11118 * ipa-inline.c (edge_badness): Fix integer underflow.
11120 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
11122 * gtm-builtins.def (_ITM_free): Declare leaf.
11124 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11127 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
11128 (*tls_local_dynamic_base_64_largepic): Likewise.
11129 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
11130 Remove predicate from call operand.
11131 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
11132 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
11134 2013-08-28 Jeff Law <law@redhat.com>
11136 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
11137 checks for the number of predecessors and successors allowed.
11138 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
11139 which require copying a joiner block if there is a request which
11140 is a subpath that requires no joiner block copying.
11142 2013-08-28 Jan Hubicka <jh@suse.cz>
11144 * lto-streamer-out.c (DFS_write_tree_body): Drop
11145 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
11146 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
11147 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
11148 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
11149 Do not read DECL_ERROR_ISSUED.
11150 (unpack_ts_decl_with_vis_value_fields): Do not read
11152 (lto_input_ts_binfo_tree_pointers): Do not read
11153 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
11154 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
11155 write DECL_ERROR_ISSUED..
11156 (pack_ts_decl_with_vis_value_fields): Do not write
11158 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
11159 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
11160 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
11161 * tree.h (tree_decl_common): Update comment.
11162 (DECL_ERROR_ISSUED): Remove.
11164 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11166 PR middle-end/58257
11167 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
11169 2013-08-28 Jan Hubicka <jh@suse.cz>
11171 * builtins.def (free): Declare leaf.
11173 2013-08-27 David Malcolm <dmalcolm@redhat.com>
11175 * gdbhooks.py: New.
11176 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
11177 * configure: Regenerate.
11179 2013-08-27 Martin Jambor <mjambor@suse.cz>
11181 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
11182 (ipa_ancestor_jf_data): Likewise.
11183 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
11184 (ipa_get_jf_pass_through_type_preserved): New function.
11185 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
11186 (ipa_get_jf_ancestor_type_preserved): New function.
11187 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
11188 (ipa_get_jf_ancestor_result): Likewise.
11189 (propagate_vals_accross_pass_through): Use
11190 ipa_get_jf_pass_through_result to do all the value mappings.
11191 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
11192 type_preserved flag.
11193 (ipa_set_jf_cst_copy): New function.
11194 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
11195 (ipa_set_jf_arith_pass_through): Likewise.
11196 (ipa_set_ancestor_jf): Likewise.
11197 (compute_complex_assign_jump_func): Set type_preserved instead of
11199 (ipa_compute_jump_functions_for_edge): Likewise.
11200 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
11201 (update_jump_functions_after_inlining): Update type_preserved.
11202 Explicitely create jump functions when combining one with pass_through.
11203 (ipa_write_jump_function): Stream the type_preserved flags.
11204 (ipa_read_jump_function): Likewise.
11206 2013-08-27 Jakub Jelinek <jakub@redhat.com>
11207 Aldy Hernandez <aldyh@redhat.com>
11209 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
11210 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
11211 * function.h (struct function): Add has_force_vect_loops and
11213 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
11214 * gimple.c (gimple_build_omp_critical): Add KIND argument and
11216 * gimple.def: Update CLAUSES comments.
11217 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
11218 (gimple_build_omp_for): Add argument to prototype.
11219 (gimple_omp_for_kind): New.
11220 (gimple_omp_for_set_kind): New.
11221 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
11222 GOVD_DATA_SHARE_CLASS.
11223 (enum omp_region_type): Add ORT_SIMD.
11224 (gimple_add_tmp_var): Handle ORT_SIMD.
11225 (gimplify_var_or_parm_decl): Same.
11226 (is_gimple_stmt): Same.
11227 (omp_firstprivatize_variable): Same.
11228 (omp_add_variable): Only use splay_tree_insert if lookup failed.
11229 (omp_notice_variable): Handle ORT_SIMD.
11230 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
11231 (omp_check_private): Handle ORT_SIMD.
11232 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11233 OMP_CLAUSE_SAFELEN.
11234 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
11235 Handle OMP_CLAUSE_LASTPRIVATE.
11236 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11237 OMP_CLAUSE_SAFELEN.
11238 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
11239 (gimplify_expr): Handle OMP_SIMD.
11240 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
11241 (expand_GOMP_SIMD_VF): New.
11242 (expand_GOMP_SIMD_LAST_LANE): New.
11243 * internal-fn.def (GOMP_SIMD_LANE): New.
11244 (GOMP_SIMD_VF): New.
11245 (GOMP_SIMD_LAST_LANE): New.
11246 * omp-low.c: Include target.h.
11247 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
11248 OMP_CLAUSE_SAFELEN.
11249 (check_omp_nesting_restrictions): Same.
11251 (lower_rec_simd_input_clauses): New.
11252 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
11254 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
11255 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
11256 (expand_omp_build_assign): New.
11257 (expand_omp_for_init_counts): New.
11258 (expand_omp_for_init_vars): New.
11259 (extract_omp_for_update_vars): New.
11260 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
11261 and rewrite accordingly.
11262 (expand_omp_simd): New.
11263 (expand_omp_for): Use expand_omp_simd.
11264 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
11265 (lower_omp_for): Do not lower the body.
11266 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
11267 in their own loops.
11268 * tree-flow.h (find_omp_clause): Remove prototype.
11269 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
11270 forcing vectorization of the loop, or if flag_tree_vectorize.
11271 (gate_tree_if_conversion): Similarly.
11272 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
11273 gimple_build_omp_for.
11274 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
11275 * tree-parloops (create_parallel_loop): Pass kind argument to
11276 gimple_build_omp_for.
11277 * tree-pretty-print.c (dump_omp_clause): Add cases for
11278 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
11279 OMP_CLAUSE__SIMDUID_.
11280 (dump_generic_node): Handle OMP_SIMD.
11281 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
11282 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
11283 unroll OMP_SIMD loops here.
11284 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
11285 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
11287 (vect_analyze_data_refs): Handle simd loops.
11288 * tree-vect-loop.c (vectorizable_live_operation): Handle
11290 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
11291 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
11292 (vectorizable_load): Same.
11293 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
11294 (struct simduid_to_vf): New.
11295 (simduid_to_vf::hash): New.
11296 (simduid_to-vf::equal): New.
11297 (struct simd_array_to_simduid): New.
11298 (simd_array_to_simduid::hash): New.
11299 (simd_array_to_simduid::equal): New.
11300 (adjust_simduid_builtins): New.
11301 (struct note_simd_array_uses_struct): New.
11302 (note_simd_array_uses_cb): New.
11303 (note_simd_array_uses): New.
11304 (vectorize_loops): Handle simd hints and adjust simd builtins
11306 * tree-vectorizer.h (struct _stmt_vec_info): Add
11307 simd_lane_access_p field.
11308 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
11309 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
11310 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
11311 (omp_clause_code_name): Same.
11312 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
11313 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
11314 * tree.def (OMP_SIMD): New entry.
11315 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
11316 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
11317 (OMP_CLAUSE_DECL): Adjust range for new clauses.
11318 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
11319 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
11320 (OMP_CLAUSE_LINEAR_STEP): New.
11321 (OMP_CLAUSE_SAFELEN_EXPR): New.
11322 (OMP_CLAUSE__SIMDUID__DECL): New.
11323 (find_omp_clause): New prototype.
11325 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
11327 * config/i386/driver-i386.c (host_detect_local_cpu): Update
11328 Haswell processor detection.
11330 2013-08-27 Christian Widmer <shadow@umbrox.de>
11333 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
11334 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
11335 AVX2 capable processors.
11337 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
11339 * config/aarch64/arm_neon.h: Replace all inline asm implementations
11340 of vget_low_* with implementations in terms of other intrinsics.
11342 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
11344 PR middle-end/57219
11345 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
11346 values to -1, 0 and 1.
11348 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
11350 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
11351 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
11352 (<optab><mode>3_insn): Remove.
11353 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
11354 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
11355 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
11356 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
11357 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
11358 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
11359 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
11360 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
11361 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
11362 (ror<mode>3_insn): Likewise.
11363 * config/aarch64/predicates.md (aarch64_simd_register): New.
11365 2013-08-27 Richard Biener <rguenther@suse.de>
11367 PR tree-optimization/57521
11368 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
11369 one edge is non-critical.
11370 (find_phi_replacement_condition): Make sure to use a non-critical
11371 edge. Cleanup and remove old bug workarounds.
11372 (bb_postdominates_preds): Remove.
11373 (if_convertible_loop_p_1): Do not compute post-dominators.
11374 (combine_blocks): Do not free post-dominators.
11375 (main_tree_if_conversion): Likewise.
11376 (pass_data_if_conversion): Add TODO_verify_ssa.
11378 2013-08-27 DJ Delorie <dj@redhat.com>
11380 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
11382 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
11384 * function.c (assign_parm_find_data_types): Set passed_mode and
11385 nominal_mode to the TYPE_MODE of nominal_type for the built
11386 pointer type in case of the struct-pass-by-reference.
11388 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11390 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
11391 (UINT16_TYPE): Change default to "unsigned int".
11393 * config/avr/avr.opt (mfract-convert-truncate): New option.
11394 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
11395 is set, round negative fractional integers according to n1169
11396 when converting to integer types.
11398 2013-08-26 Jan Hubicka <jh@suse.cz>
11400 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
11401 methods can not be called indirectly when their address is not taken.
11403 2013-08-26 Jan Hubicka <jh@suse.cz>
11405 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
11408 2013-08-26 Jan Hubicka <jh@suse.cz>
11410 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11413 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11415 * reload.c (find_valid_class): Allow classes that do not include
11416 FIRST_PSEUDO_REGISTER - 1.
11418 2013-08-26 Jan Hubicka <jh@suse.cz>
11420 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11421 fix edge count/frequency when speculation failed; fix type check
11422 for the direct call.
11424 2013-08-26 Jan Hubicka <jh@suse.cz>
11426 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11428 2013-08-26 Jan Hubicka <jh@suse.cz>
11430 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11431 to be changed by edge redirection.
11433 2013-08-26 Jan Hubicka <jh@suse.cz>
11435 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11436 formating; add sanity check.
11437 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11438 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11439 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11440 (resolve_noninline_speculation): Update callee keys, too.
11442 2013-08-26 Jan Hubicka <jh@suse.cz>
11444 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11445 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11447 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11449 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11452 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11454 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11455 (x86_64_elf_select_section): Ditto.
11456 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11458 (ix86_pass_by_reference): Ditto.
11459 (output_set_got): Ditto.
11460 (ix86_unary_operator_ok): Ditto.
11461 (ix86_expand_builtin): Ditto.
11463 2013-08-23 Jan Hubicka <jh@suse.cz>
11465 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11467 2013-08-23 Jan Hubicka <jh@suse.cz>
11469 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11470 (tree_decl_with_vis): Add FINAL field.
11472 2013-08-23 Jeff Law <law@redhat.com>
11474 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11475 the debugging dump when the expression is fully redundant.
11477 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11479 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11480 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11481 * pretty-print.c (pp_formatted_text_data): Likewise.
11482 (pp_write_text_to_stream): Likewise.
11483 (pp_write_text_as_dot_label_to_stream): Likewise.
11484 (pp_append_r): Likewise.
11485 (pp_format): Likewise.
11486 (pp_flush): Likewise.
11487 (pp_clear_output_area): Likewise.
11488 (pp_append_text): Likewise.
11489 (pp_formatted_text): Likewise.
11490 (pp_remaining_character_count_for_line): Likewise.
11491 (pp_newline): Likewise.
11492 (pp_character): Likewise.
11493 (output_buffer::~output_buffer): Define.
11494 (pretty_printer::~pretty_printer): Destruct output buffer.
11495 * pretty-print.h (output_buffer::~output_buffer): Declare.
11496 (pretty_printer::~pretty_printer): Declare virtual.
11498 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11501 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11502 HOST_WIDE_INT_M1U): New macros.
11503 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11504 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11505 unsigned -1 for lshift.
11506 * cse.c (cse_insn): Likewise.
11507 * double-int.c (rshift_double, lshift_double): Likewise.
11508 * builtins.c (fold_builtin_bitop): Likewise.
11509 * combine.c (force_to_mode): Likewise.
11510 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11511 * simplify-rtx.c (simplify_const_unary_operation,
11512 simplify_const_binary_operation): Likewise.
11513 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11514 check_va_list_escapes): Likewise.
11515 * rtlanal.c (nonzero_bits1): Likewise.
11516 * expmed.c (expand_smod_pow2): Likewise.
11517 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11519 2013-08-23 Jan Hubicka <jh@suse.cz>
11521 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11522 as having address taken.
11524 2013-08-23 Jan Hubicka <jh@suse.cz>
11526 * ipa-utils.h (method_class_type): Declare.
11527 * ipa-devirt.c (method_class_type): Export.
11529 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11530 do not walk base classes of anonymous types.
11532 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11534 PR rtl-optimization/58220
11535 PR regression/58221
11536 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11537 handle SEQUENCE insns properly.
11539 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11541 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11543 (pp_newline_and_indent): Likewise.
11544 (pp_separate_with): Likewise.
11545 * pretty-print.c (pp_newline_and_flush): Define.
11546 (pp_newline_and_indent): Likewise.
11547 (pp_separate_with): Likewise.
11549 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11552 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11553 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
11555 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
11557 * config/i386/predicates.md (ext_sse_reg_operand): New.
11558 * config/i386/i386.md (*movti_internal): Use
11559 predicate to determine if EVEX is needed.
11560 (*movsi_internal): Ditto.
11561 (*movdf_internal): Ditto.
11562 (*movsf_internal): Ditto.
11563 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
11565 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11567 PR tree-optimization/58209
11568 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
11569 (find_tail_calls): Give up for pointer result types if m is non-NULL.
11570 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
11571 emit POINTER_PLUS_EXPR.
11572 (create_tailcall_accumulator): For pointer result type accumulate in
11575 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
11577 * configure.ac: Add backslashes missing from the last change.
11578 * configure: Regenerate.
11580 2013-08-22 Jan Hubicka <jh@susue.cz>
11582 * ipa.c (function_and_variable_visibility): First remember function
11583 was global and then make it local.
11585 2013-08-22 Julian Brown <julian@codesourcery.com>
11587 * configure.ac: Add aarch64 to list of arches which use "nop" in
11589 * configure: Regenerate.
11591 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11593 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
11594 gnu_libc_has_function.
11595 * config/s390/tpf.h: Likewise.
11597 2013-08-22 Jan Hubicka <jh@susue.cz>
11599 * timevar.c (validate_phases): Add cast.
11601 2013-08-22 Jan Hubicka <jh@susue.cz>
11603 * timevar.c (validate_phases): Use size_t for memory.
11604 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11606 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11608 * pretty-print.h (output_buffer::output_buffer): Declare.
11609 (pretty_printer::pretty_printer): Likewise.
11610 (pp_construct): Remove.
11611 * pretty-print.c (output_buffer::output_buffer): Define.
11612 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11613 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11614 (print_gimple_expr): Likewise.
11615 (print_gimple_seq): Likewise.
11616 (gimple_dump_bb): Likewise.
11617 * sched-vis.c (dump_value_slim): Likewise.
11618 (dump_insn_slim): Likewise.
11619 (dump_rtl_slim): Likewise.
11620 (str_pattern_slim): Likewise.
11621 * tree-mudflap.c (mf_varname_tree): Likewise.
11622 * graph.c (print_graph_cfg): Likewise.
11623 (start_graph_dump): Likewise.
11624 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11626 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11627 pretty printer initialization.
11628 * coretypes.h (diagnostic_context): Remove superflous type alias
11630 (pretty_printer): Likewise. Declare directly as a class.
11631 (pretty_print_info): Remove declaration as class.
11632 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11633 and pp_clear_output_area.
11634 (asan_add_global): Likewise.
11636 2013-08-22 Jan Hubicka <jh@suse.cz>
11638 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11639 * ipa-utils.h (update_type_inheritance_graph): Declare.
11640 (possible_polymorphic_call_target_p): Declare.
11641 (possible_polymorphic_call_target_p): New.
11642 * ipa-devirt.c: Update toplevel comments.
11643 (cached_polymorphic_call_targets): Move up.
11644 (odr_type_d): Move ID down.
11645 (polymorphic_type_binfo_p): Update comment.
11646 (odr_hasher::remove): Likewise;
11647 (get_odr_type): Set anonymous_namespace.
11648 (dump_odr_type): Dump it.
11649 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11650 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11651 (record_binfo): Add comment.
11652 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11654 (devirt_node_removal_hook): Do not iCE when cache is freed.
11655 (possible_polymorphic_call_target_p): New predicate.
11656 (update_type_inheritance_graph): New function.
11658 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11659 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11660 Sergey Lega <sergey.s.lega@intel.com>
11661 Anna Tikhonova <anna.tikhonova@intel.com>
11662 Ilya Tocar <ilya.tocar@intel.com>
11663 Andrey Turetskiy <andrey.turetskiy@intel.com>
11664 Ilya Verbin <ilya.verbin@intel.com>
11665 Kirill Yukhin <kirill.yukhin@intel.com>
11666 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11668 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11669 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11670 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11671 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11672 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11673 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11674 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11675 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11676 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11677 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11678 OPT_mavx512pf, OPT_mavx512er cases.
11679 * config/i386/constraints.md (v): New constraint.
11680 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11681 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11682 (bit_AVX512CD): New.
11683 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11684 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11685 * config/i386/i386-c.c (ix86_target_macros_internal):
11686 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11688 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11689 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11690 * config/i386/i386.c (regclass_map, dbx_register_map)
11691 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11692 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11693 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11694 -mavx512pf options.
11695 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11696 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11697 -mavx512pf options. Fix formatting.
11698 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11699 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11700 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11701 -mavx512cd, -mavx512pf options.
11702 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11703 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11704 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11706 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11707 (ix86_set_reg_reg_cost): Ditto.
11708 (x86_order_regs_for_local_alloc): Ditto.
11709 (MAX_VECT_LEN): Extend to 64-byte.
11710 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11711 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11712 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11713 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11714 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11715 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11716 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11717 (SSE_REG_MODE_P): Support new modes.
11718 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11719 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11720 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11721 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11722 (REG_CLASS_CONTENTS): Add new registers.
11723 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11724 (EXT_REX_SSE_REGNO_P): New.
11725 (HI_REGISTER_NAMES): Add new registers.
11726 * config/i386/i386.md: Define constants for new registers.
11727 (mode): Add new 512-bit modes.
11728 (prefix): Support evex prefix.
11729 (isa): Support avx512f, noavx512f, fma_avx512f.
11730 (ssemodesuffix): Add new 512-bit modes.
11732 (*movxi_internal_avx512f): Ditto.
11733 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11735 (*movsi_internal): Likewise.
11736 (*movdf_internal): Likewise.
11737 (*movsf_internal): Likewise.
11738 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11739 (<code><mode>3): Likewise.
11740 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11742 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11743 with the new constraint "v".
11744 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11746 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11747 with the new constraint "v".
11748 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11749 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11750 (*<plusminus_insn><mode>3): Likewise.
11751 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11752 (*mul<mode>3): Likewise.
11753 (<sse>_vmmul<mode>3): Likewise.
11754 (<sse>_div<mode>3): Likewise.
11755 (<sse>_vmdiv<mode>3): Likewise.
11756 (<sse>_sqrt<mode>2): Likewise.
11757 (<sse>_vmsqrt<mode>2): Likewise.
11758 (*<code><mode>3_finite): Likewise.
11759 (*<code><mode>3) <smaxmin>: Likewise.
11760 (<sse>_vm<code><mode>3): Likewise.
11761 (*<code><mode>3) <any_logic>: Likewise.
11762 (*fma_fmadd_<mode>): Likewise.
11763 (*fma_fmsub_<mode>): Likewise.
11764 (*fma_fnmadd_<mode>): Likewise.
11765 (*fma_fnmsub_<mode>): Likewise.
11766 (*fma_fmaddsub_<mode>): Likewise.
11767 (*fma_fmsubadd_<mode>): Likewise.
11768 (*fmai_fmadd_<mode>): Likewise.
11769 (*fmai_fmsub_<mode>): Likewise.
11770 (*fmai_fnmadd_<mode>): Likewise.
11771 (*fmai_fnmsub_<mode>): Likewise.
11772 (sse_cvtsi2ss): Likewise.
11773 (sse_cvtsi2ssq): Likewise.
11774 (sse_cvtss2si): Likewise.
11775 (sse_cvtss2si_2): Likewise.
11776 (sse_cvtss2siq): Likewise.
11777 (sse_cvtss2siq_2): Likewise.
11778 (sse_cvttss2si): Likewise.
11779 (sse_cvtss2siq_2): Likewise.
11780 (float<sseintvecmodelower><mode>2): Likewise.
11781 (sse2_cvtsd2si_2): Likewise.
11782 (sse2_cvtsd2siq_2): Likewise.
11783 (*<plusminus_insn><mode>3): Likewise.
11784 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11785 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11786 (ashr<mode>3): Likewise.
11787 (<shift_insn><mode>3): Likewise.
11788 (avx2_<code><mode>3): Likewise.
11789 (*avx2_<code><mode>3): Likewise.
11790 (*andnot<mode>3): Likewise.
11791 (*<code><mode>3) <any_logic>: Likewise.
11792 (abs<mode>2): Likewise.
11793 (avx2_permvar<mode>): Likewise.
11794 (avx2_perm<mode>_1): Likewise.
11795 (*avx_vpermilp<mode>): Likewise.
11796 (avx_vpermilvar<mode>3): Likewise.
11797 (avx2_ashrv<mode>): Likewise.
11798 (avx2_<shift_insn>v<mode>): Likewise.
11799 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11801 * doc/rtl.texi: Document XImode.
11803 2013-08-21 Jeff Law <law@redhat.com>
11805 * tree-flow.h (register_jump_thread): Pass vector of edges
11806 instead of each important edge.
11807 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11808 thread path into a vector and pass that to register_jump_thread.
11809 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11810 passed in edge vector to the current 3-edge form.
11813 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11815 * dce.c (fini_dce): Call df_analyze again just in case
11816 delete_unmarked_insns removed anything.
11818 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11820 * reload.h (struct reg_equivs): Rename to ..
11821 (struct reg_equivs_s): .. this.
11823 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11825 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11827 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11829 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11831 2013-08-21 Jeff Law <law@redhat.com>
11833 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11834 simplify assignments too. If the RHS collapses to a singleton
11835 range, then return the value for the range.
11837 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11839 * config/i386/sse.md (V16): Rename to...
11841 (mov<mode>): Update iterator name.
11842 (*mov<mode>_internal): Ditto.
11843 (push<mode>1): Ditto.
11844 (movmisalign<mode>): Ditto.
11846 2013-08-20 Jan Hubicka <jh@suse.cz>
11849 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11850 entry for direct edges.
11851 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11853 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11855 Revert my last two changes, r201865 and r201864:
11858 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11860 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11861 instances can own GC refs.
11863 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11864 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11865 (gcc::context::gt_pch_nx): Likewise.
11866 (gcc::context::gt_pch_nx): Likewise.
11867 * ggc.h (gt_ggc_mx <T>): New.
11868 (gt_pch_nx_with_op <T>): New.
11869 (gt_pch_nx <T>): New.
11870 * passes.c (opt_pass::gt_ggc_mx): New.
11871 (opt_pass::gt_pch_nx): New.
11872 (opt_pass::gt_pch_nx_with_op): New.
11873 (pass_manager::gt_ggc_mx): New.
11874 (pass_manager::gt_pch_nx): New.
11875 (pass_manager::gt_pch_nx_with_op): New.
11876 (pass_manager::operator new): Use
11877 ggc_internal_cleared_alloc_stat rather than xcalloc.
11878 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11879 (pass_manager::gt_ggc_mx): New.
11880 (pass_manager::gt_pch_nx): New.
11881 (pass_manager::gt_pch_nx_with_op): New.
11882 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11883 (opt_pass::operator new): New.
11884 (opt_pass::gt_ggc_mx): New.
11885 (opt_pass::gt_pch_nx): New.
11886 (opt_pass::gt_pch_nx_with_op): New.
11889 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11891 * Makefile.in (GTFILES): Add context.h.
11892 * context.c (gcc::context::operator new): New.
11893 (gcc::context::gt_ggc_mx): New.
11894 (gcc::context::gt_pch_nx): New.
11895 (gcc::context::gt_pch_nx): New.
11896 * context.h (gcc::context): Add GTY((user)) marking.
11897 (gcc::context::operator new): New.
11898 (gcc::context::gt_ggc_mx): New.
11899 (gcc::context::gt_pch_nx): New.
11900 (gcc::context::gt_pch_nx): New.
11901 (g): Add GTY marking.
11902 (gt_ggc_mx (gcc::context *)): New.
11903 (gt_pch_nx (gcc::context *)): New.
11904 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11905 void *cookie)): New.
11906 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11908 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11910 * dce.c (fini_dce): Call df_analyze again just in case
11911 delete_unmarked_insns removed anything.
11913 2013-08-20 Teresa Johnson <tejohnson@google.com>
11915 PR rtl-optimizations/57451
11916 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11917 from crossing split section boundaries.
11919 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11921 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11922 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11923 with MULTLIB_DEFAULTS.
11925 2013-08-20 Nick Clifton <nickc@redhat.com>
11927 * target.def (narrow_volatile_bitfield): Note that the default
11928 value is false, not !TARGET_STRICT_ALIGN.
11929 * doc/tm.texi: Regenerate.
11931 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11933 Fix LIB_SPEC for systems without libpthread.
11935 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11936 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11938 * config/i386/linux-common.h: Likewise.
11939 * config/mips/linux-common.h: Likewise.
11941 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11943 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11946 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11948 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11949 instances can own GC refs.
11951 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11952 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11953 (gcc::context::gt_pch_nx): Likewise.
11954 (gcc::context::gt_pch_nx): Likewise.
11955 * ggc.h (gt_ggc_mx <T>): New.
11956 (gt_pch_nx_with_op <T>): New.
11957 (gt_pch_nx <T>): New.
11958 * passes.c (opt_pass::gt_ggc_mx): New.
11959 (opt_pass::gt_pch_nx): New.
11960 (opt_pass::gt_pch_nx_with_op): New.
11961 (pass_manager::gt_ggc_mx): New.
11962 (pass_manager::gt_pch_nx): New.
11963 (pass_manager::gt_pch_nx_with_op): New.
11964 (pass_manager::operator new): Use
11965 ggc_internal_cleared_alloc_stat rather than xcalloc.
11966 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11967 (pass_manager::gt_ggc_mx): New.
11968 (pass_manager::gt_pch_nx): New.
11969 (pass_manager::gt_pch_nx_with_op): New.
11970 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11971 (opt_pass::operator new): New.
11972 (opt_pass::gt_ggc_mx): New.
11973 (opt_pass::gt_pch_nx): New.
11974 (opt_pass::gt_pch_nx_with_op): New.
11976 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11978 * Makefile.in (GTFILES): Add context.h.
11979 * context.c (gcc::context::operator new): New.
11980 (gcc::context::gt_ggc_mx): New.
11981 (gcc::context::gt_pch_nx): New.
11982 (gcc::context::gt_pch_nx): New.
11983 * context.h (gcc::context): Add GTY((user)) marking.
11984 (gcc::context::operator new): New.
11985 (gcc::context::gt_ggc_mx): New.
11986 (gcc::context::gt_pch_nx): New.
11987 (gcc::context::gt_pch_nx): New.
11988 (g): Add GTY marking.
11989 (gt_ggc_mx (gcc::context *)): New.
11990 (gt_pch_nx (gcc::context *)): New.
11991 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11992 void *cookie)): New.
11993 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11995 2013-08-20 Alan Modra <amodra@gmail.com>
11998 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11999 (rs6000_emit_epilogue): Likewise.
12001 2013-08-19 Dehao Chen <dehao@google.com>
12003 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
12005 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
12006 Jakub Jelinek <jakub@redhat.com>
12008 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
12009 (BUILT_IN_FABSD64): Likewise.
12010 (BUILT_IN_FABSD128): Likewise.
12011 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
12012 (fold_builtin_1): Likewise.
12013 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
12014 destination and source operands.
12015 (*abstd2_fpr): Likewise.
12016 (*nabstd2_fpr): Likewise.
12018 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
12020 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
12023 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
12025 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
12028 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
12030 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
12031 * builtins.c (default_libc_has_function): New.
12032 (gnu_libc_has_function): Ditto.
12033 (no_c99_libc_has_function): Ditto.
12034 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
12035 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
12036 (fold_builtin_sincos): Likewise.
12037 (fold_builtin_cexp): Likewise.
12038 * builtins.def (DEF_C94_BUILTIN): Likewise.
12039 (DEF_C99_BUILTIN): Likewise.
12040 (DEF_C99_C90RES_BUILTIN): Likewise.
12041 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
12042 definitions to using this define.
12043 * config/darwin-protos.h (darwin_libc_has_function): New.
12044 * config/darwin.c (darwin_libc_has_function): Ditto.
12045 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
12046 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
12047 * config/darwin.h: Ditto.
12048 * config/elfos.h: Ditto.
12049 * config/freebsd.h: Ditto.
12050 * config/i386/cygming.h: Ditto.
12051 * config/i386/djgpp.h: Ditto.
12052 * config/i386/i386-interix.h: Ditto.
12053 * config/microblaze/microblaze.h: Ditto.
12054 * config/mmix/mmix.h: Ditto.
12055 * config/gnu-user.h: Ditto.
12056 * config/ia64/hpux.h: Ditto.
12057 * config/pa/pa-hpux.h: Ditto.
12058 * config/pdp11/pdp11.h: Ditto.
12059 * config/picochip/picochip.h: Ditto.
12060 * config/linux.h: Ditto.
12061 * config/netbsd.h: Ditto.
12062 * config/openbsd.h: Ditto.
12063 * config/rs6000/aix43.h: Ditto.
12064 * config/rs6000/aix51.h: Ditto.
12065 * config/rs6000/aix52.h: Ditto.
12066 * config/rs6000/aix53.h: Ditto.
12067 * config/rs6000/aix61.h: Ditto.
12068 * config/rs6000/darwin.h: Ditto.
12069 * config/rs6000/linux.h: Ditto.
12070 * config/rs6000/linux64.h: Ditto.
12071 * config/s390/tpf.h: Ditto.
12072 * config/sol2-10.h: Ditto.
12073 * config/sol2.h: Ditto.
12074 * config/vms/vms.h: Ditto.
12075 * config/vxworks.h: Ditto.
12076 * config/linux-android.c (linux_android_libc_has_function):
12077 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
12078 * config/linux-protos.h (linux_android_libc_has_function):
12080 * config/i386/i386.c (ix86_libc_has_function): New.
12081 * config/i386/i386-protos.h
12082 (ix86_libc_has_function): New declaration.
12083 * config/i386/i386.md
12084 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
12085 ("isinf<mode>2): Likewise.
12086 * convert.c (convert_to_integer): Using new target hook
12087 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
12088 TARGET_C99_FUNCTIONS.
12089 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
12090 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
12091 * coretypes.h (function_class): New enum for different
12092 classes of functions.
12093 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
12094 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
12095 (TARGET_HAS_SINCOS): Likewise.
12096 (TARGET_LIBC_HAS_FUNCTION): New.
12097 * doc/tm.texi: Regenerated.
12098 * targhooks.h (default_libc_has_function): New declaration.
12099 (no_c99_libc_has_function): Ditto.
12100 (gnu_libc_has_function): Ditto.
12101 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
12102 and TARGET_HAS_SINCOS.
12104 2013-08-18 Jan Hubicka <jh@suse.cz>
12106 * Makeifle-in (ipa-devirt.o): New.
12107 (GTFILES): Add ipa-utils.h and ipa-devirt.c
12108 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
12109 (analyze_functions): Look into possible targets of polymorphic call.
12110 * dumpfile.c (dump_files): Add type-inheritance dump.
12111 * dumpfile.h (TDI_inheritance): New.
12112 * ipa-devirt.c: New file.
12113 * ipa-utils.h (odr_type_d): Forward declare.
12114 (odr_type): New type.
12115 (build_type_inheritance_graph): Declare.
12116 (possible_polymorphic_call_targets): Declare and introduce inline
12117 variant when only edge is pased.
12118 (dump_possible_polymorphic_call_targets): Likewise.
12119 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
12120 * tree.c (type_in_anonymous_namespace_p): Break out from ...
12121 (types_same_for_odr): ... here.
12122 * tree.h (type_in_anonymous_namespace_p): Declare.
12124 2013-08-18 Jakub Jelinek <jakub@redhat.com>
12126 PR tree-optimization/58006
12127 * tree-parloops.c (take_address_of): Don't ICE if get_name
12129 (eliminate_local_variables_stmt): Remove clobber stmts.
12131 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
12133 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
12134 error message is issued for an alias to undefined symbol.
12136 2013-08-18 Jan Hubicka <jh@suse.cz>
12138 * cgraph.c (cgraph_create_indirect_edge): Discover
12139 polymorphic calls and record basic info into indirect_info.
12140 * gimple-fold.c (gimple_fold_call): When doing BINFO based
12141 devirtualization, ignore objc function calls.
12142 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
12143 call with no parm index info.
12144 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
12145 * tree.c (virtual_method_call_p): New function.
12146 * tree.h (virtual_method_call_p): Declare.
12148 2013-08-16 Jan Hubicka <jh@suse.cz>
12150 PR middle-end/58179
12151 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
12153 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
12155 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
12158 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12160 * gengtype.c (type_for_name): Add special-case support for
12161 locating types within the "gcc::" namespace.
12162 (open_base_files): Emit a "using namespace gcc" directive.
12164 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12167 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
12168 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
12170 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
12171 array instead of each individual operand as a separate argument.
12172 (emit_fusion_gpr_load): Likewise.
12173 (expand_fusion_gpr_load): Add new function declaration.
12175 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
12176 signature to have the operands passed as an array, instead of as
12177 separate arguments. Allow ZERO_EXTEND to be in the memory
12178 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
12179 depend on the register live/dead flags when peepholes are run.
12180 (expand_fusion_gpr_load): New function to be called from the
12181 peephole2 pass, to change the register that addis sets to be the
12183 (emit_fusion_gpr_load): Change the calling signature to have the
12184 operands passed as an array, instead of as separate arguments.
12185 Allow ZERO_EXTEND to be in the memory address, and also
12186 SIGN_EXTEND if -mpower8-fusion-sign.
12188 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
12189 unspec enumeration.
12190 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
12191 adjust the register addis loads up in the peephole2 pass. Do not
12192 depend on the register live/dead state when the peephole pass is done.
12194 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12196 * gengtype.c (create_user_defined_type): Ensure that the kind
12197 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
12198 declaration is seen before the GTY((user)) marking.
12200 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
12203 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
12205 2013-08-16 Jan Hubicka <jh@suse.cz>
12207 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
12208 arugment expected_type.
12209 (gimple_fold_call): Use it.
12210 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
12211 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
12212 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
12213 (try_make_edge_direct_virtual_call): Likewise.
12214 * tree.c (obj_type_ref_class): New.
12215 * tree.h (obj_type_ref_class): Use it.
12217 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
12219 * sched-vis.c (rtl_slim_pp_initialized): Remove.
12220 (rtl_slim_pp): Likewise.
12221 (init_rtl_slim_pretty_print): Likewise.
12222 (dump_value_slim): Don't call it. Use local pretty printer.
12223 (dump_insn_slim): Likewise.
12224 (dump_rtl_slim): Likewise.
12225 (str_pattern_slim): Likewise.
12226 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
12229 2013-08-16 Jakub Jelinek <jakub@redhat.com>
12231 PR tree-optimization/58164
12232 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
12233 walk gimple_goto_dest of GIMPLE_GOTO.
12235 PR tree-optimization/58165
12236 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
12237 bi_call must be the last stmt in a bb, don't split_block, instead
12238 use fallthru edge from it and give up if there is none.
12239 Release conds vector when returning early.
12241 2013-08-14 Xinliang David Li <davidxl@google.com>
12243 * config/i386/i386.c (ix86_option_override_internal):
12244 Remove unused variable and field.
12246 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12249 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
12250 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
12251 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
12252 and Linux, correct BLKmode alignment when 128-bit alignment is
12253 required and compatibility flag is not set.
12254 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
12255 for zero-size arguments when compatibility flag is not set.
12257 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12259 PR tree-optimization/58145
12260 * tree-sra.c (build_ref_for_offset): If prev_base has
12261 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
12263 2013-08-14 Xinliang David Li <davidxl@google.com>
12265 * config/i386/i386.c (ix86_option_override_internal):
12266 Fix uninitialized variable error.
12268 2013-08-14 Xinliang David Li <davidxl@google.com>
12270 * config/i386/i386.opt: Define two new options.
12271 * config/i386/x86-tune.def: Add arch selector field in macros.
12272 * config/i386/i386.h: Adjust macro definition.
12273 * config/i386/i386.c (ix86_option_override_internal):
12275 (parse_mtune_ctrl_str): New function.
12276 (set_ix86_tune_features): New function.
12277 (ix86_function_specific_restore): Call the new helper function.
12279 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
12281 PR rtl-optimization/57662
12282 * sel-sched.c (code_motion_process_successors): When the current insn
12283 is removed after the recursive traversal, break from the loop.
12284 Add comments and debug printouts.
12286 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12287 Alexandre Oliva <aoliva@redhat.com>
12290 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
12291 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
12292 there also UNSPEC_PLTOFF.
12294 2013-08-14 Marek Polacek <polacek@redhat.com>
12296 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
12297 NUM_CONDITIONS bit positions.
12299 2013-08-13 Cary Coutant <ccoutant@google.com>
12301 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
12302 (attr_checksum): Hash vector contents instead of pointer.
12303 (attr_checksum_ordered): Likewise.
12305 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
12307 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
12308 when Pmode != word_mode. Add length_address attribute.
12309 (sse3_monitor_<mode>): Merge from sse3_monitor and
12310 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
12311 Pmode != word_mode. Update insn length attribute.
12312 * config/i386/i386.c (ix86_option_override_internal): Update
12313 ix86_gen_monitor selection for merged sse3_monitor insn.
12315 2013-08-13 Julian Brown <julian@codesourcery.com>
12317 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
12318 perform invalid legitimization on greater-than-word-size modes for
12319 TARGET_E500_DOUBLE.
12321 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
12323 * ira.c (setup_class_translate_array): Use aclass instead of cl
12324 for classes not fully covered by allocno classes.
12326 2013-08-13 Jakub Jelinek <jakub@redhat.com>
12328 PR tree-optimization/57661
12329 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
12330 * tree-inline.c (tree_function_versioning): Initialize it.
12331 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
12332 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
12333 that is not being copied.
12336 * asan.c (instrument_strlen_call): Fix typo in comment.
12337 Use char * type even for the lhs of POINTER_PLUS_EXPR.
12339 2013-08-13 Steve Ellcey <sellcey@mips.com>
12341 * config/mips/mips.md (prefetch): Use lw instead of ld on
12342 loongson in 32bit mode.
12344 2013-08-13 Nick Clifton <nickc@redhat.com>
12346 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
12348 2013-08-13 Jan Hubicka <jh@suse.cz>
12350 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
12351 introduced edge; fix typo in sanity check.
12352 (cgraph_resolve_speculation): Export; improve diagnostic.
12353 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
12354 speculation at type mismatch.
12355 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
12356 (cgraph_resolve_speculation): Declare.
12357 (symtab_can_be_discarded): New function.
12358 * value-prof.c (gimple_ic_transform): Remove actual transform code.
12359 * ipa-inline-transform.c (speculation_removed): New global var.
12360 (clone_inlined_nodes): See if speculation can be removed.
12361 (inline_call): If speculations was removed, we growths may not match.
12362 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
12363 (speculation_useful_p): New function.
12364 (resolve_noninline_speculation): New function.
12365 (inline_small_functions): Resolve useless speculations.
12366 * ipa-inline.h (speculation_useful_p): Declare
12367 * ipa.c (can_replace_by_local_alias): Simplify.
12368 (ipa_profile): Produce speculative calls in non-lto, too;
12369 add simple cost model; produce local aliases.
12371 2013-08-13 David Malcolm <dmalcolm@redhat.com>
12373 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
12376 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
12378 * config/i386/i386.c (ix86_function_versions): Use error + inform.
12380 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
12382 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
12383 iterator instead of X87MODEF.
12385 2013-08-12 Perez Read <netfirewall@gmail.com>
12388 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
12389 operand 0 for intel asm alternative.
12390 (*movabs<mode>_2): Ditto for operand 1.
12392 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
12394 * config/aarch64/arm_none.h
12395 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
12397 2013-08-12 Nick Clifton <nickc@redhat.com>
12399 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
12401 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12403 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
12404 expand for QI/HImode operand to produce more effictive code for
12405 unsigned char(short) --> float(double) conversion.
12407 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
12409 * doc/invoke.texi: Mention that -ftls-model does not force the final
12412 2013-08-12 Marek Polacek <polacek@redhat.com>
12413 Marc Glisse <marc.glisse@inria.fr>
12415 PR tree-optimization/57980
12416 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12417 when creating -1 constant.
12419 2013-08-10 Jan Hubicka <jh@suse.cz>
12421 Workaround binutils PR14342.
12422 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12423 (gimple_init_edge_profiler): Likewise.
12424 (gimple_gen_ic_func_profiler): Likewise.
12426 2013-08-09 Jan Hubicka <jh@suse.cz>
12428 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12430 2013-08-09 Xinliang David Li <davidxl@google.com>
12432 * config/i386/stringop.def: New file.
12433 * config/i386/stringop.opt: New file.
12434 * config/i386/i386-opts.h: Include stringopt.def.
12435 * config/i386/i386.opt: Include stringopt.opt.
12436 * config/i386/i386.c (ix86_option_override_internal):
12437 Override default size based stringop inline strategies with options.
12438 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12441 2013-08-09 Jan Hubicka <jh@suse.cz>
12443 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12445 2013-08-09 Jan Hubicka <jh@suse.cz>
12447 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12449 (dump_cgraph_node): Dump profile-id.
12450 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12451 and common_target_probability.
12452 * lto-cgraph.c (lto_output_edge): Stream common targets.
12453 (lto_output_node): Stream profile ids.
12454 (input_node): Stream profile ids.
12455 (input_edge): Stream common targets.
12456 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12457 * ipa.c: Include value-prof.h
12458 (ipa_profile_generate_summary): Turn indirect call statement histograms
12459 into common targets.
12460 (ipa_profile): Turn common targets into speculative edges.
12462 2013-08-09 Jan Hubicka <jh@suse.cz>
12464 * cgraph.h (cgraph_node): Add profile_id.
12465 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12466 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12467 (del_node_map): Update.
12468 (find_func_by_funcdef_no): Replace by ...
12469 (find_func_by_profile_id): ... this one.
12470 (gimple_ic_transform): Do not remove useful histograms when
12471 speculation is not done; dump info when indirect call removal
12473 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12474 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12475 (__gcov_indirect_call_profiler_v2): .. this one.
12476 * profile.h (init_node_map): Update.
12477 * coverage.c (coverage_compute_profile_id): New function.
12478 * coverage.h (coverage_compute_profile_id): Declare.
12479 * tree-profile.c (init_ic_make_global_vars): Make
12480 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12481 (gimple_init_edge_profiler): Update prototype of
12482 __gcov_indirect_call_profiler.
12483 (gimple_gen_ic_func_profiler): Simplify.
12484 (tree_profiling): Use init_node_map
12486 2013-08-09 Jan Hubicka <jh@suse.cz>
12488 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12489 non-speculative refs.
12490 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12491 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12492 (cgraph_set_call_stmt): Likewise.
12493 (cgraph_create_edge_1): Fix release checking compilatoin;
12494 clear lto_stmt_uid.
12495 (cgraph_free_edge): Free indirect info.
12496 (cgraph_turn_edge_to_speculative): New function.
12497 (cgraph_speculative_call_info): New function.
12498 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12499 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12500 (dump_cgraph_node): Dump speculation.
12501 (verify_edge_count_and_frequency): Accept speculative edges.
12502 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12503 (verify_cgraph_node): Handle speculation.
12504 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12505 (cgraph_set_call_stmt): Update prototype.
12506 (cgraph_make_edge_direct): Update prototype.
12507 (cgraph_speculative_call_info): Declare.
12508 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12509 to change; update call of ipa_find_references.
12510 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12511 lto_stmt_uid and speculative flags.
12512 (ipa_dump_references): Dump speculation.
12513 (ipa_clone_references): Clone speculative flag.
12514 (ipa_clone_referring): Likewise.
12515 (ipa_clone_ref): New function.
12516 (ipa_find_reference): Look into lto_stmt_uids
12517 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12518 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12519 (ipa_find_reference): Update declaration.
12520 (ipa_clone_ref): Declare.
12521 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12522 stream speculative flag.
12523 (lto_output_ref): Stream statements uids and speculation.
12524 (input_ref): Likewise.
12525 (input_edge): Stream speuclation.
12526 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12527 (cgraph_set_call_stmt_including_clones): Handle speculation.
12528 * ipa-inline.c (heap_edge_removal_hook): New function.
12529 (inline_small_functions): Register it.
12530 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12531 also initialize refs.
12532 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12534 (try_make_edge_direct_simple_call): Likewise.
12535 (try_make_edge_direct_simple_call): Likewise.
12536 (update_indirect_edges_after_inlining): Likewise.
12537 (remove_described_reference): Look proper lto_stmt_uid.
12538 (propagate_controlled_uses): Likewise.
12539 (propagate_controlled_uses): Liekwise.
12540 * tree-inline.c (copy_bb): Copy speculative edges.
12541 (redirect_all_calls): New function.
12542 (copy_cfg_body): Do redirection after loop info is updated.
12543 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12545 2013-08-09 Jan Hubicka <jh@suse.cz>
12547 * lto-streamer-out.c (output_function): Renumber PHIs.
12548 * lto-streamer-in.c (input_function): Likewise.
12550 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12552 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12553 (get_lane_unsigned): Likewise.
12554 (dup_lane_scalar): Likewise.
12555 (get_lane): enable for VALL.
12556 * config/aarch64/aarch64-simd.md
12557 (aarch64_dup_lane_scalar<mode>): Remove.
12558 (aarch64_get_lane_signed<mode>): Likewise.
12559 (aarch64_get_lane_unsigned<mode>): Likewise.
12560 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
12561 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
12562 (aarch64_get_lane<mode>): Enable for all vector modes.
12563 (aarch64_get_lanedi): Remove misleading constraints.
12564 * config/aarch64/arm_neon.h
12565 (__aarch64_vget_lane_any): Define.
12566 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
12567 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
12568 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
12569 * config/aarch64/iterators.md (VDQQH): New.
12570 (VDQQHS): Likewise.
12571 (vwcore): Likewise.
12573 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
12575 * configure.ac: Add GAS check for LEON instructions on SPARC.
12576 * configure: Regenerate.
12577 * config.in: Likewise.
12578 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
12580 * config/sparc/sparc.opt (LEON, LEON3): New masks.
12581 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
12583 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
12584 (AS_LEON_FLAG): New macro.
12585 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
12586 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
12587 Deal with LEON and LEON3 for the memory model.
12588 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
12589 (atomic_compare_and_swap<mode>_1): Likewise.
12590 (*atomic_compare_and_swap<mode>_1): Likewise.
12592 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12594 * config/arm/neon.md (vcond): Fix floating-point vector
12595 comparisons against 0.
12597 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
12599 * lra-constraints.c (emit_spill_move): Remove assert.
12600 (process_alt_operands): Add more debugging
12601 output. Increase reject for spilling into memory. Decrease
12602 reject for reloading scratch.
12603 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12605 2013-08-08 Steve Ellcey <sellcey@mips.com>
12607 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12608 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12609 micromips incompatible. Add nan2008.
12610 (MULTILIB_DIRNAMES): Add nan2008.
12611 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12612 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12613 and micromips incompatible. Add nan2008.
12614 (MULTILIB_DIRNAMES): Add nan2008.
12615 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12617 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12619 PR rtl-optimization/58079
12620 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12621 simplify_comparison has widened a comparison with an integer.
12623 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12625 * config/arm/neon.md (movmisalign<mode>): Disable when we
12626 don't allow unaligned accesses.
12627 (*movmisalign<mode>_neon_store): Likewise.
12628 (*movmisalign<mode>_neon_load): Likewise.
12629 (*movmisalign<mode>_neon_store): Likewise.
12630 (*movmisalign<mode>_neon_load): Likewise.
12632 2013-08-08 Jan Hubicka <jh@suse.cz>
12634 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12635 (make_pass_rebuild_cgraph_edges): Also clear references.
12636 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12637 * ipa-inline-transform.c (inline_transform): Remove all references
12639 * cgraphunit.c (expand_function): Remove all references after
12641 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12642 (ipa_find_reference): Rewrite to iterator.
12643 (remove_stmt_references): Likewise.
12644 (ipa_clear_stmts_in_references): New function.
12645 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12646 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12648 * ipa-split.c (split_function): Remove references in split function.
12650 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12653 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12654 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12656 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12659 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12660 suggested mode for the assignment isn't compatible with the
12661 registers required.
12663 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12666 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12668 2013-08-07 Xinliang David Li <davidxl@google.com>
12670 * config/i386/i386.opt: New option -mtune-ctrl=.
12671 * config/i386/x86-tune.def: New file.
12672 * config/i386/i386.h: include x86-tune.def.
12673 * config/i386/i386.c (ix86_option_override_internal):
12674 Parsing -mtune-ctrl= option and set tune features.
12676 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12679 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12680 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12683 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12685 * rtl.h (update_alignments): Declare.
12686 * final.c (grow_label_align): New function extracted from...
12687 (shorten_branches): ...here. Call it.
12688 (update_alignments): New function.
12689 * reorg.c (sibling_labels): New variable.
12690 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12691 the new label along with it onto the sibling_labels vector.
12692 (fill_simple_delay_slots): Adjust call to get_label_before.
12693 (fill_slots_from_thread): Likewise.
12694 (relax_delay_slots): Likewise.
12695 (make_return_insns): Likewise.
12696 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12698 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12700 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12701 document its semantics.
12702 (diagnostic_report_diagnostic): Adjust accordingly.
12704 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12706 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12707 (sparc_option_override): ...and port to new C++ pass API.
12708 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12710 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12712 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12714 2013-08-06 Caroline Tice <cmtice@google.com>
12716 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12717 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12718 * tree-pass.h: Add pass_vtable_verify.
12719 * varasm.c (assemble_variable): Add code to properly set the comdat
12720 section and name for the .vtable_map_vars section.
12721 (assemble_vtyv_preinit_initializer): New function.
12722 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12724 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12725 * vtable-verify.c: New file.
12726 * vtable-verify.h: New file.
12727 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12728 initialiation levels.
12729 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12730 * passes.def: Insert pass_vtable_verify.
12731 * aclocal.m4: Reorder includes.
12732 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12733 -fvtv-counts options.
12734 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12735 as appropriate, if -fvtable-verify=... is used.
12736 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12737 -fvtable-verify=... is used.
12738 * Makefile.in (OBJS): Add vtable-verify.o to list.
12739 (vtable-verify.o): Add new build rule.
12740 (GTFILES): Add vtable-verify.c to list.
12741 * common.opt (fvtable-verify=): New flag.
12742 (vtv_priority): Values for fvtable-verify= flag.
12743 (fvtv-counts): New flag.
12744 (fvtv-debug): New flag.
12745 * tree.h (save_vtable_map_decl): New extern function decl.
12747 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12749 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12750 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12751 (pass_data_rl78_devirt): ...new pass_data instance and...
12752 (make_pass_rl78_devirt): ...new function.
12753 (rl78_asm_file_start): Port pass registration to new C++ API.
12755 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12757 * coretypes.h (rtl_opt_pass): Add.
12758 (gcc::context): Add.
12759 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12760 (epiphany_init): Port to new C++ pass API.
12761 (epiphany_optimize_mode_switching): Likewise.
12762 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12763 (pass_manager::get_pass_mode_switching): New.
12764 (pass_manager::get_pass_peephole2): New.
12765 * mode-switching.c (pass_mode_switching): Add clone method.
12766 * recog.c (pass_peephole2): Add clone method.
12767 (pass_split_all_insns): Add clone method.
12769 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12771 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12772 (mips_option_override): ...here, porting to new C++ API for passes.
12774 2013-08-06 Jan Hubicka <jh@suse.cz>
12776 * cgraph.c (cgraph_get_body): New function based on lto.c
12778 * cgraph.h (cgraph_get_body): Declare.
12779 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12781 * cgraphunit.c (expand_function): Get body prior expanding.
12782 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12783 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12785 * passes.c (do_per_function_toporder): Get body.
12786 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12787 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12789 2013-08-06 Martin Jambor <mjambor@suse.cz>
12792 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12793 is not re-finalized. Rename second parameter to no_collect.
12795 2013-08-06 Martin Jambor <mjambor@suse.cz>
12797 PR middle-end/58041
12798 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12799 MEM_REF has proper alignment information.
12801 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12804 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12806 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12807 rtx (*) (rtx, ...) with insn_gen_fn.
12808 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12809 insn_gen_fn::stored_funcptr. Add initializer braces.
12811 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12813 Rewrite how instances of passes are cloned to remove assumptions
12814 about their sizes (thus allowing pass subclasses to have
12815 additional data fields, albeit non-GC-managed ones at this point).
12817 * passes.c (make_pass_instance): Now that passes have clone
12818 methods, rewrite this function to eliminate XNEW and memcpy
12819 calls that used hardcoded sizes. Since this function no longer
12820 creates pass instances, rename it to...
12821 (add_pass_instance): ...this. Document the old way that passes were
12822 numbered and flagged, and rework this function to continue using it.
12823 (next_pass_1): Add an initial_pass argument for use by
12825 (position_pass): When adding multiple instances of a pass, use
12826 the pass's clone method, rather than relying on the XNEW/memcpy
12827 within the former make_pass_instance (now add_pass_instance).
12828 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12829 the initial instance of the current pass within the pass manager.
12831 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12833 This is the automated part of the conversion of passes from C
12834 structs to C++ classes.
12836 Patch autogenerated by refactor_passes.py from
12837 https://github.com/davidmalcolm/gcc-refactoring-scripts
12838 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12840 * asan.c (pass_asan): Convert from a global struct to a subclass of
12841 gimple_opt_pass along with...
12842 (pass_data_asan): ...new pass_data instance and...
12843 (make_pass_asan): ...new function.
12844 (pass_asan_O0): Convert from a global struct to a subclass of
12845 gimple_opt_pass along with...
12846 (pass_data_asan_O0): ...new pass_data instance and...
12847 (make_pass_asan_O0): ...new function.
12848 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12849 subclass of rtl_opt_pass along with...
12850 (pass_data_inc_dec): ...new pass_data instance and...
12851 (make_pass_inc_dec): ...new function.
12852 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12853 a subclass of rtl_opt_pass along with...
12854 (pass_data_reorder_blocks): ...new pass_data instance and...
12855 (make_pass_reorder_blocks): ...new function.
12856 (pass_duplicate_computed_gotos): Convert from a global struct to a
12857 subclass of rtl_opt_pass along with...
12858 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12859 (make_pass_duplicate_computed_gotos): ...new function.
12860 (pass_partition_blocks): Convert from a global struct to a subclass of
12861 rtl_opt_pass along with...
12862 (pass_data_partition_blocks): ...new pass_data instance and...
12863 (make_pass_partition_blocks): ...new function.
12864 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12865 struct to a subclass of rtl_opt_pass along with...
12866 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12868 (make_pass_branch_target_load_optimize1): ...new function.
12869 (pass_branch_target_load_optimize2): Convert from a global struct to a
12870 subclass of rtl_opt_pass along with...
12871 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12873 (make_pass_branch_target_load_optimize2): ...new function.
12874 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12875 of rtl_opt_pass along with...
12876 (pass_data_jump): ...new pass_data instance and...
12877 (make_pass_jump): ...new function.
12878 (pass_jump2): Convert from a global struct to a subclass of
12879 rtl_opt_pass along with...
12880 (pass_data_jump2): ...new pass_data instance and...
12881 (make_pass_jump2): ...new function.
12882 * cfgexpand.c (pass_expand): Convert from a global struct to a
12883 subclass of rtl_opt_pass along with...
12884 (pass_data_expand): ...new pass_data instance and...
12885 (make_pass_expand): ...new function.
12886 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12887 of rtl_opt_pass along with...
12888 (pass_data_free_cfg): ...new pass_data instance and...
12889 (make_pass_free_cfg): ...new function.
12890 (pass_into_cfg_layout_mode): Convert from a global struct to a
12891 subclass of rtl_opt_pass along with...
12892 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12893 (make_pass_into_cfg_layout_mode): ...new function.
12894 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12895 subclass of rtl_opt_pass along with...
12896 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12897 (make_pass_outof_cfg_layout_mode): ...new function.
12898 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12899 struct to a subclass of gimple_opt_pass along with...
12900 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12901 (make_pass_build_cgraph_edges): ...new function.
12902 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12903 subclass of gimple_opt_pass along with...
12904 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12905 (make_pass_rebuild_cgraph_edges): ...new function.
12906 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12907 subclass of gimple_opt_pass along with...
12908 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12910 (make_pass_remove_cgraph_callee_edges): ...new function.
12911 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12912 struct to a subclass of rtl_opt_pass along with...
12913 (pass_data_stack_adjustments): ...new pass_data instance and...
12914 (make_pass_stack_adjustments): ...new function.
12915 * combine.c (pass_combine): Convert from a global struct to a subclass
12916 of rtl_opt_pass along with...
12917 (pass_data_combine): ...new pass_data instance and...
12918 (make_pass_combine): ...new function.
12919 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12920 global struct to a subclass of rtl_opt_pass along with...
12921 (pass_data_compare_elim_after_reload): ...new pass_data instance
12923 (make_pass_compare_elim_after_reload): ...new function.
12924 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12925 of rtl_opt_pass along with...
12926 (pass_data_rtl_cprop): ...new pass_data instance and...
12927 (make_pass_rtl_cprop): ...new function.
12928 * cse.c (pass_cse): Convert from a global struct to a subclass of
12929 rtl_opt_pass along with...
12930 (pass_data_cse): ...new pass_data instance and...
12931 (make_pass_cse): ...new function.
12932 (pass_cse2): Convert from a global struct to a subclass of
12933 rtl_opt_pass along with...
12934 (pass_data_cse2): ...new pass_data instance and...
12935 (make_pass_cse2): ...new function.
12936 (pass_cse_after_global_opts): Convert from a global struct to a
12937 subclass of rtl_opt_pass along with...
12938 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12939 (make_pass_cse_after_global_opts): ...new function.
12940 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12941 of rtl_opt_pass along with...
12942 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12943 (make_pass_ud_rtl_dce): ...new function.
12944 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12945 rtl_opt_pass along with...
12946 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12947 (make_pass_fast_rtl_dce): ...new function.
12948 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12949 a subclass of rtl_opt_pass along with...
12950 (pass_data_df_initialize_opt): ...new pass_data instance and...
12951 (make_pass_df_initialize_opt): ...new function.
12952 (pass_df_initialize_no_opt): Convert from a global struct to a
12953 subclass of rtl_opt_pass along with...
12954 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12955 (make_pass_df_initialize_no_opt): ...new function.
12956 (pass_df_finish): Convert from a global struct to a subclass of
12957 rtl_opt_pass along with...
12958 (pass_data_df_finish): ...new pass_data instance and...
12959 (make_pass_df_finish): ...new function.
12960 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12961 rtl_opt_pass along with...
12962 (pass_data_rtl_dse1): ...new pass_data instance and...
12963 (make_pass_rtl_dse1): ...new function.
12964 (pass_rtl_dse2): Convert from a global struct to a subclass of
12965 rtl_opt_pass along with...
12966 (pass_data_rtl_dse2): ...new pass_data instance and...
12967 (make_pass_rtl_dse2): ...new function.
12968 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12969 subclass of rtl_opt_pass along with...
12970 (pass_data_dwarf2_frame): ...new pass_data instance and...
12971 (make_pass_dwarf2_frame): ...new function.
12972 * except.c (pass_set_nothrow_function_flags): Convert from a global
12973 struct to a subclass of rtl_opt_pass along with...
12974 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12976 (make_pass_set_nothrow_function_flags): ...new function.
12977 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12978 subclass of rtl_opt_pass along with...
12979 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12981 (make_pass_convert_to_eh_region_ranges): ...new function.
12982 * final.c (pass_compute_alignments): Convert from a global struct to a
12983 subclass of rtl_opt_pass along with...
12984 (pass_data_compute_alignments): ...new pass_data instance and...
12985 (make_pass_compute_alignments): ...new function.
12986 (pass_final): Convert from a global struct to a subclass of
12987 rtl_opt_pass along with...
12988 (pass_data_final): ...new pass_data instance and...
12989 (make_pass_final): ...new function.
12990 (pass_shorten_branches): Convert from a global struct to a subclass of
12991 rtl_opt_pass along with...
12992 (pass_data_shorten_branches): ...new pass_data instance and...
12993 (make_pass_shorten_branches): ...new function.
12994 (pass_clean_state): Convert from a global struct to a subclass of
12995 rtl_opt_pass along with...
12996 (pass_data_clean_state): ...new pass_data instance and...
12997 (make_pass_clean_state): ...new function.
12998 * function.c (pass_instantiate_virtual_regs): Convert from a global
12999 struct to a subclass of rtl_opt_pass along with...
13000 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
13001 (make_pass_instantiate_virtual_regs): ...new function.
13002 (pass_leaf_regs): Convert from a global struct to a subclass of
13003 rtl_opt_pass along with...
13004 (pass_data_leaf_regs): ...new pass_data instance and...
13005 (make_pass_leaf_regs): ...new function.
13006 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
13007 subclass of rtl_opt_pass along with...
13008 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
13010 (make_pass_thread_prologue_and_epilogue): ...new function.
13011 (pass_match_asm_constraints): Convert from a global struct to a
13012 subclass of rtl_opt_pass along with...
13013 (pass_data_match_asm_constraints): ...new pass_data instance and...
13014 (make_pass_match_asm_constraints): ...new function.
13015 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
13016 subclass of rtl_opt_pass along with...
13017 (pass_data_rtl_fwprop): ...new pass_data instance and...
13018 (make_pass_rtl_fwprop): ...new function.
13019 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
13020 rtl_opt_pass along with...
13021 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
13022 (make_pass_rtl_fwprop_addr): ...new function.
13023 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
13024 rtl_opt_pass along with...
13025 (pass_data_rtl_pre): ...new pass_data instance and...
13026 (make_pass_rtl_pre): ...new function.
13027 (pass_rtl_hoist): Convert from a global struct to a subclass of
13028 rtl_opt_pass along with...
13029 (pass_data_rtl_hoist): ...new pass_data instance and...
13030 (make_pass_rtl_hoist): ...new function.
13031 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
13032 subclass of gimple_opt_pass along with...
13033 (pass_data_lower_cf): ...new pass_data instance and...
13034 (make_pass_lower_cf): ...new function.
13035 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
13036 from a global struct to a subclass of gimple_opt_pass along with...
13037 (pass_data_strength_reduction): ...new pass_data instance and...
13038 (make_pass_strength_reduction): ...new function.
13039 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
13040 of rtl_opt_pass along with...
13041 (pass_data_rtl_ifcvt): ...new pass_data instance and...
13042 (make_pass_rtl_ifcvt): ...new function.
13043 (pass_if_after_combine): Convert from a global struct to a subclass of
13044 rtl_opt_pass along with...
13045 (pass_data_if_after_combine): ...new pass_data instance and...
13046 (make_pass_if_after_combine): ...new function.
13047 (pass_if_after_reload): Convert from a global struct to a subclass of
13048 rtl_opt_pass along with...
13049 (pass_data_if_after_reload): ...new pass_data instance and...
13050 (make_pass_if_after_reload): ...new function.
13051 * init-regs.c (pass_initialize_regs): Convert from a global struct to
13052 a subclass of rtl_opt_pass along with...
13053 (pass_data_initialize_regs): ...new pass_data instance and...
13054 (make_pass_initialize_regs): ...new function.
13055 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
13056 of ipa_opt_pass_d along with...
13057 (pass_data_ipa_cp): ...new pass_data instance and...
13058 (make_pass_ipa_cp): ...new function.
13059 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
13060 global struct to a subclass of gimple_opt_pass along with...
13061 (pass_data_inline_parameters): ...new pass_data instance and...
13062 (make_pass_inline_parameters): ...new function.
13063 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
13064 subclass of gimple_opt_pass along with...
13065 (pass_data_early_inline): ...new pass_data instance and...
13066 (make_pass_early_inline): ...new function.
13067 (pass_ipa_inline): Convert from a global struct to a subclass of
13068 ipa_opt_pass_d along with...
13069 (pass_data_ipa_inline): ...new pass_data instance and...
13070 (make_pass_ipa_inline): ...new function.
13071 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
13072 struct to a subclass of gimple_opt_pass along with...
13073 (pass_data_local_pure_const): ...new pass_data instance and...
13074 (make_pass_local_pure_const): ...new function.
13075 (pass_ipa_pure_const): Convert from a global struct to a subclass of
13076 ipa_opt_pass_d along with...
13077 (pass_data_ipa_pure_const): ...new pass_data instance and...
13078 (make_pass_ipa_pure_const): ...new function.
13079 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
13080 to a subclass of ipa_opt_pass_d along with...
13081 (pass_data_ipa_reference): ...new pass_data instance and...
13082 (make_pass_ipa_reference): ...new function.
13083 * ipa-split.c (pass_split_functions): Convert from a global struct to
13084 a subclass of gimple_opt_pass along with...
13085 (pass_data_split_functions): ...new pass_data instance and...
13086 (make_pass_split_functions): ...new function.
13087 (pass_feedback_split_functions): Convert from a global struct to a
13088 subclass of gimple_opt_pass along with...
13089 (pass_data_feedback_split_functions): ...new pass_data instance and...
13090 (make_pass_feedback_split_functions): ...new function.
13091 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
13092 global struct to a subclass of simple_ipa_opt_pass along with...
13093 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
13095 (make_pass_ipa_function_and_variable_visibility): ...new function.
13096 (pass_ipa_free_inline_summary): Convert from a global struct to a
13097 subclass of simple_ipa_opt_pass along with...
13098 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
13099 (make_pass_ipa_free_inline_summary): ...new function.
13100 (pass_ipa_whole_program_visibility): Convert from a global struct to a
13101 subclass of ipa_opt_pass_d along with...
13102 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
13104 (make_pass_ipa_whole_program_visibility): ...new function.
13105 (pass_ipa_profile): Convert from a global struct to a subclass of
13106 ipa_opt_pass_d along with...
13107 (pass_data_ipa_profile): ...new pass_data instance and...
13108 (make_pass_ipa_profile): ...new function.
13109 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
13110 ipa_opt_pass_d along with...
13111 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
13112 (make_pass_ipa_cdtor_merge): ...new function.
13113 * ira.c (pass_ira): Convert from a global struct to a subclass of
13114 rtl_opt_pass along with...
13115 (pass_data_ira): ...new pass_data instance and...
13116 (make_pass_ira): ...new function.
13117 (pass_reload): Convert from a global struct to a subclass of
13118 rtl_opt_pass along with...
13119 (pass_data_reload): ...new pass_data instance and...
13120 (make_pass_reload): ...new function.
13121 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
13122 subclass of rtl_opt_pass along with...
13123 (pass_data_cleanup_barriers): ...new pass_data instance and...
13124 (make_pass_cleanup_barriers): ...new function.
13125 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
13126 of rtl_opt_pass along with...
13127 (pass_data_loop2): ...new pass_data instance and...
13128 (make_pass_loop2): ...new function.
13129 (pass_rtl_loop_init): Convert from a global struct to a subclass of
13130 rtl_opt_pass along with...
13131 (pass_data_rtl_loop_init): ...new pass_data instance and...
13132 (make_pass_rtl_loop_init): ...new function.
13133 (pass_rtl_loop_done): Convert from a global struct to a subclass of
13134 rtl_opt_pass along with...
13135 (pass_data_rtl_loop_done): ...new pass_data instance and...
13136 (make_pass_rtl_loop_done): ...new function.
13137 (pass_rtl_move_loop_invariants): Convert from a global struct to a
13138 subclass of rtl_opt_pass along with...
13139 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
13140 (make_pass_rtl_move_loop_invariants): ...new function.
13141 (pass_rtl_unswitch): Convert from a global struct to a subclass of
13142 rtl_opt_pass along with...
13143 (pass_data_rtl_unswitch): ...new pass_data instance and...
13144 (make_pass_rtl_unswitch): ...new function.
13145 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
13146 subclass of rtl_opt_pass along with...
13147 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
13149 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13150 (pass_rtl_doloop): Convert from a global struct to a subclass of
13151 rtl_opt_pass along with...
13152 (pass_data_rtl_doloop): ...new pass_data instance and...
13153 (make_pass_rtl_doloop): ...new function.
13154 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
13155 a subclass of rtl_opt_pass along with...
13156 (pass_data_lower_subreg): ...new pass_data instance and...
13157 (make_pass_lower_subreg): ...new function.
13158 (pass_lower_subreg2): Convert from a global struct to a subclass of
13159 rtl_opt_pass along with...
13160 (pass_data_lower_subreg2): ...new pass_data instance and...
13161 (make_pass_lower_subreg2): ...new function.
13162 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
13163 struct to a subclass of ipa_opt_pass_d along with...
13164 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
13165 (make_pass_ipa_lto_gimple_out): ...new function.
13166 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
13167 of ipa_opt_pass_d along with...
13168 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
13169 (make_pass_ipa_lto_finish_out): ...new function.
13170 * mode-switching.c (pass_mode_switching): Convert from a global struct
13171 to a subclass of rtl_opt_pass along with...
13172 (pass_data_mode_switching): ...new pass_data instance and...
13173 (make_pass_mode_switching): ...new function.
13174 * modulo-sched.c (pass_sms): Convert from a global struct to a
13175 subclass of rtl_opt_pass along with...
13176 (pass_data_sms): ...new pass_data instance and...
13177 (make_pass_sms): ...new function.
13178 * omp-low.c (pass_expand_omp): Convert from a global struct to a
13179 subclass of gimple_opt_pass along with...
13180 (pass_data_expand_omp): ...new pass_data instance and...
13181 (make_pass_expand_omp): ...new function.
13182 (pass_lower_omp): Convert from a global struct to a subclass of
13183 gimple_opt_pass along with...
13184 (pass_data_lower_omp): ...new pass_data instance and...
13185 (make_pass_lower_omp): ...new function.
13186 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
13187 of gimple_opt_pass along with...
13188 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
13189 (make_pass_diagnose_omp_blocks): ...new function.
13190 * passes.c (pass_early_local_passes): Convert from a global struct to
13191 a subclass of simple_ipa_opt_pass along with...
13192 (pass_data_early_local_passes): ...new pass_data instance and...
13193 (make_pass_early_local_passes): ...new function.
13194 (pass_all_early_optimizations): Convert from a global struct to a
13195 subclass of gimple_opt_pass along with...
13196 (pass_data_all_early_optimizations): ...new pass_data instance and...
13197 (make_pass_all_early_optimizations): ...new function.
13198 (pass_all_optimizations): Convert from a global struct to a subclass
13199 of gimple_opt_pass along with...
13200 (pass_data_all_optimizations): ...new pass_data instance and...
13201 (make_pass_all_optimizations): ...new function.
13202 (pass_all_optimizations_g): Convert from a global struct to a subclass
13203 of gimple_opt_pass along with...
13204 (pass_data_all_optimizations_g): ...new pass_data instance and...
13205 (make_pass_all_optimizations_g): ...new function.
13206 (pass_rest_of_compilation): Convert from a global struct to a subclass
13207 of rtl_opt_pass along with...
13208 (pass_data_rest_of_compilation): ...new pass_data instance and...
13209 (make_pass_rest_of_compilation): ...new function.
13210 (pass_postreload): Convert from a global struct to a subclass of
13211 rtl_opt_pass along with...
13212 (pass_data_postreload): ...new pass_data instance and...
13213 (make_pass_postreload): ...new function.
13214 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
13215 subclass of rtl_opt_pass along with...
13216 (pass_data_gcse2): ...new pass_data instance and...
13217 (make_pass_gcse2): ...new function.
13218 * postreload.c (pass_postreload_cse): Convert from a global struct to
13219 a subclass of rtl_opt_pass along with...
13220 (pass_data_postreload_cse): ...new pass_data instance and...
13221 (make_pass_postreload_cse): ...new function.
13222 * predict.c (pass_profile): Convert from a global struct to a subclass
13223 of gimple_opt_pass along with...
13224 (pass_data_profile): ...new pass_data instance and...
13225 (make_pass_profile): ...new function.
13226 (pass_strip_predict_hints): Convert from a global struct to a subclass
13227 of gimple_opt_pass along with...
13228 (pass_data_strip_predict_hints): ...new pass_data instance and...
13229 (make_pass_strip_predict_hints): ...new function.
13230 * recog.c (pass_peephole2): Convert from a global struct to a subclass
13231 of rtl_opt_pass along with...
13232 (pass_data_peephole2): ...new pass_data instance and...
13233 (make_pass_peephole2): ...new function.
13234 (pass_split_all_insns): Convert from a global struct to a subclass of
13235 rtl_opt_pass along with...
13236 (pass_data_split_all_insns): ...new pass_data instance and...
13237 (make_pass_split_all_insns): ...new function.
13238 (pass_split_after_reload): Convert from a global struct to a subclass
13239 of rtl_opt_pass along with...
13240 (pass_data_split_after_reload): ...new pass_data instance and...
13241 (make_pass_split_after_reload): ...new function.
13242 (pass_split_before_regstack): Convert from a global struct to a
13243 subclass of rtl_opt_pass along with...
13244 (pass_data_split_before_regstack): ...new pass_data instance and...
13245 (make_pass_split_before_regstack): ...new function.
13246 (pass_split_before_sched2): Convert from a global struct to a subclass
13247 of rtl_opt_pass along with...
13248 (pass_data_split_before_sched2): ...new pass_data instance and...
13249 (make_pass_split_before_sched2): ...new function.
13250 (pass_split_for_shorten_branches): Convert from a global struct to a
13251 subclass of rtl_opt_pass along with...
13252 (pass_data_split_for_shorten_branches): ...new pass_data instance
13254 (make_pass_split_for_shorten_branches): ...new function.
13255 * ree.c (pass_ree): Convert from a global struct to a subclass of
13256 rtl_opt_pass along with...
13257 (pass_data_ree): ...new pass_data instance and...
13258 (make_pass_ree): ...new function.
13259 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
13260 subclass of rtl_opt_pass along with...
13261 (pass_data_stack_regs): ...new pass_data instance and...
13262 (make_pass_stack_regs): ...new function.
13263 (pass_stack_regs_run): Convert from a global struct to a subclass of
13264 rtl_opt_pass along with...
13265 (pass_data_stack_regs_run): ...new pass_data instance and...
13266 (make_pass_stack_regs_run): ...new function.
13267 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
13268 subclass of rtl_opt_pass along with...
13269 (pass_data_cprop_hardreg): ...new pass_data instance and...
13270 (make_pass_cprop_hardreg): ...new function.
13271 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
13272 subclass of rtl_opt_pass along with...
13273 (pass_data_reginfo_init): ...new pass_data instance and...
13274 (make_pass_reginfo_init): ...new function.
13275 * regmove.c (pass_regmove): Convert from a global struct to a subclass
13276 of rtl_opt_pass along with...
13277 (pass_data_regmove): ...new pass_data instance and...
13278 (make_pass_regmove): ...new function.
13279 * regrename.c (pass_regrename): Convert from a global struct to a
13280 subclass of rtl_opt_pass along with...
13281 (pass_data_regrename): ...new pass_data instance and...
13282 (make_pass_regrename): ...new function.
13283 * reorg.c (pass_delay_slots): Convert from a global struct to a
13284 subclass of rtl_opt_pass along with...
13285 (pass_data_delay_slots): ...new pass_data instance and...
13286 (make_pass_delay_slots): ...new function.
13287 (pass_machine_reorg): Convert from a global struct to a subclass of
13288 rtl_opt_pass along with...
13289 (pass_data_machine_reorg): ...new pass_data instance and...
13290 (make_pass_machine_reorg): ...new function.
13291 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
13292 of rtl_opt_pass along with...
13293 (pass_data_sched): ...new pass_data instance and...
13294 (make_pass_sched): ...new function.
13295 (pass_sched2): Convert from a global struct to a subclass of
13296 rtl_opt_pass along with...
13297 (pass_data_sched2): ...new pass_data instance and...
13298 (make_pass_sched2): ...new function.
13299 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
13300 to a subclass of rtl_opt_pass along with...
13301 (pass_data_stack_ptr_mod): ...new pass_data instance and...
13302 (make_pass_stack_ptr_mod): ...new function.
13303 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
13304 to a subclass of rtl_opt_pass along with...
13305 (pass_data_rtl_store_motion): ...new pass_data instance and...
13306 (make_pass_rtl_store_motion): ...new function.
13307 * tracer.c (pass_tracer): Convert from a global struct to a subclass
13308 of gimple_opt_pass along with...
13309 (pass_data_tracer): ...new pass_data instance and...
13310 (make_pass_tracer): ...new function.
13311 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
13312 to a subclass of gimple_opt_pass along with...
13313 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
13314 (make_pass_diagnose_tm_blocks): ...new function.
13315 (pass_lower_tm): Convert from a global struct to a subclass of
13316 gimple_opt_pass along with...
13317 (pass_data_lower_tm): ...new pass_data instance and...
13318 (make_pass_lower_tm): ...new function.
13319 (pass_tm_init): Convert from a global struct to a subclass of
13320 gimple_opt_pass along with...
13321 (pass_data_tm_init): ...new pass_data instance and...
13322 (make_pass_tm_init): ...new function.
13323 (pass_tm_mark): Convert from a global struct to a subclass of
13324 gimple_opt_pass along with...
13325 (pass_data_tm_mark): ...new pass_data instance and...
13326 (make_pass_tm_mark): ...new function.
13327 (pass_tm_edges): Convert from a global struct to a subclass of
13328 gimple_opt_pass along with...
13329 (pass_data_tm_edges): ...new pass_data instance and...
13330 (make_pass_tm_edges): ...new function.
13331 (pass_tm_memopt): Convert from a global struct to a subclass of
13332 gimple_opt_pass along with...
13333 (pass_data_tm_memopt): ...new pass_data instance and...
13334 (make_pass_tm_memopt): ...new function.
13335 (pass_ipa_tm): Convert from a global struct to a subclass of
13336 simple_ipa_opt_pass along with...
13337 (pass_data_ipa_tm): ...new pass_data instance and...
13338 (make_pass_ipa_tm): ...new function.
13339 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
13340 subclass of gimple_opt_pass along with...
13341 (pass_data_call_cdce): ...new pass_data instance and...
13342 (make_pass_call_cdce): ...new function.
13343 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
13344 subclass of gimple_opt_pass along with...
13345 (pass_data_build_cfg): ...new pass_data instance and...
13346 (make_pass_build_cfg): ...new function.
13347 (pass_split_crit_edges): Convert from a global struct to a subclass of
13348 gimple_opt_pass along with...
13349 (pass_data_split_crit_edges): ...new pass_data instance and...
13350 (make_pass_split_crit_edges): ...new function.
13351 (pass_warn_function_return): Convert from a global struct to a
13352 subclass of gimple_opt_pass along with...
13353 (pass_data_warn_function_return): ...new pass_data instance and...
13354 (make_pass_warn_function_return): ...new function.
13355 (pass_warn_function_noreturn): Convert from a global struct to a
13356 subclass of gimple_opt_pass along with...
13357 (pass_data_warn_function_noreturn): ...new pass_data instance and...
13358 (make_pass_warn_function_noreturn): ...new function.
13359 (pass_warn_unused_result): Convert from a global struct to a subclass
13360 of gimple_opt_pass along with...
13361 (pass_data_warn_unused_result): ...new pass_data instance and...
13362 (make_pass_warn_unused_result): ...new function.
13363 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
13364 a subclass of gimple_opt_pass along with...
13365 (pass_data_merge_phi): ...new pass_data instance and...
13366 (make_pass_merge_phi): ...new function.
13367 * tree-complex.c (pass_lower_complex): Convert from a global struct to
13368 a subclass of gimple_opt_pass along with...
13369 (pass_data_lower_complex): ...new pass_data instance and...
13370 (make_pass_lower_complex): ...new function.
13371 (pass_lower_complex_O0): Convert from a global struct to a subclass of
13372 gimple_opt_pass along with...
13373 (pass_data_lower_complex_O0): ...new pass_data instance and...
13374 (make_pass_lower_complex_O0): ...new function.
13375 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
13376 subclass of gimple_opt_pass along with...
13377 (pass_data_lower_eh): ...new pass_data instance and...
13378 (make_pass_lower_eh): ...new function.
13379 (pass_refactor_eh): Convert from a global struct to a subclass of
13380 gimple_opt_pass along with...
13381 (pass_data_refactor_eh): ...new pass_data instance and...
13382 (make_pass_refactor_eh): ...new function.
13383 (pass_lower_resx): Convert from a global struct to a subclass of
13384 gimple_opt_pass along with...
13385 (pass_data_lower_resx): ...new pass_data instance and...
13386 (make_pass_lower_resx): ...new function.
13387 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
13388 of gimple_opt_pass along with...
13389 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
13390 (make_pass_lower_eh_dispatch): ...new function.
13391 (pass_cleanup_eh): Convert from a global struct to a subclass of
13392 gimple_opt_pass along with...
13393 (pass_data_cleanup_eh): ...new pass_data instance and...
13394 (make_pass_cleanup_eh): ...new function.
13395 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
13396 to a subclass of simple_ipa_opt_pass along with...
13397 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
13398 (make_pass_ipa_lower_emutls): ...new function.
13399 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
13400 a subclass of gimple_opt_pass along with...
13401 (pass_data_if_conversion): ...new pass_data instance and...
13402 (make_pass_if_conversion): ...new function.
13403 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
13404 subclass of gimple_opt_pass along with...
13405 (pass_data_build_ssa): ...new pass_data instance and...
13406 (make_pass_build_ssa): ...new function.
13407 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
13408 global struct to a subclass of gimple_opt_pass along with...
13409 (pass_data_loop_distribution): ...new pass_data instance and...
13410 (make_pass_loop_distribution): ...new function.
13411 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13412 subclass of gimple_opt_pass along with...
13413 (pass_data_mudflap_1): ...new pass_data instance and...
13414 (make_pass_mudflap_1): ...new function.
13415 (pass_mudflap_2): Convert from a global struct to a subclass of
13416 gimple_opt_pass along with...
13417 (pass_data_mudflap_2): ...new pass_data instance and...
13418 (make_pass_mudflap_2): ...new function.
13419 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13420 subclass of gimple_opt_pass along with...
13421 (pass_data_mudflap_1): ...new pass_data instance and...
13422 (make_pass_mudflap_1): ...new function.
13423 (pass_mudflap_2): Convert from a global struct to a subclass of
13424 gimple_opt_pass along with...
13425 (pass_data_mudflap_2): ...new pass_data instance and...
13426 (make_pass_mudflap_2): ...new function.
13427 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13428 gimple_opt_pass along with...
13429 (pass_data_nrv): ...new pass_data instance and...
13430 (make_pass_nrv): ...new function.
13431 (pass_return_slot): Convert from a global struct to a subclass of
13432 gimple_opt_pass along with...
13433 (pass_data_return_slot): ...new pass_data instance and...
13434 (make_pass_return_slot): ...new function.
13435 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13436 to a subclass of gimple_opt_pass along with...
13437 (pass_data_object_sizes): ...new pass_data instance and...
13438 (make_pass_object_sizes): ...new function.
13439 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13440 global struct to a subclass of gimple_opt_pass along with...
13441 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13443 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13444 (pass_fixup_cfg): Convert from a global struct to a subclass of
13445 gimple_opt_pass along with...
13446 (pass_data_fixup_cfg): ...new pass_data instance and...
13447 (make_pass_fixup_cfg): ...new function.
13448 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13449 (make_pass_mudflap_1): ...new function.
13450 (pass_mudflap_2): Replace declaration with that of...
13451 (make_pass_mudflap_2): ...new function.
13452 (pass_asan): Replace declaration with that of...
13453 (make_pass_asan): ...new function.
13454 (pass_asan_O0): Replace declaration with that of...
13455 (make_pass_asan_O0): ...new function.
13456 (pass_tsan): Replace declaration with that of...
13457 (make_pass_tsan): ...new function.
13458 (pass_tsan_O0): Replace declaration with that of...
13459 (make_pass_tsan_O0): ...new function.
13460 (pass_lower_cf): Replace declaration with that of...
13461 (make_pass_lower_cf): ...new function.
13462 (pass_refactor_eh): Replace declaration with that of...
13463 (make_pass_refactor_eh): ...new function.
13464 (pass_lower_eh): Replace declaration with that of...
13465 (make_pass_lower_eh): ...new function.
13466 (pass_lower_eh_dispatch): Replace declaration with that of...
13467 (make_pass_lower_eh_dispatch): ...new function.
13468 (pass_lower_resx): Replace declaration with that of...
13469 (make_pass_lower_resx): ...new function.
13470 (pass_build_cfg): Replace declaration with that of...
13471 (make_pass_build_cfg): ...new function.
13472 (pass_early_tree_profile): Replace declaration with that of...
13473 (make_pass_early_tree_profile): ...new function.
13474 (pass_cleanup_eh): Replace declaration with that of...
13475 (make_pass_cleanup_eh): ...new function.
13476 (pass_sra): Replace declaration with that of...
13477 (make_pass_sra): ...new function.
13478 (pass_sra_early): Replace declaration with that of...
13479 (make_pass_sra_early): ...new function.
13480 (pass_early_ipa_sra): Replace declaration with that of...
13481 (make_pass_early_ipa_sra): ...new function.
13482 (pass_tail_recursion): Replace declaration with that of...
13483 (make_pass_tail_recursion): ...new function.
13484 (pass_tail_calls): Replace declaration with that of...
13485 (make_pass_tail_calls): ...new function.
13486 (pass_tree_loop): Replace declaration with that of...
13487 (make_pass_tree_loop): ...new function.
13488 (pass_tree_loop_init): Replace declaration with that of...
13489 (make_pass_tree_loop_init): ...new function.
13490 (pass_lim): Replace declaration with that of...
13491 (make_pass_lim): ...new function.
13492 (pass_tree_unswitch): Replace declaration with that of...
13493 (make_pass_tree_unswitch): ...new function.
13494 (pass_predcom): Replace declaration with that of...
13495 (make_pass_predcom): ...new function.
13496 (pass_iv_canon): Replace declaration with that of...
13497 (make_pass_iv_canon): ...new function.
13498 (pass_scev_cprop): Replace declaration with that of...
13499 (make_pass_scev_cprop): ...new function.
13500 (pass_empty_loop): Replace declaration with that of...
13501 (make_pass_empty_loop): ...new function.
13502 (pass_record_bounds): Replace declaration with that of...
13503 (make_pass_record_bounds): ...new function.
13504 (pass_graphite): Replace declaration with that of...
13505 (make_pass_graphite): ...new function.
13506 (pass_graphite_transforms): Replace declaration with that of...
13507 (make_pass_graphite_transforms): ...new function.
13508 (pass_if_conversion): Replace declaration with that of...
13509 (make_pass_if_conversion): ...new function.
13510 (pass_loop_distribution): Replace declaration with that of...
13511 (make_pass_loop_distribution): ...new function.
13512 (pass_vectorize): Replace declaration with that of...
13513 (make_pass_vectorize): ...new function.
13514 (pass_slp_vectorize): Replace declaration with that of...
13515 (make_pass_slp_vectorize): ...new function.
13516 (pass_complete_unroll): Replace declaration with that of...
13517 (make_pass_complete_unroll): ...new function.
13518 (pass_complete_unrolli): Replace declaration with that of...
13519 (make_pass_complete_unrolli): ...new function.
13520 (pass_parallelize_loops): Replace declaration with that of...
13521 (make_pass_parallelize_loops): ...new function.
13522 (pass_loop_prefetch): Replace declaration with that of...
13523 (make_pass_loop_prefetch): ...new function.
13524 (pass_iv_optimize): Replace declaration with that of...
13525 (make_pass_iv_optimize): ...new function.
13526 (pass_tree_loop_done): Replace declaration with that of...
13527 (make_pass_tree_loop_done): ...new function.
13528 (pass_ch): Replace declaration with that of...
13529 (make_pass_ch): ...new function.
13530 (pass_ccp): Replace declaration with that of...
13531 (make_pass_ccp): ...new function.
13532 (pass_phi_only_cprop): Replace declaration with that of...
13533 (make_pass_phi_only_cprop): ...new function.
13534 (pass_build_ssa): Replace declaration with that of...
13535 (make_pass_build_ssa): ...new function.
13536 (pass_build_alias): Replace declaration with that of...
13537 (make_pass_build_alias): ...new function.
13538 (pass_build_ealias): Replace declaration with that of...
13539 (make_pass_build_ealias): ...new function.
13540 (pass_dominator): Replace declaration with that of...
13541 (make_pass_dominator): ...new function.
13542 (pass_dce): Replace declaration with that of...
13543 (make_pass_dce): ...new function.
13544 (pass_dce_loop): Replace declaration with that of...
13545 (make_pass_dce_loop): ...new function.
13546 (pass_cd_dce): Replace declaration with that of...
13547 (make_pass_cd_dce): ...new function.
13548 (pass_call_cdce): Replace declaration with that of...
13549 (make_pass_call_cdce): ...new function.
13550 (pass_merge_phi): Replace declaration with that of...
13551 (make_pass_merge_phi): ...new function.
13552 (pass_split_crit_edges): Replace declaration with that of...
13553 (make_pass_split_crit_edges): ...new function.
13554 (pass_pre): Replace declaration with that of...
13555 (make_pass_pre): ...new function.
13556 (pass_profile): Replace declaration with that of...
13557 (make_pass_profile): ...new function.
13558 (pass_strip_predict_hints): Replace declaration with that of...
13559 (make_pass_strip_predict_hints): ...new function.
13560 (pass_lower_complex_O0): Replace declaration with that of...
13561 (make_pass_lower_complex_O0): ...new function.
13562 (pass_lower_complex): Replace declaration with that of...
13563 (make_pass_lower_complex): ...new function.
13564 (pass_lower_vector): Replace declaration with that of...
13565 (make_pass_lower_vector): ...new function.
13566 (pass_lower_vector_ssa): Replace declaration with that of...
13567 (make_pass_lower_vector_ssa): ...new function.
13568 (pass_lower_omp): Replace declaration with that of...
13569 (make_pass_lower_omp): ...new function.
13570 (pass_diagnose_omp_blocks): Replace declaration with that of...
13571 (make_pass_diagnose_omp_blocks): ...new function.
13572 (pass_expand_omp): Replace declaration with that of...
13573 (make_pass_expand_omp): ...new function.
13574 (pass_expand_omp_ssa): Replace declaration with that of...
13575 (make_pass_expand_omp_ssa): ...new function.
13576 (pass_object_sizes): Replace declaration with that of...
13577 (make_pass_object_sizes): ...new function.
13578 (pass_strlen): Replace declaration with that of...
13579 (make_pass_strlen): ...new function.
13580 (pass_fold_builtins): Replace declaration with that of...
13581 (make_pass_fold_builtins): ...new function.
13582 (pass_stdarg): Replace declaration with that of...
13583 (make_pass_stdarg): ...new function.
13584 (pass_early_warn_uninitialized): Replace declaration with that of...
13585 (make_pass_early_warn_uninitialized): ...new function.
13586 (pass_late_warn_uninitialized): Replace declaration with that of...
13587 (make_pass_late_warn_uninitialized): ...new function.
13588 (pass_cse_reciprocals): Replace declaration with that of...
13589 (make_pass_cse_reciprocals): ...new function.
13590 (pass_cse_sincos): Replace declaration with that of...
13591 (make_pass_cse_sincos): ...new function.
13592 (pass_optimize_bswap): Replace declaration with that of...
13593 (make_pass_optimize_bswap): ...new function.
13594 (pass_optimize_widening_mul): Replace declaration with that of...
13595 (make_pass_optimize_widening_mul): ...new function.
13596 (pass_warn_function_return): Replace declaration with that of...
13597 (make_pass_warn_function_return): ...new function.
13598 (pass_warn_function_noreturn): Replace declaration with that of...
13599 (make_pass_warn_function_noreturn): ...new function.
13600 (pass_cselim): Replace declaration with that of...
13601 (make_pass_cselim): ...new function.
13602 (pass_phiopt): Replace declaration with that of...
13603 (make_pass_phiopt): ...new function.
13604 (pass_forwprop): Replace declaration with that of...
13605 (make_pass_forwprop): ...new function.
13606 (pass_phiprop): Replace declaration with that of...
13607 (make_pass_phiprop): ...new function.
13608 (pass_tree_ifcombine): Replace declaration with that of...
13609 (make_pass_tree_ifcombine): ...new function.
13610 (pass_dse): Replace declaration with that of...
13611 (make_pass_dse): ...new function.
13612 (pass_nrv): Replace declaration with that of...
13613 (make_pass_nrv): ...new function.
13614 (pass_rename_ssa_copies): Replace declaration with that of...
13615 (make_pass_rename_ssa_copies): ...new function.
13616 (pass_sink_code): Replace declaration with that of...
13617 (make_pass_sink_code): ...new function.
13618 (pass_fre): Replace declaration with that of...
13619 (make_pass_fre): ...new function.
13620 (pass_check_data_deps): Replace declaration with that of...
13621 (make_pass_check_data_deps): ...new function.
13622 (pass_copy_prop): Replace declaration with that of...
13623 (make_pass_copy_prop): ...new function.
13624 (pass_vrp): Replace declaration with that of...
13625 (make_pass_vrp): ...new function.
13626 (pass_uncprop): Replace declaration with that of...
13627 (make_pass_uncprop): ...new function.
13628 (pass_return_slot): Replace declaration with that of...
13629 (make_pass_return_slot): ...new function.
13630 (pass_reassoc): Replace declaration with that of...
13631 (make_pass_reassoc): ...new function.
13632 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13633 (make_pass_rebuild_cgraph_edges): ...new function.
13634 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13635 (make_pass_remove_cgraph_callee_edges): ...new function.
13636 (pass_build_cgraph_edges): Replace declaration with that of...
13637 (make_pass_build_cgraph_edges): ...new function.
13638 (pass_local_pure_const): Replace declaration with that of...
13639 (make_pass_local_pure_const): ...new function.
13640 (pass_tracer): Replace declaration with that of...
13641 (make_pass_tracer): ...new function.
13642 (pass_warn_unused_result): Replace declaration with that of...
13643 (make_pass_warn_unused_result): ...new function.
13644 (pass_diagnose_tm_blocks): Replace declaration with that of...
13645 (make_pass_diagnose_tm_blocks): ...new function.
13646 (pass_lower_tm): Replace declaration with that of...
13647 (make_pass_lower_tm): ...new function.
13648 (pass_tm_init): Replace declaration with that of...
13649 (make_pass_tm_init): ...new function.
13650 (pass_tm_mark): Replace declaration with that of...
13651 (make_pass_tm_mark): ...new function.
13652 (pass_tm_memopt): Replace declaration with that of...
13653 (make_pass_tm_memopt): ...new function.
13654 (pass_tm_edges): Replace declaration with that of...
13655 (make_pass_tm_edges): ...new function.
13656 (pass_split_functions): Replace declaration with that of...
13657 (make_pass_split_functions): ...new function.
13658 (pass_feedback_split_functions): Replace declaration with that of...
13659 (make_pass_feedback_split_functions): ...new function.
13660 (pass_strength_reduction): Replace declaration with that of...
13661 (make_pass_strength_reduction): ...new function.
13662 (pass_ipa_lower_emutls): Replace declaration with that of...
13663 (make_pass_ipa_lower_emutls): ...new function.
13664 (pass_ipa_function_and_variable_visibility): Replace declaration with
13666 (make_pass_ipa_function_and_variable_visibility): ...new function.
13667 (pass_ipa_tree_profile): Replace declaration with that of...
13668 (make_pass_ipa_tree_profile): ...new function.
13669 (pass_early_local_passes): Replace declaration with that of...
13670 (make_pass_early_local_passes): ...new function.
13671 (pass_ipa_whole_program_visibility): Replace declaration with that
13673 (make_pass_ipa_whole_program_visibility): ...new function.
13674 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13675 (make_pass_ipa_lto_gimple_out): ...new function.
13676 (pass_ipa_increase_alignment): Replace declaration with that of...
13677 (make_pass_ipa_increase_alignment): ...new function.
13678 (pass_ipa_inline): Replace declaration with that of...
13679 (make_pass_ipa_inline): ...new function.
13680 (pass_ipa_free_lang_data): Replace declaration with that of...
13681 (make_pass_ipa_free_lang_data): ...new function.
13682 (pass_ipa_free_inline_summary): Replace declaration with that of...
13683 (make_pass_ipa_free_inline_summary): ...new function.
13684 (pass_ipa_cp): Replace declaration with that of...
13685 (make_pass_ipa_cp): ...new function.
13686 (pass_ipa_reference): Replace declaration with that of...
13687 (make_pass_ipa_reference): ...new function.
13688 (pass_ipa_pure_const): Replace declaration with that of...
13689 (make_pass_ipa_pure_const): ...new function.
13690 (pass_ipa_pta): Replace declaration with that of...
13691 (make_pass_ipa_pta): ...new function.
13692 (pass_ipa_lto_finish_out): Replace declaration with that of...
13693 (make_pass_ipa_lto_finish_out): ...new function.
13694 (pass_ipa_tm): Replace declaration with that of...
13695 (make_pass_ipa_tm): ...new function.
13696 (pass_ipa_profile): Replace declaration with that of...
13697 (make_pass_ipa_profile): ...new function.
13698 (pass_ipa_cdtor_merge): Replace declaration with that of...
13699 (make_pass_ipa_cdtor_merge): ...new function.
13700 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13702 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13703 (pass_init_datastructures): Replace declaration with that of...
13704 (make_pass_init_datastructures): ...new function.
13705 (pass_fixup_cfg): Replace declaration with that of...
13706 (make_pass_fixup_cfg): ...new function.
13707 (pass_expand): Replace declaration with that of...
13708 (make_pass_expand): ...new function.
13709 (pass_instantiate_virtual_regs): Replace declaration with that of...
13710 (make_pass_instantiate_virtual_regs): ...new function.
13711 (pass_rtl_fwprop): Replace declaration with that of...
13712 (make_pass_rtl_fwprop): ...new function.
13713 (pass_rtl_fwprop_addr): Replace declaration with that of...
13714 (make_pass_rtl_fwprop_addr): ...new function.
13715 (pass_jump): Replace declaration with that of...
13716 (make_pass_jump): ...new function.
13717 (pass_jump2): Replace declaration with that of...
13718 (make_pass_jump2): ...new function.
13719 (pass_lower_subreg): Replace declaration with that of...
13720 (make_pass_lower_subreg): ...new function.
13721 (pass_cse): Replace declaration with that of...
13722 (make_pass_cse): ...new function.
13723 (pass_fast_rtl_dce): Replace declaration with that of...
13724 (make_pass_fast_rtl_dce): ...new function.
13725 (pass_ud_rtl_dce): Replace declaration with that of...
13726 (make_pass_ud_rtl_dce): ...new function.
13727 (pass_rtl_dce): Replace declaration with that of...
13728 (make_pass_rtl_dce): ...new function.
13729 (pass_rtl_dse1): Replace declaration with that of...
13730 (make_pass_rtl_dse1): ...new function.
13731 (pass_rtl_dse2): Replace declaration with that of...
13732 (make_pass_rtl_dse2): ...new function.
13733 (pass_rtl_dse3): Replace declaration with that of...
13734 (make_pass_rtl_dse3): ...new function.
13735 (pass_rtl_cprop): Replace declaration with that of...
13736 (make_pass_rtl_cprop): ...new function.
13737 (pass_rtl_pre): Replace declaration with that of...
13738 (make_pass_rtl_pre): ...new function.
13739 (pass_rtl_hoist): Replace declaration with that of...
13740 (make_pass_rtl_hoist): ...new function.
13741 (pass_rtl_store_motion): Replace declaration with that of...
13742 (make_pass_rtl_store_motion): ...new function.
13743 (pass_cse_after_global_opts): Replace declaration with that of...
13744 (make_pass_cse_after_global_opts): ...new function.
13745 (pass_rtl_ifcvt): Replace declaration with that of...
13746 (make_pass_rtl_ifcvt): ...new function.
13747 (pass_into_cfg_layout_mode): Replace declaration with that of...
13748 (make_pass_into_cfg_layout_mode): ...new function.
13749 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13750 (make_pass_outof_cfg_layout_mode): ...new function.
13751 (pass_loop2): Replace declaration with that of...
13752 (make_pass_loop2): ...new function.
13753 (pass_rtl_loop_init): Replace declaration with that of...
13754 (make_pass_rtl_loop_init): ...new function.
13755 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13756 (make_pass_rtl_move_loop_invariants): ...new function.
13757 (pass_rtl_unswitch): Replace declaration with that of...
13758 (make_pass_rtl_unswitch): ...new function.
13759 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13760 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13761 (pass_rtl_doloop): Replace declaration with that of...
13762 (make_pass_rtl_doloop): ...new function.
13763 (pass_rtl_loop_done): Replace declaration with that of...
13764 (make_pass_rtl_loop_done): ...new function.
13765 (pass_web): Replace declaration with that of...
13766 (make_pass_web): ...new function.
13767 (pass_cse2): Replace declaration with that of...
13768 (make_pass_cse2): ...new function.
13769 (pass_df_initialize_opt): Replace declaration with that of...
13770 (make_pass_df_initialize_opt): ...new function.
13771 (pass_df_initialize_no_opt): Replace declaration with that of...
13772 (make_pass_df_initialize_no_opt): ...new function.
13773 (pass_reginfo_init): Replace declaration with that of...
13774 (make_pass_reginfo_init): ...new function.
13775 (pass_inc_dec): Replace declaration with that of...
13776 (make_pass_inc_dec): ...new function.
13777 (pass_stack_ptr_mod): Replace declaration with that of...
13778 (make_pass_stack_ptr_mod): ...new function.
13779 (pass_initialize_regs): Replace declaration with that of...
13780 (make_pass_initialize_regs): ...new function.
13781 (pass_combine): Replace declaration with that of...
13782 (make_pass_combine): ...new function.
13783 (pass_if_after_combine): Replace declaration with that of...
13784 (make_pass_if_after_combine): ...new function.
13785 (pass_ree): Replace declaration with that of...
13786 (make_pass_ree): ...new function.
13787 (pass_partition_blocks): Replace declaration with that of...
13788 (make_pass_partition_blocks): ...new function.
13789 (pass_match_asm_constraints): Replace declaration with that of...
13790 (make_pass_match_asm_constraints): ...new function.
13791 (pass_regmove): Replace declaration with that of...
13792 (make_pass_regmove): ...new function.
13793 (pass_split_all_insns): Replace declaration with that of...
13794 (make_pass_split_all_insns): ...new function.
13795 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13796 (make_pass_fast_rtl_byte_dce): ...new function.
13797 (pass_lower_subreg2): Replace declaration with that of...
13798 (make_pass_lower_subreg2): ...new function.
13799 (pass_mode_switching): Replace declaration with that of...
13800 (make_pass_mode_switching): ...new function.
13801 (pass_sms): Replace declaration with that of...
13802 (make_pass_sms): ...new function.
13803 (pass_sched): Replace declaration with that of...
13804 (make_pass_sched): ...new function.
13805 (pass_ira): Replace declaration with that of...
13806 (make_pass_ira): ...new function.
13807 (pass_reload): Replace declaration with that of...
13808 (make_pass_reload): ...new function.
13809 (pass_clean_state): Replace declaration with that of...
13810 (make_pass_clean_state): ...new function.
13811 (pass_branch_prob): Replace declaration with that of...
13812 (make_pass_branch_prob): ...new function.
13813 (pass_value_profile_transformations): Replace declaration with that
13815 (make_pass_value_profile_transformations): ...new function.
13816 (pass_postreload_cse): Replace declaration with that of...
13817 (make_pass_postreload_cse): ...new function.
13818 (pass_gcse2): Replace declaration with that of...
13819 (make_pass_gcse2): ...new function.
13820 (pass_split_after_reload): Replace declaration with that of...
13821 (make_pass_split_after_reload): ...new function.
13822 (pass_branch_target_load_optimize1): Replace declaration with that
13824 (make_pass_branch_target_load_optimize1): ...new function.
13825 (pass_thread_prologue_and_epilogue): Replace declaration with that
13827 (make_pass_thread_prologue_and_epilogue): ...new function.
13828 (pass_stack_adjustments): Replace declaration with that of...
13829 (make_pass_stack_adjustments): ...new function.
13830 (pass_peephole2): Replace declaration with that of...
13831 (make_pass_peephole2): ...new function.
13832 (pass_if_after_reload): Replace declaration with that of...
13833 (make_pass_if_after_reload): ...new function.
13834 (pass_regrename): Replace declaration with that of...
13835 (make_pass_regrename): ...new function.
13836 (pass_cprop_hardreg): Replace declaration with that of...
13837 (make_pass_cprop_hardreg): ...new function.
13838 (pass_reorder_blocks): Replace declaration with that of...
13839 (make_pass_reorder_blocks): ...new function.
13840 (pass_branch_target_load_optimize2): Replace declaration with that
13842 (make_pass_branch_target_load_optimize2): ...new function.
13843 (pass_leaf_regs): Replace declaration with that of...
13844 (make_pass_leaf_regs): ...new function.
13845 (pass_split_before_sched2): Replace declaration with that of...
13846 (make_pass_split_before_sched2): ...new function.
13847 (pass_compare_elim_after_reload): Replace declaration with that of...
13848 (make_pass_compare_elim_after_reload): ...new function.
13849 (pass_sched2): Replace declaration with that of...
13850 (make_pass_sched2): ...new function.
13851 (pass_stack_regs): Replace declaration with that of...
13852 (make_pass_stack_regs): ...new function.
13853 (pass_stack_regs_run): Replace declaration with that of...
13854 (make_pass_stack_regs_run): ...new function.
13855 (pass_df_finish): Replace declaration with that of...
13856 (make_pass_df_finish): ...new function.
13857 (pass_compute_alignments): Replace declaration with that of...
13858 (make_pass_compute_alignments): ...new function.
13859 (pass_duplicate_computed_gotos): Replace declaration with that of...
13860 (make_pass_duplicate_computed_gotos): ...new function.
13861 (pass_variable_tracking): Replace declaration with that of...
13862 (make_pass_variable_tracking): ...new function.
13863 (pass_free_cfg): Replace declaration with that of...
13864 (make_pass_free_cfg): ...new function.
13865 (pass_machine_reorg): Replace declaration with that of...
13866 (make_pass_machine_reorg): ...new function.
13867 (pass_cleanup_barriers): Replace declaration with that of...
13868 (make_pass_cleanup_barriers): ...new function.
13869 (pass_delay_slots): Replace declaration with that of...
13870 (make_pass_delay_slots): ...new function.
13871 (pass_split_for_shorten_branches): Replace declaration with that of...
13872 (make_pass_split_for_shorten_branches): ...new function.
13873 (pass_split_before_regstack): Replace declaration with that of...
13874 (make_pass_split_before_regstack): ...new function.
13875 (pass_convert_to_eh_region_ranges): Replace declaration with that
13877 (make_pass_convert_to_eh_region_ranges): ...new function.
13878 (pass_shorten_branches): Replace declaration with that of...
13879 (make_pass_shorten_branches): ...new function.
13880 (pass_set_nothrow_function_flags): Replace declaration with that of...
13881 (make_pass_set_nothrow_function_flags): ...new function.
13882 (pass_dwarf2_frame): Replace declaration with that of...
13883 (make_pass_dwarf2_frame): ...new function.
13884 (pass_final): Replace declaration with that of...
13885 (make_pass_final): ...new function.
13886 (pass_rtl_seqabstr): Replace declaration with that of...
13887 (make_pass_rtl_seqabstr): ...new function.
13888 (pass_release_ssa_names): Replace declaration with that of...
13889 (make_pass_release_ssa_names): ...new function.
13890 (pass_early_inline): Replace declaration with that of...
13891 (make_pass_early_inline): ...new function.
13892 (pass_inline_parameters): Replace declaration with that of...
13893 (make_pass_inline_parameters): ...new function.
13894 (pass_update_address_taken): Replace declaration with that of...
13895 (make_pass_update_address_taken): ...new function.
13896 (pass_convert_switch): Replace declaration with that of...
13897 (make_pass_convert_switch): ...new function.
13898 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13899 to a subclass of simple_ipa_opt_pass along with...
13900 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13901 (make_pass_ipa_tree_profile): ...new function.
13902 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13903 subclass of gimple_opt_pass along with...
13904 (pass_data_sra_early): ...new pass_data instance and...
13905 (make_pass_sra_early): ...new function.
13906 (pass_sra): Convert from a global struct to a subclass of
13907 gimple_opt_pass along with...
13908 (pass_data_sra): ...new pass_data instance and...
13909 (make_pass_sra): ...new function.
13910 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13911 gimple_opt_pass along with...
13912 (pass_data_early_ipa_sra): ...new pass_data instance and...
13913 (make_pass_early_ipa_sra): ...new function.
13914 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13915 subclass of gimple_opt_pass along with...
13916 (pass_data_ccp): ...new pass_data instance and...
13917 (make_pass_ccp): ...new function.
13918 (pass_fold_builtins): Convert from a global struct to a subclass of
13919 gimple_opt_pass along with...
13920 (pass_data_fold_builtins): ...new pass_data instance and...
13921 (make_pass_fold_builtins): ...new function.
13922 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13923 subclass of gimple_opt_pass along with...
13924 (pass_data_copy_prop): ...new pass_data instance and...
13925 (make_pass_copy_prop): ...new function.
13926 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13927 global struct to a subclass of gimple_opt_pass along with...
13928 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13929 (make_pass_rename_ssa_copies): ...new function.
13930 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13931 subclass of gimple_opt_pass along with...
13932 (pass_data_dce): ...new pass_data instance and...
13933 (make_pass_dce): ...new function.
13934 (pass_dce_loop): Convert from a global struct to a subclass of
13935 gimple_opt_pass along with...
13936 (pass_data_dce_loop): ...new pass_data instance and...
13937 (make_pass_dce_loop): ...new function.
13938 (pass_cd_dce): Convert from a global struct to a subclass of
13939 gimple_opt_pass along with...
13940 (pass_data_cd_dce): ...new pass_data instance and...
13941 (make_pass_cd_dce): ...new function.
13942 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13943 subclass of gimple_opt_pass along with...
13944 (pass_data_dominator): ...new pass_data instance and...
13945 (make_pass_dominator): ...new function.
13946 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13947 gimple_opt_pass along with...
13948 (pass_data_phi_only_cprop): ...new pass_data instance and...
13949 (make_pass_phi_only_cprop): ...new function.
13950 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13951 subclass of gimple_opt_pass along with...
13952 (pass_data_dse): ...new pass_data instance and...
13953 (make_pass_dse): ...new function.
13954 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13955 a subclass of gimple_opt_pass along with...
13956 (pass_data_forwprop): ...new pass_data instance and...
13957 (make_pass_forwprop): ...new function.
13958 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13959 struct to a subclass of gimple_opt_pass along with...
13960 (pass_data_tree_ifcombine): ...new pass_data instance and...
13961 (make_pass_tree_ifcombine): ...new function.
13962 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13963 subclass of gimple_opt_pass along with...
13964 (pass_data_ch): ...new pass_data instance and...
13965 (make_pass_ch): ...new function.
13966 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13967 subclass of gimple_opt_pass along with...
13968 (pass_data_tree_loop): ...new pass_data instance and...
13969 (make_pass_tree_loop): ...new function.
13970 (pass_tree_loop_init): Convert from a global struct to a subclass of
13971 gimple_opt_pass along with...
13972 (pass_data_tree_loop_init): ...new pass_data instance and...
13973 (make_pass_tree_loop_init): ...new function.
13974 (pass_lim): Convert from a global struct to a subclass of
13975 gimple_opt_pass along with...
13976 (pass_data_lim): ...new pass_data instance and...
13977 (make_pass_lim): ...new function.
13978 (pass_tree_unswitch): Convert from a global struct to a subclass of
13979 gimple_opt_pass along with...
13980 (pass_data_tree_unswitch): ...new pass_data instance and...
13981 (make_pass_tree_unswitch): ...new function.
13982 (pass_predcom): Convert from a global struct to a subclass of
13983 gimple_opt_pass along with...
13984 (pass_data_predcom): ...new pass_data instance and...
13985 (make_pass_predcom): ...new function.
13986 (pass_vectorize): Convert from a global struct to a subclass of
13987 gimple_opt_pass along with...
13988 (pass_data_vectorize): ...new pass_data instance and...
13989 (make_pass_vectorize): ...new function.
13990 (pass_graphite): Convert from a global struct to a subclass of
13991 gimple_opt_pass along with...
13992 (pass_data_graphite): ...new pass_data instance and...
13993 (make_pass_graphite): ...new function.
13994 (pass_graphite_transforms): Convert from a global struct to a subclass
13995 of gimple_opt_pass along with...
13996 (pass_data_graphite_transforms): ...new pass_data instance and...
13997 (make_pass_graphite_transforms): ...new function.
13998 (pass_check_data_deps): Convert from a global struct to a subclass of
13999 gimple_opt_pass along with...
14000 (pass_data_check_data_deps): ...new pass_data instance and...
14001 (make_pass_check_data_deps): ...new function.
14002 (pass_iv_canon): Convert from a global struct to a subclass of
14003 gimple_opt_pass along with...
14004 (pass_data_iv_canon): ...new pass_data instance and...
14005 (make_pass_iv_canon): ...new function.
14006 (pass_scev_cprop): Convert from a global struct to a subclass of
14007 gimple_opt_pass along with...
14008 (pass_data_scev_cprop): ...new pass_data instance and...
14009 (make_pass_scev_cprop): ...new function.
14010 (pass_record_bounds): Convert from a global struct to a subclass of
14011 gimple_opt_pass along with...
14012 (pass_data_record_bounds): ...new pass_data instance and...
14013 (make_pass_record_bounds): ...new function.
14014 (pass_complete_unroll): Convert from a global struct to a subclass of
14015 gimple_opt_pass along with...
14016 (pass_data_complete_unroll): ...new pass_data instance and...
14017 (make_pass_complete_unroll): ...new function.
14018 (pass_complete_unrolli): Convert from a global struct to a subclass of
14019 gimple_opt_pass along with...
14020 (pass_data_complete_unrolli): ...new pass_data instance and...
14021 (make_pass_complete_unrolli): ...new function.
14022 (pass_parallelize_loops): Convert from a global struct to a subclass
14023 of gimple_opt_pass along with...
14024 (pass_data_parallelize_loops): ...new pass_data instance and...
14025 (make_pass_parallelize_loops): ...new function.
14026 (pass_loop_prefetch): Convert from a global struct to a subclass of
14027 gimple_opt_pass along with...
14028 (pass_data_loop_prefetch): ...new pass_data instance and...
14029 (make_pass_loop_prefetch): ...new function.
14030 (pass_iv_optimize): Convert from a global struct to a subclass of
14031 gimple_opt_pass along with...
14032 (pass_data_iv_optimize): ...new pass_data instance and...
14033 (make_pass_iv_optimize): ...new function.
14034 (pass_tree_loop_done): Convert from a global struct to a subclass of
14035 gimple_opt_pass along with...
14036 (pass_data_tree_loop_done): ...new pass_data instance and...
14037 (make_pass_tree_loop_done): ...new function.
14038 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
14039 struct to a subclass of gimple_opt_pass along with...
14040 (pass_data_cse_reciprocals): ...new pass_data instance and...
14041 (make_pass_cse_reciprocals): ...new function.
14042 (pass_cse_sincos): Convert from a global struct to a subclass of
14043 gimple_opt_pass along with...
14044 (pass_data_cse_sincos): ...new pass_data instance and...
14045 (make_pass_cse_sincos): ...new function.
14046 (pass_optimize_bswap): Convert from a global struct to a subclass of
14047 gimple_opt_pass along with...
14048 (pass_data_optimize_bswap): ...new pass_data instance and...
14049 (make_pass_optimize_bswap): ...new function.
14050 (pass_optimize_widening_mul): Convert from a global struct to a
14051 subclass of gimple_opt_pass along with...
14052 (pass_data_optimize_widening_mul): ...new pass_data instance and...
14053 (make_pass_optimize_widening_mul): ...new function.
14054 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
14055 subclass of gimple_opt_pass along with...
14056 (pass_data_phiopt): ...new pass_data instance and...
14057 (make_pass_phiopt): ...new function.
14058 (pass_cselim): Convert from a global struct to a subclass of
14059 gimple_opt_pass along with...
14060 (pass_data_cselim): ...new pass_data instance and...
14061 (make_pass_cselim): ...new function.
14062 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
14063 subclass of gimple_opt_pass along with...
14064 (pass_data_phiprop): ...new pass_data instance and...
14065 (make_pass_phiprop): ...new function.
14066 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
14067 subclass of gimple_opt_pass along with...
14068 (pass_data_pre): ...new pass_data instance and...
14069 (make_pass_pre): ...new function.
14070 (pass_fre): Convert from a global struct to a subclass of
14071 gimple_opt_pass along with...
14072 (pass_data_fre): ...new pass_data instance and...
14073 (make_pass_fre): ...new function.
14074 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
14075 subclass of gimple_opt_pass along with...
14076 (pass_data_reassoc): ...new pass_data instance and...
14077 (make_pass_reassoc): ...new function.
14078 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
14079 subclass of gimple_opt_pass along with...
14080 (pass_data_sink_code): ...new pass_data instance and...
14081 (make_pass_sink_code): ...new function.
14082 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
14083 subclass of gimple_opt_pass along with...
14084 (pass_data_strlen): ...new pass_data instance and...
14085 (make_pass_strlen): ...new function.
14086 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
14087 struct to a subclass of gimple_opt_pass along with...
14088 (pass_data_build_alias): ...new pass_data instance and...
14089 (make_pass_build_alias): ...new function.
14090 (pass_build_ealias): Convert from a global struct to a subclass of
14091 gimple_opt_pass along with...
14092 (pass_data_build_ealias): ...new pass_data instance and...
14093 (make_pass_build_ealias): ...new function.
14094 (pass_ipa_pta): Convert from a global struct to a subclass of
14095 simple_ipa_opt_pass along with...
14096 (pass_data_ipa_pta): ...new pass_data instance and...
14097 (make_pass_ipa_pta): ...new function.
14098 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
14099 subclass of gimple_opt_pass along with...
14100 (pass_data_uncprop): ...new pass_data instance and...
14101 (make_pass_uncprop): ...new function.
14102 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
14103 global struct to a subclass of gimple_opt_pass along with...
14104 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
14105 (make_pass_late_warn_uninitialized): ...new function.
14106 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
14107 to a subclass of gimple_opt_pass along with...
14108 (pass_data_init_datastructures): ...new pass_data instance and...
14109 (make_pass_init_datastructures): ...new function.
14110 (pass_early_warn_uninitialized): Convert from a global struct to a
14111 subclass of gimple_opt_pass along with...
14112 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
14113 (make_pass_early_warn_uninitialized): ...new function.
14114 (pass_update_address_taken): Convert from a global struct to a
14115 subclass of gimple_opt_pass along with...
14116 (pass_data_update_address_taken): ...new pass_data instance and...
14117 (make_pass_update_address_taken): ...new function.
14118 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
14119 struct to a subclass of gimple_opt_pass along with...
14120 (pass_data_release_ssa_names): ...new pass_data instance and...
14121 (make_pass_release_ssa_names): ...new function.
14122 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
14123 subclass of gimple_opt_pass along with...
14124 (pass_data_stdarg): ...new pass_data instance and...
14125 (make_pass_stdarg): ...new function.
14126 * tree-switch-conversion.c (pass_convert_switch): Convert from a
14127 global struct to a subclass of gimple_opt_pass along with...
14128 (pass_data_convert_switch): ...new pass_data instance and...
14129 (make_pass_convert_switch): ...new function.
14130 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
14131 to a subclass of gimple_opt_pass along with...
14132 (pass_data_tail_recursion): ...new pass_data instance and...
14133 (make_pass_tail_recursion): ...new function.
14134 (pass_tail_calls): Convert from a global struct to a subclass of
14135 gimple_opt_pass along with...
14136 (pass_data_tail_calls): ...new pass_data instance and...
14137 (make_pass_tail_calls): ...new function.
14138 * tree-vect-generic.c (pass_lower_vector): Convert from a global
14139 struct to a subclass of gimple_opt_pass along with...
14140 (pass_data_lower_vector): ...new pass_data instance and...
14141 (make_pass_lower_vector): ...new function.
14142 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
14143 gimple_opt_pass along with...
14144 (pass_data_lower_vector_ssa): ...new pass_data instance and...
14145 (make_pass_lower_vector_ssa): ...new function.
14146 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
14147 to a subclass of gimple_opt_pass along with...
14148 (pass_data_slp_vectorize): ...new pass_data instance and...
14149 (make_pass_slp_vectorize): ...new function.
14150 (pass_ipa_increase_alignment): Convert from a global struct to a
14151 subclass of simple_ipa_opt_pass along with...
14152 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
14153 (make_pass_ipa_increase_alignment): ...new function.
14154 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
14155 gimple_opt_pass along with...
14156 (pass_data_vrp): ...new pass_data instance and...
14157 (make_pass_vrp): ...new function.
14158 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
14159 subclass of simple_ipa_opt_pass along with...
14160 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
14161 (make_pass_ipa_free_lang_data): ...new function.
14162 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
14163 gimple_opt_pass along with...
14164 (pass_data_tsan): ...new pass_data instance and...
14165 (make_pass_tsan): ...new function.
14166 (pass_tsan_O0): Convert from a global struct to a subclass of
14167 gimple_opt_pass along with...
14168 (pass_data_tsan_O0): ...new pass_data instance and...
14169 (make_pass_tsan_O0): ...new function.
14170 * var-tracking.c (pass_variable_tracking): Convert from a global
14171 struct to a subclass of rtl_opt_pass along with...
14172 (pass_data_variable_tracking): ...new pass_data instance and...
14173 (make_pass_variable_tracking): ...new function.
14174 * web.c (pass_web): Convert from a global struct to a subclass of
14175 rtl_opt_pass along with...
14176 (pass_data_web): ...new pass_data instance and...
14177 (make_pass_web): ...new function.
14178 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
14179 declaration with that of...
14180 (make_pass_mode_switch_use): ...new function.
14181 (pass_resolve_sw_modes): Replace declaration with that of...
14182 (make_pass_resolve_sw_modes): ...new function.
14183 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
14184 from a global struct to a subclass of rtl_opt_pass along with...
14185 (pass_data_mode_switch_use): ...new pass_data instance and...
14186 (make_pass_mode_switch_use): ...new function.
14187 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
14188 from a global struct to a subclass of rtl_opt_pass along with...
14189 (pass_data_resolve_sw_modes): ...new pass_data instance and...
14190 (make_pass_resolve_sw_modes): ...new function.
14191 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
14192 struct to a subclass of rtl_opt_pass along with...
14193 (pass_data_insert_vzeroupper): ...new pass_data instance and...
14194 (make_pass_insert_vzeroupper): ...new function.
14195 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
14196 global struct to a subclass of rtl_opt_pass along with...
14197 (pass_data_work_around_errata): ...new pass_data instance and...
14198 (make_pass_work_around_errata): ...new function.
14199 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
14200 struct to a subclass of rtl_opt_pass along with...
14201 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
14202 (make_pass_mips_machine_reorg2): ...new function.
14204 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14206 * passes.c (pass_manager::operator new): New.
14208 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14210 Handwritten part of conversion of passes to C++ classes.
14212 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
14213 (toplev.o): Add dep on PASS_MANAGER_H.
14214 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
14215 of early local pases to reflect this moving from a global to a
14216 member of gcc::pass_manager.
14217 (cgraph_add_new_function): Likewise.
14218 * lto-cgraph.c (lto_output_node): Update for conversion of
14219 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
14220 * passes.c (opt_pass::clone): New.
14221 (opt_pass::gate): New.
14222 (opt_pass::execute): New.
14223 (opt_pass::opt_pass): New.
14224 (pass_manager::execute_early_local_passes): New.
14225 (pass_manager::execute_pass_mode_switching): new.
14226 (finish_optimization_passes): Convert to...
14227 (pass_manager::finish_optimization_passes): ...this.
14228 (finish_optimization_passes): Update for conversion of passes to
14230 (register_dump_files_1): Use has_gate since we cannot portably
14231 check a vtable entry against NULL.
14232 (dump_one_pass): Likewise.
14233 (ipa_write_summaries_2): Likewise.
14234 (ipa_write_optimization_summaries_1): Likewise.
14235 (ipa_read_summaries_1): Likewise.
14236 (ipa_read_optimization_summaries_1): Likewise.
14237 (execute_ipa_stmt_fixups): Likewise.
14238 (pass_manager::pass_manager): Rewrite pass-creation, invoking
14239 pass-creation functions rather than wiring up globals, and
14240 storing the results in fields of pass_manager generated using
14241 pass-instances.def.
14242 (pass_manager::dump_profile_report): Update for conversion of
14243 passes to C++ classes.
14244 (pass_manager::execute_ipa_summary_passes): Likewise.
14245 (execute_one_ipa_transform_pass): Likewise.
14246 (execute_one_pass): Use has_gate and has_execute since we cannot
14247 portably check a vtable entry against NULL.
14248 * pass_manager.h (pass_manager::finish_optimization_passes): New.
14249 (pass_manager): Use pass-instances.def to add fields for the
14250 various pass instances.
14251 * toplev.c (finalize): Update for move of
14252 finish_optimization_passes to a method of gcc::pass_manager.
14253 * toplev.h (finish_optimization_passes): Move to method of class
14255 * tree-pass.h (struct pass_data): New.
14256 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
14257 (opt_pass::gate): Convert to virtual function.
14258 (opt_pass::~opt_pass): New.
14259 (opt_pass::clone): New.
14260 (opt_pass::execute): Convert to virtual function.
14261 (opt_pass::opt_pass): New.
14262 (opt_pass::ctxt_): new.
14263 (gimple_opt_pass): Convert to subclass of opt_pass.
14264 (gimple_opt_pass::gimple_opt_pass): New.
14265 (rtl_opt_pass): Convert to subclass of opt_pass.
14266 (rtl_opt_pass::rtl_opt_pass): New.
14267 (ipa_opt_pass_d): Convert to subclass of opt_pass.
14268 (ipa_opt_pass_d::ipa_opt_pass_d): New.
14269 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
14270 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
14271 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
14272 invocation of pass_mode_switching to reflect this moving from a
14273 global to a member of gcc::pass_manager.
14274 (ix86_option_override): Rework how pass_insert_vzeroupper is
14275 added to the pass_manager to reflect autogenerated changes.
14276 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
14278 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
14280 PR rtl-optimization/57708
14281 * recog.c (peep2_find_free_register): Validate all regs in a
14284 2013-08-05 Jan Hubicka <jh@suse.cz>
14287 * cgraph.c (verify_cgraph_node): Accept local flags from other
14289 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
14290 (function_and_variable_visibility): Likewise.
14291 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
14293 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14295 * graph.c (init_graph_slim_pretty_print): Remove.
14296 (print_graph_cfg): Do not call it. Use local pretty printer.
14297 (start_graph_dump): Likewise.
14299 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14301 * gimple-pretty-print.c (buffer): Remove.
14302 (initialized): Likewise.
14303 (maybe_init_pretty_print): Likewise.
14304 (print_gimple_stmt): Do not call it. Use non-static local
14305 pretty_printer variable.
14306 (print_gimple_expr): Likewise.
14307 (print_gimple_seq): Likewise.
14308 (gimple_dump_bb): Likewise.
14310 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14312 * asan.c (asan_pp): Remove.
14313 (asan_pp_initialized): Likewise.
14314 (asan_pp_initialize): Likewise.
14315 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
14316 (asan_emit_stack_protection): Tidy. Use local pretty printer.
14317 (asan_add_global): Likewise.
14319 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
14321 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
14322 * diagnostic.h (diagnostic_flush_buffer): Adjust.
14323 * pretty-print.c (pp_formatted_text_data): Likewise.
14324 (pp_indent): Rename from pp_base_indent.
14325 (pp_format): Rename from pp_base_format.
14326 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
14327 (pp_format_verbatim): Rename from pp_base_format_verbatim.
14328 (pp_flush): Rename from pp_base_flush.
14329 (pp_set_line_maximum_length): Rename from
14330 pp_base_set_line_maximum_length.
14331 (pp_clear_output_area): Rename from pp_base_clear_output_area.
14332 (pp_set_prefix): Rename from pp_base_set_prefix.
14333 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
14334 (pp_emit_prefix): Rename from pp_base_emit_prefix.
14335 (pp_append_text): Rename from pp_base_append_text.
14336 (pp_formatted_text): Rename from pp_base_formatted_text.
14337 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
14338 (pp_remaining_character_count_for_line): Rename from
14339 pp_base_remaining_character_count_for_line.
14340 (pp_newline): Rename from pp_base_newline.
14341 (pp_character): Rename from pp_base_character.
14342 (pp_string): Rename from pp_base_string.
14343 (pp_maybe_space): Rename from pp_base_maybe_space.
14344 * asan.c (asan_pp_string): Adjust.
14345 (asan_emit_stack_protection): Likewise.
14346 (asan_add_global): Likewise.
14347 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
14348 * tree-mudflap.c (mf_varname_tree): Likewise.
14349 * tree-pretty-print.c (pp_tree_identifier): Rename from
14350 pp_base_tree_identifier.
14351 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
14352 Declare as function.
14354 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14356 * pretty-print.h (pp_bar_bar): New.
14357 (pp_ampersand_ampersand): Likewise.
14358 (pp_less_equal): Likewise.
14359 (pp_greater_equal): Likewise.
14360 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
14361 printer functions instead of pp_string or operators and punctuators.
14362 (dump_gimple_call): Likewise.
14363 (dump_gimple_omp_for): Likewise.
14364 (dump_gimple_transaction): Likewise.
14365 (dump_gimple_phi): Likewise.
14366 (pp_gimple_stmt_1): Likewise.
14367 * sched-vis.c (print_insn): Likewise.
14368 * tree-mudflap.c (mf_varname_tree): Likewise.
14369 * tree-pretty-print.c (dump_block_node): Likewise.
14370 (dump_generic_node): Likewise.
14372 2013-08-02 Jan Hubicka <jh@suse.cz>
14374 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
14376 * lto-streamer-out.c (tree_is_indexable): Results decls and
14377 parm decls are not indexable.
14378 (DFS_write_tree_body): Do not follow args and results.
14379 (hash_tree): Likewise.
14380 (output_functions): Rearrange so struct function is needed
14381 only when real body is output; be able to also ouptut abstract
14382 functions; output DECL_ARGUMENTS and DECL_RESULT.
14383 (lto_output): When not in WPA, ale store abstract functions.
14384 (write_symbol): Do not care about RESULT_DECL.
14385 (output_symbol_p): Handle correctly sbtract decls.
14386 * lto-streamer-in.c (input_function): Rearrange so struct
14387 function can be NULL at entry; allow streaming of
14388 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
14389 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
14390 sanity check during LTO.
14391 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
14392 RESULT_DECl and DECL_ARGUMENTS.
14393 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
14396 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14398 * pretty-print.h (pp_underscore): New.
14400 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
14401 printer functions instead of pp_character.
14402 (dump_binary_rhs): Likewise.
14403 (dump_ternary_rhs): Likewise.
14404 (dump_gimple_call_args): Likewise.
14405 (pp_points_to_solution): Likewise.
14406 (dump_gimple_call): Likewise.
14407 (dump_gimple_switch): Likewise.
14408 (dump_gimple_cond): Likewise.
14409 (dump_gimple_bind): Likewise.
14410 (dump_gimple_try): Likewise.
14411 (dump_gimple_omp_for): Likewise.
14412 (dump_gimple_omp_continue): Likewise.
14413 (dump_gimple_omp_single): Likewise.
14414 (dump_gimple_omp_sections): Likewise.
14415 (dump_gimple_omp_block): Likewise.
14416 (dump_gimple_omp_critical): Likewise.
14417 (dump_gimple_transaction): Likewise.
14418 (dump_gimple_asm): Likewise.
14419 (dump_gimple_phi): Likewise.
14420 (dump_gimple_omp_parallel): Likewise.
14421 (dump_gimple_omp_task): Likewise.
14422 (dump_gimple_omp_atomic_load): Likewise.
14423 (dump_gimple_omp_atomic_store): Likewise.
14424 (dump_gimple_mem_ops): Likewise.
14425 (pp_gimple_stmt_1): Likewise.
14426 (pp_cfg_jump): Likewise.
14427 (dump_implicit_edges): Likewise.
14428 (gimple_dump_bb_for_graph): Likewise.
14429 * graph.c (draw_cfg_node): Likewise.
14430 * langhooks.c (lhd_print_error_function): Likewise.
14431 * sched-vis.c (print_exp): Likewise.
14432 (print_value): Likewise.
14433 (print_pattern): Likewise.
14434 (print_insn): Likewise.
14435 (rtl_dump_bb_for_graph): Likewise.
14436 * tree-pretty-print.c (dump_function_declaration): Likewise.
14437 (dump_array_domain): Likewise.
14438 (dump_omp_clause): Likewise.
14439 (dump_location): Likewise.
14440 (dump_generic_node): Likewise.
14441 (print_struct_decl): Likewise.
14442 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14444 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14446 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14447 candidate table when replacing a candidate statement.
14448 (replace_rhs_if_not_dup): Likewise.
14449 (replace_one_candidate): Likewise.
14451 2013-08-02 Jan Hubicka <jh@suse.cz>
14452 Martin Liska <marxin.liska@gmail.com>
14454 * cgraphunit.c (add_new_function): Fix logic when adding from
14456 (assemble_thunk): Rename to ...
14457 (expand_thunk); .. this one; export; get it working with
14458 general functions; make produced gimple valid.
14459 * cgraph.h (expand_thunk): Declare.
14461 2013-08-02 Jan Hubicka <jh@suse.cz>
14463 * ipa-cp.c (gather_context_independent_values): Use
14464 ipa_get_param_move_cost.
14465 (get_replacement_map): Remove PARAM; move parameter folding
14467 (create_specialized_node): Update.
14468 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14469 assert that we have gimple body; update move_cost.
14470 (count_formal_params): Assert that we have gimple body.
14471 (ipa_dump_param): New function.
14472 (ipa_alloc_node_params): Break out from ...
14473 (ipa_initialize_node_params): ... here.
14474 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14475 (ipa_write_node_info): Stream move costs.
14476 (ipa_read_node_info): Read move costs.
14477 (ipa_update_after_lto_read): Do not recompute node params.
14478 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14479 (ipa_get_param): Check we are not in WPA.
14480 (ipa_get_param_move_cost): New.
14481 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14482 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14483 parm numbers to be present.
14485 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14487 PR rtl-optimization/58048
14488 * lra-constraints.c (process_alt_operands): Don't check asm
14489 operand on register.
14491 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14493 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14494 the implied StoreLoad barrier for atomic operations if before.
14496 2013-08-02 Jan Hubicka <jh@suse.cz>
14497 Martin Liska <marxin.liska@gmail.com>
14499 * cgraph.c (cgraph_function_body_availability): Do not check
14501 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14502 symtab_node_availability): Declare.
14503 * ipa.c (can_replace_by_local_alias): New.
14504 (function_and_variable_visibility): Use it.
14505 * symtab.c (symtab_for_node_and_aliases,
14506 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14508 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14510 PR rtl-optimization/57963
14511 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14512 (lra_constraints): Use them.
14514 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14516 * config/arm/types.md (define_attr "type"): Add "load_acq"
14518 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14520 (cortex_a53_store1): Likewise.
14522 2013-08-01 Jan Hubicka <jh@suse.cz>
14524 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14525 partitions are not needed.
14527 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14529 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14530 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14531 MAYBE_NON_Q_CLASS_P where appropriate.
14533 2013-08-01 Jan Hubicka <jh@suse.cz>
14535 * cgraph.h (release_function_body): Declare.
14536 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14537 of unused delcarations.
14539 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14541 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14542 RTL form when subtracting a constant.
14544 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14546 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14547 Generate canonical plus rtx with negated immediate instead of minus
14549 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14551 2013-08-01 Jan Hubicka <jh@suse.cz>
14553 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14554 (cgraph_release_function_body): Likewise.
14555 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
14556 * cgraph.h (cgrpah_node): Rename abstract_and_needed
14557 to used_as_abstract_origin.
14558 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
14559 symbols used as abstract origins.
14560 * cgraphunit.c (analyze_functions): Update.
14561 * ipa.c (symtab_remove_unreachable_nodes): Recompute
14562 used_as_abstract_origin.
14563 * tree-inline.c (tree_function_versioning): Update
14564 used_as_abstract_origin; be ready for DECL_RESULT and
14565 DECL_ARGUMENTS to be NULL.
14567 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
14568 for abstract functions.
14569 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
14572 2013-08-01 Jan Hubicka <jh@suse.cz>
14574 * profile.c (compute_value_histograms): Fix thinko.
14576 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14578 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
14579 aarch-common-protos.h to extra_headers.
14580 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14581 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
14582 * config/aarch64/t-aarch64 (aarch-common.o): Define.
14584 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14586 * config/aarch64/aarch64.md (define_attr "type"): Delete.
14587 Include "../arm/types.md". Define "type" attribute for all patterns.
14588 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
14591 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
14593 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
14594 to support power8 load fusion.
14595 (fusion_gpr_mem_load): Likewise.
14597 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
14599 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
14600 declarations for power8 load fusion.
14601 (emit_fusion_gpr_load): Likewise.
14603 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14604 tuning for power8, turn on fusion mode by default. Turn on sign
14605 extending fusion mode if normal fusion mode is on, and we are at
14607 (fusion_gpr_load_p): New function, return true if we can fuse an
14608 addis instruction with a dependent load to a GPR.
14609 (emit_fusion_gpr_load): Emit the instructions for power8 load
14612 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14613 (VSX load fusion peepholes): New peepholes to fuse together an
14614 addi instruction with a VSX load instruction.
14616 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14617 peepholes to fuse an addis instruction with a load to a GPR base
14618 register. If we are supporting sign extending fusions, convert
14619 sign extending loads to zero extending loads and add an explicit
14622 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14624 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14625 aarch-common-protos.h to extra_headers.
14626 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14627 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14628 (arm_early_store_addr_dep): Likewise.
14629 (arm_no_early_alu_shift_dep): Likewise.
14630 (arm_no_early_alu_shift_value_dep): Likewise.
14631 (arm_no_early_mul_dep): Likewise.
14632 (arm_no_early_store_addr_dep): Likewise.
14633 (arm_mac_accumulator_is_mul_result): Likewise.
14634 (arm_mac_accumulator_is_result): Likewise.
14635 * config/arm/aarch-common.c: ... here. New file.
14636 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14638 (arm_early_store_addr_dep): Likewise.
14639 (arm_no_early_alu_shift_dep): Likewise.
14640 (arm_no_early_alu_shift_value_dep): Likewise.
14641 (arm_no_early_mul_dep): Likewise.
14642 (arm_no_early_store_addr_dep): Likewise.
14643 (arm_mac_accumulator_is_mul_result): Likewise.
14644 (arm_mac_accumulator_is_result): Likewise.
14645 * config/arm/aarch-common-protos.h: ... here. New file.
14646 * config/arm/t-arm (aarch-common.o): Define.
14648 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14650 * config/arm/arm.md: Include new file "types.md".
14651 (define_attr "type"): Move from here to ...
14652 (define_attr "mul32"): Likewise.
14653 (define_attr "mul64"): Likewise.
14654 * config/arm/types.md: ... here. New file.
14656 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14658 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14659 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14661 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14663 * gen-pass-instances.awk: Fix offset of substr().
14665 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14667 * Makefile.in (pass-instances.def): New.
14668 (passes.o): Replace dependency on passes.def with one on
14671 * gen-pass-instances.awk: New.
14673 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14674 rather than passes.def, updating local definition of NEXT_PASS
14675 macro to add an extra NUM parameter (currently unused).
14677 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14679 * Makefile.in (PASS_MANAGER_H): New.
14680 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14681 (passes.o): Likewise.
14682 (statistics.o): Likewise.
14683 (cgraphunit.o): Likewise.
14684 (context.o): Depend on PASS_MANAGER_H.
14686 * pass_manager.h: New.
14688 * cgraphunit.c (cgraph_add_new_function): Update for moves
14689 of globals to fields of pass_manager.
14690 (analyze_function): Likewise.
14691 (expand_function): Likewise.
14692 (ipa_passes): Likewise.
14693 (compile): Likewise.
14695 * context.c (context::context): New.
14696 * context.h (context::context): New.
14697 (context::get_passes): New.
14698 (context::passes_): New.
14700 * lto-cgraph.c (input_node): Update for moves of globals to
14701 fields of pass_manager.
14703 * passes.c (all_passes): Remove, in favor of a field of the
14704 same name within the new class pass_manager.
14705 (all_small_ipa_passes): Likewise.
14706 (all_lowering_passes): Likewise.
14707 (all_regular_ipa_passes): Likewise.
14708 (all_late_ipa_passes): Likewise.
14709 (all_lto_gen_passes): Likewise.
14710 (passes_by_id): Likewise.
14711 (passes_by_id_size): Likewise.
14712 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14713 the new class pass_manager.
14714 (set_pass_for_id): Convert to...
14715 (pass_manager::set_pass_for_id): ...method.
14716 (get_pass_for_id): Convert to...
14717 (pass_manager::get_pass_for_id): ...method.
14718 (register_one_dump_file): Move body of implementation into...
14719 (pass_manager::register_one_dump_file): ...here.
14720 (register_dump_files_1): Convert to...
14721 (pass_manager::register_dump_files_1): ...method.
14722 (register_dump_files): Convert to...
14723 (pass_manager::register_dump_files): ...method.
14724 (create_pass_tab): Update for moves of globals to fields of
14726 (dump_passes): Move body of implementation into...
14727 (pass_manager::dump_passes): ...here.
14728 (register_pass): Move body of implementation into...
14729 (pass_manager::register_pass): ...here.
14730 (init_optimization_passes): Convert into...
14731 (pass_manager::pass_manager): ...constructor for new
14732 pass_manager class, and initialize the pass_lists array.
14733 (check_profile_consistency): Update for moves of globals to
14734 fields of pass_manager.
14735 (dump_profile_report): Move body of implementation into...
14736 (pass_manager::dump_profile_report): ...here.
14737 (ipa_write_summaries_1): Update for moves of pass lists from
14738 being globals to fields of pass_manager.
14739 (ipa_write_optimization_summaries): Likewise.
14740 (ipa_read_summaries): Likewise.
14741 (ipa_read_optimization_summaries): Likewise.
14742 (execute_all_ipa_stmt_fixups): Likewise.
14744 * statistics.c (statistics_fini): Update for moves of globals to
14745 fields of pass_manager.
14747 * toplev.c (general_init): Replace call to
14748 init_optimization_passes with construction of the pass_manager
14751 * tree-pass.h (all_passes): Remove, in favor of a field of the
14752 same name within the new class pass_manager.
14753 (all_small_ipa_passes): Likewise.
14754 (all_lowering_passes): Likewise.
14755 (all_regular_ipa_passes): Likewise.
14756 (all_lto_gen_passes): Likewise.
14757 (all_late_ipa_passes): Likewise.
14758 (passes_by_id): Likewise.
14759 (passes_by_id_size): Likewise.
14760 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14761 the new class pass_manager.
14762 (get_pass_for_id): Remove.
14764 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14766 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14769 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14771 * arm.md (mulhi3): New expand pattern.
14773 2013-07-30 Jan Hubicka <jh@suse.cz>
14774 Martin Liska <marxin.liska@gmail.com>
14776 * profile.c (compute_value_histograms): Do not ICE when
14777 there is mismatch only on some counters.
14779 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14781 PR rtl-optimization/57637
14782 * function.c (move_insn_for_shrink_wrap): Also check the
14783 GEN set of the LIVE problem for the liveness analysis
14784 if it exists, otherwise give up.
14786 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14788 PR tree-optimization/57993
14789 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14790 replaced statement in the candidate table.
14791 (phi_add_costs): Return infinite cost when the hidden basis does
14792 not dominate all phis on which the candidate is dependent.
14793 (replace_one_candidate): Record replaced statement in the
14796 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14798 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14799 (ashlv2si3): New expander.
14800 (*ashlv2si3_i): New define_insn_and_split.
14801 * predicates.md (float_operation): Allow patterns with three
14802 basic sub-patterns.
14804 PR rtl-optimization/58021
14805 * mode-switching.c (create_pre_exit): Always split off preceding
14806 insns if we are not at the basic block head.
14808 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14810 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14811 (UCLIBC_DYNAMIC_LINKER): New macro.
14812 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14814 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14815 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14817 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14818 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14819 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14820 for SF and DF modes. Use ieee_quad_format for TF mode.
14821 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14822 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14823 (mips_option_override): Handle `-mnan=legacy'.
14824 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14825 `-mabs=2008' and `-mnan=2008'.
14826 (OPTION_DEFAULT_SPECS): Add "nan" default.
14827 (ASM_SPEC): Handle `-mnan='.
14828 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14829 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14830 comment accordingly.
14831 (neg<mode>2): Likewise.
14832 * config/mips/mips.opt (mabs, mnan): New options.
14833 * doc/install.texi (Configuration): Document `--with-nan=' option.
14834 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14836 (MIPS Options): Document them.
14837 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14838 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14839 * configure: Regenerate.
14840 * config.in: Regenerate.
14842 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14844 * config/i386/i386.md (float post-reload splitters): Do not check
14845 for subregs of SSE registers.
14847 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14848 H.J. Lu <hongjiu.lu@intel.com>
14852 * config/i386/i386.md (post-reload splitter
14853 to avoid partial SSE reg dependency stalls): New pattern.
14855 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14857 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14859 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14861 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14862 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14864 2013-07-26 Cary Coutant <ccoutant@google.com>
14866 * dwarf2out.c (die_checksum_ordered): Don't include template
14867 instantiations in signature.
14868 (is_template_parameter): New function.
14869 (is_template_instantiation): New function.
14870 (generate_skeleton_bottom_up): Don't include template instantiations
14872 (generate_skeleton): Likewise.
14873 (break_out_comdat_types): Move recursive call to break out nested
14875 (prune_unused_types_mark_generic_parms_dies): Call
14876 is_template_parameter.
14878 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14880 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14881 uses vector registers.
14882 * config/aarch64/iterators.md: Add attributes rtn and vas.
14884 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14885 Richard Earnshaw <richard.earnshaw@arm.com>
14887 * combine.c (simplify_comparison): Re-canonicalize operands
14889 * config/arm/arm.md (movcond_addsi): New splitter.
14891 2013-07-25 Sterling Augustine <saugustine@google.com>
14893 * dwarf2out.c (size_of_pubnames): Move code to...
14894 (include_pubname_in_output): ...here. New.
14895 (want_pubnames): Rearrange.
14896 (output_pubnames): Call include_pubname_in_output. Move assertion.
14898 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14900 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14902 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14905 * doc/extend.texi: Remove obsolete builtins. Fix
14906 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14908 2013-07-25 Jan Hubicka <jh@suse.cz>
14910 * cgraph.c (release_function_body): Break out from ...
14911 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14912 and DECL_ARGUMENTS.
14913 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14914 old_tree in the map.
14915 (create_specialized_node): Update.
14916 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14918 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14919 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14920 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14921 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14924 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14926 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14927 addsi3_carryin_alt2_<optab>): Correct output template.
14929 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14931 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14932 Adjust for arm_restrict_it.
14933 Remove trailing whitespace.
14935 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14937 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14938 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14940 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14942 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14944 PR rtl-optimization/57960
14945 * lra-constraints.c (process_alt_operands): Use the right mode
14946 when checking strict_low.
14948 2013-07-25 Jan Hubicka <jh@suse.cz>
14950 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14951 * cgraph.c (cgraph_remove_node): Do not release function body
14952 when in cgraph streaming.
14953 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14954 in other partitions are not considered reachable; fix handling of
14957 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14959 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14961 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14966 * config/arm/arm.md ("*sibcall_insn): Replace use of
14967 Ss with US. Adjust output for v5 and v4t.
14968 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14970 * config/arm/constraints.md ("Ss"): Rename to US.
14972 2013-07-25 Terry Guo <terry.guo@arm.com>
14974 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14975 shift_add/shift_sub0/shift_sub1 RTXs.
14977 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14978 Anton Blanchard <anton@au1.ibm.com>
14980 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14981 (altivec_vpks<VI_char>ss): Likewise.
14982 (altivec_vpks<VI_char>us): Likewise.
14983 (altivec_vpku<VI_char>us): Likewise.
14984 (altivec_vpku<VI_char>um): Likewise.
14986 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14988 Introduce context class.
14990 * Makefile.in (CONTEXT_H): New.
14991 (OBJS): Add context.o.
14992 (toplev.o): Add CONTEXT_H to dependencies.
14995 * toplev.c (general_init): Create the singleton gcc::context instance.
15001 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
15003 PR rtl-optimization/57968
15004 * mode-switching.c (create_pre_exit): Allow instructions that
15005 don't set a return register to need a non-exit mode.
15007 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15008 Anton Blanchard <anton@au1.ibm.com>
15010 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
15011 operands to vperm for little endian.
15012 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
15013 of lvsl to create the control mask for a vperm for little endian.
15015 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15016 Anton Blanchard <anton@au1.ibm.com>
15018 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
15019 two operands for little-endian.
15021 2013-07-23 Steve Ellcey <sellcey@mips.com>
15023 * config/mips/mips.c (mips_case_values_threshold): New.
15024 (TARGET_CASE_VALUES_THRESHOLD): Define.
15026 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15027 Anton Blanchard <anton@au1.ibm.com>
15029 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
15030 selection of field for vector splat in little endian mode.
15032 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
15034 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
15035 expanders to rs6000.md.
15036 (ior<mode>3): Likewise.
15037 (and<mode>3): Likewise.
15038 (one_cmpl<mode>2): Likewise.
15039 (nor<mode>3): Likewise.
15040 (andc<mode>3): Likewise.
15041 (eqv<mode>3): Likewise.
15042 (nand<mode>3): Likewise.
15043 (orc<mode>3): Likewise.
15045 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
15048 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
15049 to split multi-word logical operations.
15050 (rs6000_split_logical_di): Likewise.
15051 (rs6000_split_logical): Likewise.
15053 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
15054 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
15055 and allow TImode operations in 32-bit.
15056 (vsx_and<mode>3_64bit): Likewise.
15057 (vsx_ior<mode>3_32bit): Likewise.
15058 (vsx_ior<mode>3_64bit): Likewise.
15059 (vsx_xor<mode>3_32bit): Likewise.
15060 (vsx_xor<mode>3_64bit): Likewise.
15061 (vsx_one_cmpl<mode>2_32bit): Likewise.
15062 (vsx_one_cmpl<mode>2_64bit): Likewise.
15063 (vsx_nor<mode>3_32bit): Likewise.
15064 (vsx_nor<mode>3_64bit): Likewise.
15065 (vsx_andc<mode>3_32bit): Likewise.
15066 (vsx_andc<mode>3_64bit): Likewise.
15067 (vsx_eqv<mode>3_32bit): Likewise.
15068 (vsx_eqv<mode>3_64bit): Likewise.
15069 (vsx_nand<mode>3_32bit): Likewise.
15070 (vsx_nand<mode>3_64bit): Likewise.
15071 (vsx_orc<mode>3_32bit): Likewise.
15072 (vsx_orc<mode>3_64bit): Likewise.
15074 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
15075 logical types in GPRs.
15077 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
15078 logical insns to rs6000.md, and allow TImode operations in
15080 (altivec_ior<mode>3): Likewise.
15081 (altivec_xor<mode>3): Likewise.
15082 (altivec_one_cmpl<mode>2): Likewise.
15083 (altivec_nor<mode>3): Likewise.
15084 (altivec_andc<mode>3): Likewise.
15086 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
15087 attributes for moving the 128-bit logical operations into
15089 (BOOL_REGS_OUTPUT): Likewise.
15090 (BOOL_REGS_OP1): Likewise.
15091 (BOOL_REGS_OP2): Likewise.
15092 (BOOL_REGS_UNARY): Likewise.
15093 (BOOL_REGS_AND_CR0): Likewise.
15094 (one_cmpl<mode>2): Add support for DI logical operations on
15095 32-bit, splitting the operations to 32-bit.
15096 (anddi3): Likewise.
15097 (iordi3): Likewise.
15098 (xordi3): Likewise.
15099 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
15100 changes to combine the 32/64-bit code, allow logical operations on
15101 TI mode in 32-bit, and to use similar match_operator patterns like
15102 scalar mode uses. Combine the Altivec and VSX code for logical
15103 operations, and move it here.
15104 (ior<mode>3, 128-bit types): Likewise.
15105 (xor<mode>3, 128-bit types): Likewise.
15106 (one_cmpl<mode>3, 128-bit types): Likewise.
15107 (nor<mode>3, 128-bit types): Likewise.
15108 (andc<mode>3, 128-bit types): Likewise.
15109 (eqv<mode>3, 128-bit types): Likewise.
15110 (nand<mode>3, 128-bit types): Likewise.
15111 (orc<mode>3, 128-bit types): Likewise.
15112 (and<mode>3_internal): Likewise.
15113 (bool<mode>3_internal): Likewise.
15114 (boolc<mode>3_internal1): Likewise.
15115 (boolc<mode>3_internal2): Likewise.
15116 (boolcc<mode>3_internal1): Likewise.
15117 (boolcc<mode>3_internal2): Likewise.
15118 (eqv<mode>3_internal1): Likewise.
15119 (eqv<mode>3_internal2): Likewise.
15120 (one_cmpl1<mode>3_internal): Likewise.
15122 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15124 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15125 Rename flag_stack_usage to flag_stack_usage_info.
15127 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15129 * config/microblaze/sync.md: New file.
15130 * config/microblaze/microblaze.md: Include sync.md
15131 * config/microblaze/microblaze.c: Add print_operand 'y'.
15132 * config/microblaze/constraints.md: Add memory_contraint
15133 'Q' which is a single register.
15135 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
15137 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
15139 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15141 * reload.c (find_reloads): Exit loop once we find this operand
15142 cannot be reloaded somehow for this alternative.
15144 * reload.c (find_reloads): Exit loop once we find a hard register.
15146 * rtlanal.c (computed_jump_p): Exit loop once we find label
15149 * i386.c (ix86_pad_returns): Exit loop after setting replace.
15151 * cfgloopmanip.c (remove_path): Exit loop after setting
15154 * gensupport.c (subst_dup): Avoid loop if code is not
15155 MATCH_DUP nor MATCH_OP_DUP.
15157 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
15159 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
15161 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15163 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
15164 true for SP_REGNUM if mode == ptr_mode.
15165 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
15166 with value R0_REGNUM + 31.
15168 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15170 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
15171 pad pointer-typed argument downward.
15173 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15175 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
15176 and __ILP32__ when the ILP32 model is in use.
15178 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15180 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
15181 (aarch64_load_symref_appropriately): In the case of
15182 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
15183 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
15184 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
15185 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
15186 if 'mode' doesn't equal to 'ptr_mode'.
15187 (aarch64_output_mi_thunk): Add an assertion on the alignment of
15188 'vcall_offset'; change to call aarch64_emit_move differently depending
15189 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
15190 to calculate the upper bound of 'vcall_offset'.
15191 (aarch64_cannot_force_const_mem): Change to also return true if
15193 (aarch64_legitimize_reload_address): In the case of large
15194 displacements, add new local variable 'xmode' and an assertion
15195 based on it; change to use 'xmode' to generate the new rtx and
15197 (aarch64_asm_trampoline_template): Change to generate the template
15198 differently depending on TARGET_ILP32 or not; change to use
15199 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
15200 (aarch64_trampoline_size): Removed.
15201 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
15202 and replace immediate literals with it. Change to use 'ptr_mode'
15203 instead of 'DImode' and call convert_memory_address if the mode
15204 of 'fnaddr' doesn't equal to 'ptr_mode'.
15205 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
15207 (aarch64_elf_asm_destructor): Likewise.
15208 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
15209 on TARGET_ILP32 instead of aarch64_trampoline_size.
15210 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
15211 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
15212 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
15213 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
15214 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
15215 (storewb_pair<GPI:mode>_<P:mode>): ... this.
15216 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
15217 depending on the value of 'mode'.
15218 (add_losym_<mode>): New.
15219 (ldr_got_small_<mode>): New, based on ldr_got_small.
15220 (ldr_got_small): Remove.
15221 (ldr_got_small_sidi): New.
15222 * config/aarch64/iterators.md (P): New.
15223 (PTR): Change to 'ptr_mode' in the condition.
15225 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15227 * config.gcc (aarch64*-*-*): Support --with-abi.
15228 (aarch64*-*-elf): Support --with-multilib-list.
15229 (aarch64*-*-linux*): Likewise.
15230 (supported_defaults): Add abi to aarch64*-*-*.
15231 * configure.ac: Mention AArch64 for --with-multilib-list.
15232 * configure: Re-generated.
15233 * config/aarch64/biarchilp32.h: New file.
15234 * config/aarch64/biarchlp64.h: New file.
15235 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
15237 (MULTILIB_DEFAULTS): Ditto.
15238 (DRIVER_SELF_SPECS): Ditto.
15239 (ASM_SPEC): Update to also substitute -mabi.
15240 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
15241 file whose name depends on -mabi= and -mbig-endian.
15242 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
15244 (POINTER_SIZE): New define.
15245 (POINTERS_EXTEND_UNSIGNED): Ditto.
15246 (enum aarch64_abi_type): New enumeration tag.
15247 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
15248 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
15249 (TARGET_ILP32): New define.
15250 * config/aarch64/aarch64.opt (mabi): New.
15251 (aarch64_abi): New.
15252 (ilp32, lp64): New values for -mabi.
15253 * config/aarch64/t-aarch64 (comma): New define.
15254 (MULTILIB_OPTIONS): Ditto.
15255 (MULTILIB_DIRNAMES): Ditto.
15256 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
15257 * doc/invoke.texi: Document -mabi for AArch64.
15259 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
15261 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
15263 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15264 Anton Blanchard <anton@au1.ibm.com>
15266 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
15267 endianness when selecting field to splat.
15269 2013-07-22 Eric Christopher <echristo@gmail.com>
15271 * dwarf2out.c (die_odr_checksum): New function to use
15272 CHECKSUM_ macros and ULEB128 for DIE tag.
15273 (generate_type_signature): Use.
15275 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
15277 * config.gcc (sparc*-*-*): Accept leon3 processor.
15278 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
15279 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
15280 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
15281 * config/sparc/sparc.opt (enum processor_type): Add leon3.
15282 (mfix-ut699): Adjust comment.
15283 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
15284 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
15285 (CPP_CPU_SPEC): Likewise.
15286 (ASM_CPU_SPEC): Likewise.
15287 * config/sparc/sparc.c (leon3_cost): New constant.
15288 (sparc_option_override): Add leon3 support.
15289 (mem_ref): New function.
15290 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
15291 (sparc_do_work_around_errata): Look into the instruction in the delay
15292 slot and adjust accordingly. Add fix for the data cache nullify issues
15293 of the UT699. Change insertion position for the NOP.
15294 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
15295 (leon3_load): New reservation.
15296 (leon_store): Bump latency to 2.
15297 (grfpu): New automaton.
15298 (grfpu_alu): New unit.
15299 (grfpu_ds): Likewise.
15300 (leon_fp_alu): Adjust.
15301 (leon_fp_mult): Delete.
15302 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
15303 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
15304 * config/sparc/sparc.md (cpu): Add leon3.
15305 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
15306 (swapsi): Likewise.
15307 (atomic_test_and_set): Likewise.
15308 (ldstub): Likewise.
15310 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
15312 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
15313 default for R5900 targets.
15314 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
15315 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
15316 * config/mips/mips.c (mips_option_override): Report an error for
15317 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
15318 for -march=r5900 -mhard-float.
15320 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15322 * df-problems.c (can_move_insns_across): Exit loop once we
15323 find a non-fixed, non-global register.
15325 * ipa-pure-const.c (propagate_nothrow): Exit loop after
15328 * omega.c (omega_eliminate_red): Break after setting red_found.
15329 (omega_problem_has_red_equations): Similarly after setting found.
15330 (omega_query_variable): Similarly after setting coupled.
15332 2013-07-22 Marek Polacek <polacek@redhat.com>
15334 * gimplify.c: Don't include gimple.h twice.
15336 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15338 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
15339 instead of TARGET_THUMB1.
15340 (Pz): New constraint.
15341 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
15343 (compare_negsi_si): Likewise.
15344 (compare_addsi2_op0): Likewise.
15345 (compare_addsi2_op1): Likewise.
15346 (addsi3_carryin_<optab>): Likewise.
15347 (addsi3_carryin_alt2_<optab>): Likewise.
15348 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
15349 for arm_restrict_it.
15350 (subsi3_carryin): Likewise.
15351 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
15352 (minmax_arithsi): Disable for arm_restrict_it.
15353 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
15354 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
15355 (satsi_<SAT:code>_shift): Likewise.
15356 (arm_shiftsi3): Add alternative for 16-bit encoding.
15357 (arm32_movhf): Disable for arm_restrict_it.
15358 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
15359 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
15361 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
15363 * config/arm/arm.md (attribute "insn"): Delete.
15364 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
15365 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
15366 (not_shiftsi): Update for attribute change.
15367 (not_shiftsi_compare0): Likewise.
15368 (not_shiftsi_compare0_scratch): Likewise.
15369 (arm_one_cmplsi2): Likewise.
15370 (thumb1_one_cmplsi2): Likewise.
15371 (notsi_compare0): Likewise.
15372 (notsi_compare0_scratch): Likewise.
15373 (thumb1_movdi_insn): Likewise.
15374 (arm_movsi_insn): Likewise.
15375 (movhi_insn_arch4): Likewise.
15376 (movhi_bytes): Likewise.
15377 (arm_movqi_insn): Likewise.
15378 (thumb1_movqi_insn): Likewise.
15379 (arm32_movhf): Likewise.
15380 (thumb1_movhf): Likewise.
15381 (arm_movsf_soft_insn): Likewise.
15382 (thumb1_movsf_insn): Likewise.
15383 (thumb_movdf_insn): Likewise.
15384 (movsicc_insn): Likewise.
15385 (movsfcc_soft_insn): Likewise.
15386 (and_scc): Likewise.
15387 (cond_move): Likewise.
15388 (if_move_not): Likewise.
15389 (if_not_move): Likewise.
15390 (if_shift_move): Likewise.
15391 (if_move_shift): Likewise.
15392 (if_shift_shift): Likewise.
15393 (if_not_arith): Likewise.
15394 (if_arith_not): Likewise.
15395 (cond_move_not): Likewise.
15396 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
15397 (neon_mov<mode>): Likewise.
15398 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
15399 (thumb2_movsi_vfp): Likewise.
15400 (movsf_vfp): Likewise.
15401 (thumb2_movsf_vfp): Likewise.
15402 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15404 (cortexa7_older_only): Likewise.
15405 (cortexa7_younger): Likewise.
15406 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
15407 (1020alu_shift_op): Likewise.
15408 (1020alu_shift_reg_op): Likewise.
15409 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15410 (alu_shift_op): Likewise.
15411 (alu_shift_reg_op): Likewise.
15412 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15413 (11_alu_shift_op): Likewise.
15414 (11_alu_shift_reg_op): Likewise.
15415 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15416 (9_alu_shift_reg_op): Likewise.
15417 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15419 (cortex_a15_alu_shift): Likewise.
15420 (cortex_a15_alu_shift_reg): Likewise.
15421 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15422 (cortex_a5_alu_shift): Likewise.
15423 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15425 (cortex_a53_alu_shift): Likewise.
15426 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15428 (cortex_a7_alu_reg): Likewise.
15429 (cortex_a7_alu_shift): Likewise.
15430 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15431 (cortex_a8_alu_shift): Likewise.
15432 (cortex_a8_alu_shift_reg): Likewise.
15433 (cortex_a8_mov): Likewise.
15434 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15435 (cortex_a9_dp_shift): Likewise.
15436 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15437 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15438 (cortex_r4_mov): Likewise.
15439 (cortex_r4_alu_shift): Likewise.
15440 (cortex_r4_alu_shift_reg): Likewise.
15441 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15442 (526_alu_shift_op): Likewise.
15443 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15444 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15445 (626te_alu_shift_op): Likewise.
15446 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15447 (726te_alu_op): Likewise.
15448 (726te_alu_shift_op): Likewise.
15449 (726te_alu_shift_reg_op): Likewise.
15450 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15451 (mp626_alu_shift_op): Likewise.
15452 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15453 (pj4_alu_e1_conds): Likewise.
15454 (pj4_alu): Likewise.
15455 (pj4_alu_conds): Likewise.
15456 (pj4_shift): Likewise.
15457 (pj4_shift_conds): Likewise.
15458 (pj4_alu_shift): Likewise.
15459 (pj4_alu_shift_conds): Likewise.
15461 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15463 * config/arm/predicates.md (shiftable_operator_strict_it):
15465 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15466 Disable cond_exec version for arm_restrict_it.
15467 (thumb2_smaxsi3): Convert to generate cond_exec.
15468 (thumb2_sminsi3): Likewise.
15469 (thumb32_umaxsi3): Likewise.
15470 (thumb2_uminsi3): Likewise.
15471 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15472 (thumb2_neg_abssi2): Likewise.
15473 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15474 (thumb2_movsicc_insn): Adjust alternatives.
15475 (thumb2_mov_negscc): Disable for arm_restrict_it.
15476 (thumb2_mov_negscc_strict_it): New pattern.
15477 (thumb2_mov_notscc_strict_it): New pattern.
15478 (thumb2_mov_notscc): Disable for arm_restrict_it.
15479 (thumb2_ior_scc): Likewise.
15480 (thumb2_ior_scc_strict_it): New pattern.
15481 (thumb2_cond_move): Adjust for arm_restrict_it.
15482 (thumb2_cond_arith): Disable for arm_restrict_it.
15483 (thumb2_cond_arith_strict_it): New pattern.
15484 (thumb2_cond_sub): Adjust for arm_restrict_it.
15485 (thumb2_movcond): Likewise.
15486 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15487 (thumb2_zero_extendhisi2_v6): Likewise.
15488 (thumb2_zero_extendqisi2_v6): Likewise.
15489 (orsi_notsi_si): Likewise.
15490 (orsi_not_shiftsi_si): Likewise.
15492 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15494 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15495 instruction sequence is 1 byte shorter.
15497 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15499 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15500 it is not needed after split.
15502 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15505 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15506 second label for nonlocal goto receivers. Don't output pic base labels
15507 unless we're producing PIC; mark that action unreachable().
15508 (ix86_save_reg): If the function contains a nonlocal label, save the
15510 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15511 * config/darwin.c (emitted_pic_label_num): New GTY.
15512 (update_pic_label_number_if_needed): New.
15513 (machopic_output_function_base_name): Adjust for nonlocal receiver
15515 (machopic_should_output_picbase_label): New.
15516 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15517 (nonlocal_goto_receiver): New insn and split.
15519 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15521 * config/aarch64/aarch64-builtins.c
15522 (aarch64_fold_builtin): Fold abs in all modes.
15523 * config/aarch64/aarch64-simd-builtins.def
15524 (abs): Enable for all modes.
15525 * config/aarch64/arm_neon.h
15526 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15527 (vabs_f64): Add missing intrinsic.
15529 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15531 * config/aarch64/arm_neon.h (vabs_s64): New function
15533 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15536 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15537 * config/avr/avr.md (adjust_len): Add `round'.
15538 * config/avr/avr-protos.h (avr_out_round): New prototype.
15539 (avr_out_plus): Add `out_label' argument.
15540 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15541 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15542 Handle the case where `insn' is just a pattern.
15543 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15544 (avr_out_round): New function.
15545 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15547 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15549 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15550 Add check for flag_stack_usage to handle -fstack-usage support
15552 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15554 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
15555 interaction for new Power8 flags and VSX.
15557 2013-07-18 Sriraman Tallam <tmsriram@google.com>
15559 PR middle-end/57698
15560 * tree-inline.c (expand_call_inline): Emit errors during
15561 early_inlining only if optimization is not turned on.
15563 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15567 * passes.c (init_optimization_passes): Move the construction of
15568 the pass hierarchy into a new passes.def file.
15570 * Makefile.in (passes.o): Add dependency on passes.def.
15572 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15574 * passes.c (init_optimization_passes): Introduce macros for
15575 constructing the tree of passes (INSERT_PASSES_AFTER,
15576 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
15577 TERMINATE_PASS_LIST).
15579 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
15580 Wei Mi <wmi@google.com>
15582 PR rtl-optimization/57878
15583 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
15585 (reload_pseudo_compare_func): Check nregs first for reload
15588 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15590 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
15592 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
15594 * read-rtl.c (validate_const_int): Once an invalid character is
15595 seen, quit the loop.
15597 * gengtype.c (write_roots): Similarly once we find the "deletable"
15598 or "if_marked" option.
15600 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15602 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15603 "xtab" and "sat". Move value "clz" from here to ...
15604 (attriubte "type"): ... here.
15605 (satsi_<SAT:code>): Delete "insn" attribute.
15606 (satsi_<SAT:code>_shift): Likewise.
15607 (arm_zero_extendqisi2addsi): Likewise.
15608 (arm_extendqisi2addsi): Likewise.
15609 (clzsi2): Update for attribute changes.
15610 (rbitsi2): Likewise.
15611 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15613 (arm_usatsihi): Likewise.
15614 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15616 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15618 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15619 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15620 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15621 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15622 in alphabetical order.
15623 (attribute "core_cycles"): Update for attribute changes.
15624 (arm_addsi3): Likewise.
15625 (addsi3_compare0): Likewise.
15626 (addsi3_compare0_scratch): Likewise.
15627 (addsi3_compare_op1): Likewise.
15628 (addsi3_compare_op2): Likewise.
15629 (compare_addsi2_op0): Likewise.
15630 (compare_addsi2_op1): Likewise.
15631 (addsi3_carryin_shift_<optab>): Likewise.
15632 (subsi3_carryin_shift): Likewise.
15633 (rsbsi3_carryin_shift): Likewise.
15634 (arm_subsi3_insn): Likewise.
15635 (subsi3_compare0): Likewise.
15636 (subsi3_compare): Likewise.
15637 (arm_andsi3_insn): Likewise.
15638 (thumb1_andsi3_insn): Likewise.
15639 (andsi3_compare0): Likewise.
15640 (andsi3_compare0_scratch): Likewise.
15641 (zeroextractsi_compare0_scratch
15642 (andsi_not_shiftsi_si): Likewise.
15643 (iorsi3_insn): Likewise.
15644 (iorsi3_compare0): Likewise.
15645 (iorsi3_compare0_scratch): Likewise.
15646 (arm_xorsi3): Likewise.
15647 (thumb1_xorsi3_insn): Likewise.
15648 (xorsi3_compare0): Likewise.
15649 (xorsi3_compare0_scratch): Likewise.
15650 (satsi_<SAT:code>_shift): Likewise.
15652 (arm_shiftsi3): Likewise.
15653 (shiftsi3_compare0): Likewise.
15654 (not_shiftsi): Likewise.
15655 (not_shiftsi_compare0): Likewise.
15656 (not_shiftsi_compare0_scratch): Likewise.
15657 (arm_one_cmplsi2): Likewise.
15658 (thumb_one_complsi2): Likewise.
15659 (notsi_compare0): Likewise.
15660 (notsi_compare0_scratch): Likewise.
15661 (thumb1_zero_extendhisi2): Likewise.
15662 (arm_zero_extendhisi2): Likewise.
15663 (arm_zero_extendhisi2_v6): Likewise.
15664 (arm_zero_extendhisi2addsi): Likewise.
15665 (thumb1_zero_extendqisi2): Likewise.
15666 (thumb1_zero_extendqisi2_v6): Likewise.
15667 (arm_zero_extendqisi2): Likewise.
15668 (arm_zero_extendqisi2_v6): Likewise.
15669 (arm_zero_extendqisi2addsi): Likewise.
15670 (thumb1_extendhisi2): Likewise.
15671 (arm_extendhisi2): Likewise.
15672 (arm_extendhisi2_v6): Likewise.
15673 (arm_extendqisi): Likewise.
15674 (arm_extendqisi_v6): Likewise.
15675 (arm_extendqisi2addsi): Likewise.
15676 (thumb1_extendqisi2): Likewise.
15677 (thumb1_movdi_insn): Likewise.
15678 (arm_movsi_insn): Likewise.
15679 (movsi_compare0): Likewise.
15680 (movhi_insn_arch4): Likewise.
15681 (movhi_bytes): Likewise.
15682 (arm_movqi_insn): Likewise.
15683 (thumb1_movqi_insn): Likewise.
15684 (arm32_movhf): Likewise.
15685 (thumb1_movhf): Likewise.
15686 (arm_movsf_soft_insn): Likewise.
15687 (thumb1_movsf_insn): Likewise.
15688 (movdf_soft_insn): Likewise.
15689 (thumb_movdf_insn): Likewise.
15690 (arm_cmpsi_insn): Likewise.
15691 (cmpsi_shiftsi): Likewise.
15692 (cmpsi_shiftsi_swp): Likewise.
15693 (arm_cmpsi_negshiftsi_si): Likewise.
15694 (movsicc_insn): Likewise.
15695 (movsfcc_soft_insn): Likewise.
15696 (arith_shiftsi): Likewise.
15697 (arith_shiftsi_compare0
15698 (arith_shiftsi_compare0_scratch
15699 (sub_shiftsi): Likewise.
15700 (sub_shiftsi_compare0
15701 (sub_shiftsi_compare0_scratch
15702 (and_scc): Likewise.
15703 (cond_move): Likewise.
15704 (if_plus_move): Likewise.
15705 (if_move_plus): Likewise.
15706 (if_move_not): Likewise.
15707 (if_not_move): Likewise.
15708 (if_shift_move): Likewise.
15709 (if_move_shift): Likewise.
15710 (if_shift_shift): Likewise.
15711 (if_not_arith): Likewise.
15712 (if_arith_not): Likewise.
15713 (cond_move_not): Likewise.
15714 (thumb1_ashlsi3): Set type attribute.
15715 (thumb1_ashrsi3): Likewise.
15716 (thumb1_lshrsi3): Likewise.
15717 (thumb1_rotrsi3): Likewise.
15718 (shiftsi3_compare0_scratch): Likewise.
15719 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15720 (neon_mov<mode>): Likewise.
15721 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15723 (thumb2_movsi_insn): Likewise.
15724 (thumb2_cmpsi_neg_shiftsi): Likewise.
15725 (thumb2_extendqisi_v6): Likewise.
15726 (thumb2_zero_extendhisi2_v6): Likewise.
15727 (thumb2_zero_extendqisi2_v6): Likewise.
15728 (thumb2_shiftsi3_short): Likewise.
15729 (thumb2_addsi3_compare0_scratch): Likewise.
15730 (orsi_not_shiftsi_si): Likewise.
15731 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15732 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15734 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15735 (1020alu_shift_op): Likewise.
15736 (1020alu_shift_reg_op): Likewise.
15737 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15738 (alu_shift_op): Likewise.
15739 (alu_shift_reg_op): Likewise.
15740 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15741 (11_alu_shift_op): Likewise.
15742 (11_alu_shift_reg_op): Likewise.
15743 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15744 (9_alu_shift_reg_op): Likewise.
15745 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15747 (cortex_a15_alu_shift): Likewise.
15748 (cortex_a15_alu_shift_reg): Likewise.
15749 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15751 (cortex_a5_alu_shift): Likewise.
15752 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15754 (cortex_a53_alu_shift): Likewise.
15755 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15757 (cortex_a7_alu_reg): Likewise.
15758 (cortex_a7_alu_shift): Likewise.
15759 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15761 (cortex_a8_alu_shift): Likewise.
15762 (cortex_a8_alu_shift_reg): Likewise.
15763 (cortex_a8_mov): Likewise.
15764 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15765 (cortex_a9_dp_shift): Likewise.
15766 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15768 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15770 (cortex_r4_mov): Likewise.
15771 (cortex_r4_alu_shift): Likewise.
15772 (cortex_r4_alu_shift_reg): Likewise.
15773 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15774 (526_alu_shift_op): Likewise.
15775 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15776 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15777 (626te_alu_shift_op): Likewise.
15778 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15779 (726te_alu_op): Likewise.
15780 (726te_alu_shift_op): Likewise.
15781 (726te_alu_shift_reg_op): Likewise.
15782 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15783 (mp626_alu_shift_op): Likewise.
15784 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15785 (pj4_alu_e1_conds): Likewise.
15786 (pj4_alu): Likewise.
15787 (pj4_alu_conds): Likewise.
15788 (pj4_shift): Likewise.
15789 (pj4_shift_conds): Likewise.
15790 (pj4_alu_shift): Likewise.
15791 (pj4_alu_shift_conds): Likewise.
15792 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15794 (cortexa7_older_only): Likewise.
15795 (cortexa7_younger): Likewise.
15797 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15799 * ipa-pure-const.c (generate_summary): Rename to...
15800 (pure_const_generate_summary): ... this.
15802 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15804 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15806 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15809 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15812 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15814 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15815 enabled without -march=zEC12.
15816 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15819 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15821 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15822 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15823 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15824 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15825 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15826 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15827 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15829 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15830 and ISA_HAS_NMADD3_NMSUB3.
15831 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15832 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15833 (nmsub4<mode>, nmsub3<mode>): Likewise.
15834 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15836 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15838 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15839 TARGET_MIPS5400 checking.
15841 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15842 Peter Bergner <bergner@vnet.ibm.com>
15844 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15845 registers in the comment.
15846 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15847 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15848 rather than FIRST_PSEUDO_REGISTERS.
15850 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15852 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15853 enable extra ISA flags with TARGET_HTM.
15855 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15857 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15860 2013-07-15 Cong Hou <congh@google.com>
15862 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15863 in compare function for sorting.
15865 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15867 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15868 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15869 * config/rs6000/rs6000.opt: Add -mhtm option.
15870 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15871 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15872 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15873 __HTM__ if the HTM instructions are available.
15874 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15875 htm_spr_reg_operand): New define_predicates.
15876 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15877 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15879 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15880 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15881 HTM builtin functions.
15882 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15883 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15884 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15885 (rs6000_builtin_mask_calculate): Likewise.
15886 (rs6000_option_override_internal): Likewise.
15887 (bdesc_htm): Add new HTM builtin support.
15888 (htm_spr_num): New function.
15889 (htm_spr_regno): Likewise.
15890 (rs6000_htm_spr_icode): Likewise.
15891 (htm_expand_builtin): Likewise.
15892 (htm_init_builtins): Likewise.
15893 (rs6000_expand_builtin): Add support for HTM builtin functions.
15894 (rs6000_init_builtins): Likewise.
15895 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15897 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15898 (TARGET_HTM, MASK_HTM): Define macros.
15899 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15900 (FIXED_REGISTERS): Likewise.
15901 (CALL_USED_REGISTERS): Likewise.
15902 (CALL_REALLY_USED_REGISTERS): Likewise.
15903 (REG_ALLOC_ORDER): Likewise.
15904 (enum reg_class): Likewise.
15905 (REG_CLASS_NAMES): Likewise.
15906 (REG_CLASS_CONTENTS): Likewise.
15907 (REGISTER_NAMES): Likewise.
15908 (ADDITIONAL_REGISTER_NAMES): Likewise.
15909 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15910 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15911 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15912 * config/rs6000/htm.md: New file.
15913 * config/rs6000/htmintrin.h: New file.
15914 * config/rs6000/htmxlintrin.h: New file.
15916 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15918 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15919 Define SYMBOL_TINY_GOT, update comment.
15920 * config/aarch64/aarch64.c
15921 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15922 (aarch64_expand_mov_immediate): Likewise.
15923 (aarch64_print_operand): Likewise.
15924 (aarch64_classify_symbol): Likewise.
15925 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15926 (ldr_got_tiny): Define.
15928 2013-07-13 Tobias Grosser <tobias@grosser.es>
15930 PR tree-optimization/54094
15931 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15932 scheduling dimension for the parallelism check from the polyhedral
15933 information in the AST.
15934 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15937 2013-07-13 Jason Merrill <jason@redhat.com>
15939 * print-tree.c (debug_vec_tree): Use debug_raw.
15940 (debug_raw (vec<tree, va_gc> &)): New.
15941 (debug_raw (vec<tree, va_gc> *)): New.
15942 * tree.h: Declare them.
15944 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15946 * ifcvt.c (ifcvt_after_combine): New static variable.
15947 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15949 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15950 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15951 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15953 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15955 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15957 2013-07-12 Michael Matz <matz@suse.de>
15959 PR middle-end/55771
15960 * convert.c (convert_to_real): Reject non-float inner types.
15962 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15964 * config/aarch64/aarch64-protos.h
15965 (aarch64_simd_immediate_valid_for_move): Remove.
15966 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15967 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15969 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15971 2013-07-11 Steve Ellcey <sellcey@mips.com>
15973 * config/mips/mips.c (mips_conditional_register_usage): Do not
15974 use t[0-7] registers in MIPS16 mode when optimizing for size.
15976 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15978 * config/i386/i386.c (dispatch_function_versions): Fix array
15979 indexing of function_version_info to match actual_versions.
15981 2013-07-11 Teresa Johnson <tejohnson@google.com>
15983 * vec.h (struct va_gc): Move release out-of-line.
15984 (va_gc::release): Call ggc_free on released vec.
15986 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15988 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15989 Require GOT register as additional operand in UNSPEC.
15990 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15991 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15992 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15993 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15994 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15995 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15996 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15998 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
16001 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
16002 name seen by assembler/linker rather if available.
16004 2013-07-11 Andreas Schwab <schwab@suse.de>
16006 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
16008 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
16010 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
16012 2013-07-10 Joseph Myers <joseph@codesourcery.com>
16014 * doc/tm.texi.in: Move hook documentation to ....
16015 * target.def: ... here.
16017 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
16018 text on @hook line.
16019 * doc/tm.texi: Regenerate.
16021 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
16024 * doc/invoke.texi: Document -Wconditionally-supported.
16026 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16029 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
16032 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16035 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
16036 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
16037 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
16038 Remove duplicate devices.
16039 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
16040 * config/avr/t-multilib: Regenerate.
16041 * config/avr/avr-tables.opt: Regenerate.
16042 * doc/avr-mmcu.texi: Regenerate.
16044 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16047 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
16049 2013-07-10 Graham Stott <graham.stott@btinternet.com>
16051 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
16052 the cost of MULT when optimizing for size.
16054 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16056 * config/cr16/cr16-protos.h: Don't include target.h.
16058 2013-07-09 Joseph Myers <joseph@codesourcery.com>
16060 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
16061 adjust register size for TDmode and TFmode for VSX registers.
16063 2013-07-08 Kai Tietz <ktietz@redhat.com>
16066 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
16067 hook_bool_const_tree_true.
16069 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16071 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
16072 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
16073 * config/s390/s390.md: Define FPR*_REGNUM constants.
16074 Fix FPR2_REGNUM constant (18 -> 17).
16075 ("*trunc<BFP:mode><DFP_ALL:mode>2")
16076 ("*trunc<DFP_ALL:mode><BFP:mode>2")
16077 ("trunc<BFP:mode><DFP_ALL:mode>2")
16078 ("trunc<DFP_ALL:mode><BFP:mode>2")
16079 ("*extend<BFP:mode><DFP_ALL:mode>2")
16080 ("*extend<DFP_ALL:mode><BFP:mode>2")
16081 ("extend<BFP:mode><DFP_ALL:mode>2")
16082 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
16085 2013-07-08 Graham Stott <graham.stott@btinternet.com>
16087 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
16089 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16091 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
16092 and cfun_fpr_bit_p to cfun_fpr_save_p.
16093 (s390_frame_area, s390_register_info, s390_frame_info)
16094 (s390_emit_prologue, s390_emit_epilogue)
16095 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
16097 * config/s390/s390.h: Define *_REGNUM macros for floating point
16100 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
16102 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
16104 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
16106 PR rtl-optimization/57786
16107 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
16108 and break out of the loop when it is set to false.
16110 2013-07-08 Jakub Jelinek <jakub@redhat.com>
16113 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
16114 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
16115 (const_int 63)) 0)).
16116 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
16117 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
16118 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
16120 PR rtl-optimization/57829
16121 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
16122 mask bits outside of mode are just sign-extension from mode to HWI.
16124 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
16126 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
16127 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
16128 adjust_address instead of change_address to keep info about alignment.
16129 (emit_strmov): Remove.
16130 (emit_memmov): New function.
16131 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
16132 (expand_movmem_epilogue): Likewise and return updated rtx for
16134 (expand_constant_movmem_prologue): Likewise and return updated rtx for
16135 destination and source.
16136 (decide_alignment): Refactor, handle vector_loop.
16137 (ix86_expand_movmem): Likewise.
16138 (ix86_expand_setmem): Likewise.
16139 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
16141 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
16143 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
16144 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
16146 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16148 * config/i386/sse.md (sse_movlhps): Change alternative 3
16149 of operand 2 to "m".
16151 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16154 * config/i386/sse.md (iptr): New mode attribute.
16155 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
16156 (<sse>_vm<plusminus_insn><mode>3): Ditto.
16157 (<sse>_vmmul<mode>3): Ditto.
16158 (<sse>_vmdiv<mode>3): Ditto.
16159 (sse_vmrcpv4sf2): Ditto.
16160 (<sse>_vmsqrt<mode>2): Ditto.
16161 (sse_vmrsqrtv4sf2): Ditto.
16162 (<sse>_vm<code><mode>3): Ditto.
16163 (avx_vmcmp<mode>3): Ditto.
16164 (<sse>_vmmaskcmp<mode>3): Ditto.
16165 (<sse>_comi): Ditto.
16166 (<sse>_ucomi): Ditto.
16167 (*xop_vmfrcz_<mode>): Ditto.
16168 (*fmai_fmadd_<mode>): Ditto.
16169 (*fmai_fmsub_<mode>): Ditto.
16170 (*fmai_fnmadd_<mode>): Ditto.
16171 (*fmai_fnmsub_<mode>): Ditto.
16172 (*fma4i_vmfmadd_<mode>): Ditto.
16173 (*fma4i_vmfmsub_<mode>): Ditto.
16174 (*fma4i_vmfnmadd_<mode>): Ditto.
16175 (*fma4i_vmfnmsub_<mode>): Ditto.
16176 (*xop_vmfrcz_<mode>): Ditto.
16177 (sse_cvtps2pi): Ditto.
16178 (sse_cvttps2pi): Ditto.
16179 (sse_cvtss2si): Ditto.
16180 (sse_cvtss2si_2): Ditto.
16181 (sse_cvtss2siq_2): Ditto.
16182 (sse_cvttss2si): Ditto.
16183 (sse_cvttss2siq): Ditto.
16184 (sse_cvtsd2si): Ditto.
16185 (sse_cvtsd2si_2): Ditto.
16186 (sse_cvtsd2siq_2): Ditto.
16187 (sse_cvttsd2si): Ditto.
16188 (sse_cvttsd2siq): Ditto.
16189 (sse_cvtsd2ss): Ditto.
16190 (sse_cvtss2sd): Ditto.
16191 (avx2_pbroadcast<mode>): Ditto.
16192 (avx2_pbroadcast<mode>_1): Ditto.
16193 (*avx_vperm_broadcast_v4sf): Ditto.
16195 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
16196 (sse_movlhps): Ditto.
16197 (sse_storehps): Ditto.
16198 (sse_loadhps): Ditto.
16199 (sse_storelps): Ditto.
16200 (sse_loadlps): Ditto.
16201 (*vec_concatv4sf): Ditto.
16202 (*vec_interleave_highv2df): Ditto.
16203 (*vec_interleave_lowv2df): Ditto.
16204 (*vec_extractv2df_1_sse): Ditto.
16205 (*vec_extractv2df_0_sse): Ditto.
16206 (sse2_storelpd): Ditto.
16207 (sse2_loadlpd): Ditto.
16208 (sse2_movsd): Ditto.
16209 (*vec_concatv4si): Ditto.
16210 (vec_concatv2di): Ditto.
16212 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
16213 for Intel asm dialect.
16214 (mmx_punpcklwd): Ditto.
16215 (mmx_punpckldq): Ditto.
16217 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
16218 for intel assembler dialect.
16220 2013-07-06 Jakub Jelinek <jakub@redhat.com>
16223 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
16224 for BUILT_IN_C{LZ,LRSB}*.
16225 * tree.h (CASE_INT_FN): Add FN##IMAX case.
16226 * tree-vrp.c (extract_range_basic): Handle
16227 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
16228 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
16229 fall thru to code calling set_value*.
16230 * builtins.c (expand_builtin): Remove *IMAX cases.
16231 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
16232 if width is bigger than 2*HWI.
16234 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
16236 PR rtl-optimization/55342
16237 * lra-int.h (lra_subreg_reload_pseudos): New.
16238 * lra.c: Add undoing optional reloads to the block diagram.
16239 (lra_subreg_reload_pseudos): New.
16240 (lra_optional_reload_pseudos): Change comments.
16241 (lra): Init and clear lra_subreg_reload_pseudos. Clear
16242 lra_optional_reload_pseudos after undo transformations.
16243 * lra-assigns.c (pseudo_prefix_title): New.
16244 (lra_setup_reg_renumber): Use it.
16245 (spill_for): Ditto. Check subreg reload pseudos too.
16246 (assign_by_spills): Consider subreg reload pseudos too.
16247 * lra-constraints.c (simplify_operand_subreg): Use
16248 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
16249 (curr_insn_transform): Recognize and do optional reloads.
16250 (undo_optional_reloads): New.
16251 (lra_undo_inheritance): Call undo_optional_reloads.
16253 2013-07-05 Thomas Quinot <quinot@adacore.com>
16255 * tree-complex.c (expand_complex_operations_1): Fix typo.
16257 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
16259 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
16260 (tune_params): New member 'const vec_costs'.
16261 * config/aarch64/aarch64.c (generic_vector_cost): New.
16262 (generic_tunings): New member 'generic_vector_cost'.
16263 (aarch64_builtin_vectorization_cost): New.
16264 (aarch64_add_stmt_cost): New.
16265 (TARGET_VECTORIZE_ADD_STMT_COST): New.
16266 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
16268 2013-07-03 Jakub Jelinek <jakub@redhat.com>
16271 * config/i386/predicates.md (vsib_address_operand): Disallow
16272 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
16274 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
16276 PR middle-end/55030
16277 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
16278 expand_builtin_setjmp_receiver.
16279 (expand_label): Adjust, call expand_builtin_setjmp_receiver
16280 with NULL for the label parameter.
16281 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
16282 the frame-pointer. Adjust comments.
16283 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
16284 only if LABEL is non-NULL.
16286 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
16288 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
16289 (ARM_ABI_AAPCS64): Ditto.
16291 (ARM_DEFAULT_ABI): Ditto.
16293 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
16295 * config/aarch64/aarch64-builtins.c
16296 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
16297 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
16299 * config/aarch64/aarch64-simd.md
16300 (aarch64_ld1<VALL:mode>): New.
16301 (aarch64_st1<VALL:mode>): Likewise.
16302 * config/aarch64/arm_neon.h
16303 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
16305 2013-07-02 Sriraman Tallam <tmsriram@google.com>
16307 * config/i386/i386.c (gate_insert_vzeroupper): Check if
16309 (ix86_option_override_internal):Turn on all -mavx target flags by
16310 default as they are dependent on AVX anyway.
16312 2013-07-02 Cary Coutant <ccoutant@google.com>
16314 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
16315 deterministic hash.
16316 (loc_checksum_ordered): Likewise.
16317 (hash_loc_operands): Remove inline keyword.
16319 2013-07-02 Jakub Jelinek <jakub@redhat.com>
16321 PR tree-optimization/57741
16322 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
16323 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
16324 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
16325 Allow REAL_CST step_exprs if flag_associative_math.
16326 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
16328 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16330 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
16332 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16334 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
16336 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16338 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
16340 (iorsi3_insn): Likewise.
16341 (arm_xorsi3): Likewise.
16343 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
16345 * arm.md (attribute "wtype"): Delete. Move attribute values from here
16347 (attribute "type"): ... here, and prefix with "wmmx_".
16348 (attribute "core_cycles"): Update for attribute changes.
16349 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
16350 (tbcstv4hi): Likewise.
16351 (tbcstv2si): Likewise.
16352 (iwmmxt_iordi3): Likewise.
16353 (iwmmxt_xordi3): Likewise.
16354 (iwmmxt_anddi3): Likewise.
16355 (iwmmxt_nanddi3): Likewise.
16356 (iwmmxt_arm_movdi): Likewise.
16357 (iwmmxt_movsi_insn): Likewise.
16358 (mov<mode>_internal): Likewise.
16359 (and<mode>3_iwmmxt): Likewise.
16360 (ior<mode>3_iwmmxt): Likewise.
16361 (xor<mode>3_iwmmxt): Likewise.
16362 (add<mode>3_iwmmxt): Likewise.
16363 (ssaddv8qi3): Likewise.
16364 (ssaddv4hi3): Likewise.
16365 (ssaddv2si3): Likewise.
16366 (usaddv8qi3): Likewise.
16367 (usaddv4hi3): Likewise.
16368 (usaddv2si3): Likewise.
16369 (sub<mode>3_iwmmxt): Likewise.
16370 (sssubv8qi3): Likewise.
16371 (sssubv4hi3): Likewise.
16372 (sssubv2si3): Likewise.
16373 (ussubv8qi3): Likewise.
16374 (ussubv4hi3): Likewise.
16375 (ussubv2si3): Likewise.
16376 (mulv4hi3_iwmmxt): Likewise.
16377 (smulv4hi3_highpart): Likewise.
16378 (umulv4hi3_highpart): Likewise.
16379 (iwmmxt_wmacs): Likewise.
16380 (iwmmxt_wmacsz): Likewise.
16381 (iwmmxt_wmacu): Likewise.
16382 (iwmmxt_wmacuz): Likewise.
16383 (iwmmxt_clrdi): Likewise.
16384 (iwmmxt_clrv8qi): Likewise.
16385 (iwmmxt_clr4hi): Likewise.
16386 (iwmmxt_clr2si): Likewise.
16387 (iwmmxt_uavgrndv8qi3): Likewise.
16388 (iwmmxt_uavgrndv4hi3): Likewise.
16389 (iwmmxt_uavgv8qi3): Likewise.
16390 (iwmmxt_uavgv4hi3): Likewise.
16391 (iwmmxt_tinsrb): Likewise.
16392 (iwmmxt_tinsrh): Likewise.
16393 (iwmmxt_tinsrw): Likewise.
16394 (iwmmxt_textrmub): Likewise.
16395 (iwmmxt_textrmsb): Likewise.
16396 (iwmmxt_textrmuh): Likewise.
16397 (iwmmxt_textrmsh): Likewise.
16398 (iwmmxt_textrmw): Likewise.
16399 (iwmxxt_wshufh): Likewise.
16400 (eqv8qi3): Likewise.
16401 (eqv4hi3): Likewise.
16402 (eqv2si3): Likewise.
16403 (gtuv8qi3): Likewise.
16404 (gtuv4hi3): Likewise.
16405 (gtuv2si3): Likewise.
16406 (gtv8qi3): Likewise.
16407 (gtv4hi3): Likewise.
16408 (gtv2si3): Likewise.
16409 (smax<mode>3_iwmmxt): Likewise.
16410 (umax<mode>3_iwmmxt): Likewise.
16411 (smin<mode>3_iwmmxt): Likewise.
16412 (umin<mode>3_iwmmxt): Likewise.
16413 (iwmmxt_wpackhss): Likewise.
16414 (iwmmxt_wpackwss): Likewise.
16415 (iwmmxt_wpackdss): Likewise.
16416 (iwmmxt_wpackhus): Likewise.
16417 (iwmmxt_wpackwus): Likewise.
16418 (iwmmxt_wpackdus): Likewise.
16419 (iwmmxt_wunpckihb): Likewise.
16420 (iwmmxt_wunpckihh): Likewise.
16421 (iwmmxt_wunpckihw): Likewise.
16422 (iwmmxt_wunpckilb): Likewise.
16423 (iwmmxt_wunpckilh): Likewise.
16424 (iwmmxt_wunpckilw): Likewise.
16425 (iwmmxt_wunpckehub): Likewise.
16426 (iwmmxt_wunpckehuh): Likewise.
16427 (iwmmxt_wunpckehuw): Likewise.
16428 (iwmmxt_wunpckehsb): Likewise.
16429 (iwmmxt_wunpckehsh): Likewise.
16430 (iwmmxt_wunpckehsw): Likewise.
16431 (iwmmxt_wunpckelub): Likewise.
16432 (iwmmxt_wunpckeluh): Likewise.
16433 (iwmmxt_wunpckeluw): Likewise.
16434 (iwmmxt_wunpckelsb): Likewise.
16435 (iwmmxt_wunpckelsh): Likewise.
16436 (iwmmxt_wunpckelsw): Likewise.
16437 (ror<mode>3): Likewise.
16438 (ashr<mode>3_iwmmxt): Likewise.
16439 (lshr<mode>3_iwmmxt): Likewise.
16440 (ashl<mode>3_iwmmxt): Likewise.
16441 (ror<mode>3_di): Likewise.
16442 (ashr<mode>3_di): Likewise.
16443 (lshr<mode>3_di): Likewise.
16444 (ashl<mode>3_di): Likewise.
16445 (iwmmxt_wmadds): Likewise.
16446 (iwmmxt_wmaddu): Likewise.
16447 (iwmmxt_tmia): Likewise.
16448 (iwmmxt_tmiaph): Likewise.
16449 (iwmmxt_tmiabb): Likewise.
16450 (iwmmxt_tmiatb): Likewise.
16451 (iwmmxt_tmiabt): Likewise.
16452 (iwmmxt_tmiatt): Likewise.
16453 (iwmmxt_tmovmskb): Likewise.
16454 (iwmmxt_tmovmskh): Likewise.
16455 (iwmmxt_tmovmskw): Likewise.
16456 (iwmmxt_waccb): Likewise.
16457 (iwmmxt_wacch): Likewise.
16458 (iwmmxt_waccw): Likewise.
16459 (iwmmxt_waligni): Likewise.
16460 (iwmmxt_walignr): Likewise.
16461 (iwmmxt_walignr0): Likewise.
16462 (iwmmxt_walignr1): Likewise.
16463 (iwmmxt_walignr2): Likewise.
16464 (iwmmxt_walignr3): Likewise.
16465 (iwmmxt_wsadb): Likewise.
16466 (iwmmxt_wsadh): Likewise.
16467 (iwmmxt_wsadbz): Likewise.
16468 (iwmmxt_wsadhz): Likewise.
16469 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16470 (iwmmxt_wabsdiffb): Likewise.
16471 (iwmmxt_wabsdiffh): Likewise.
16472 (iwmmxt_wabsdiffw): Likewise.
16473 (iwmmxt_waddsubhx): Likewise
16474 (iwmmxt_wsubaddhx): Likewise.
16475 (addc<mode>3): Likewise.
16476 (iwmmxt_avg4): Likewise.
16477 (iwmmxt_avg4r): Likewise.
16478 (iwmmxt_wmaddsx): Likewise.
16479 (iwmmxt_wmaddux): Likewise.
16480 (iwmmxt_wmaddsn): Likewise.
16481 (iwmmxt_wmaddun): Likewise.
16482 (iwmmxt_wmulwsm): Likewise.
16483 (iwmmxt_wmulwum): Likewise.
16484 (iwmmxt_wmulsmr): Likewise.
16485 (iwmmxt_wmulumr): Likewise.
16486 (iwmmxt_wmulwsmr): Likewise.
16487 (iwmmxt_wmulwumr): Likewise.
16488 (iwmmxt_wmulwl): Likewise.
16489 (iwmmxt_wqmulm): Likewise.
16490 (iwmmxt_wqmulwm): Likewise.
16491 (iwmmxt_wqmulmr): Likewise.
16492 (iwmmxt_wqmulwmr): Likewise.
16493 (iwmmxt_waddbhusm): Likewise.
16494 (iwmmxt_waddbhusl): Likewise.
16495 (iwmmxt_wqmiabb): Likewise.
16496 (iwmmxt_wqmiabt): Likewise.
16497 (iwmmxt_wqmiatb): Likewise.
16498 (iwmmxt_wqmiatt): Likewise.
16499 (iwmmxt_wqmiabbn): Likewise.
16500 (iwmmxt_wqmiabtn): Likewise.
16501 (iwmmxt_wqmiatbn): Likewise.
16502 (iwmmxt_wqmiattn): Likewise.
16503 (iwmmxt_wmiabb): Likewise.
16504 (iwmmxt_wmiabt): Likewise.
16505 (iwmmxt_wmiatb): Likewise.
16506 (iwmmxt_wmiatt): Likewise.
16507 (iwmmxt_wmiabbn): Likewise.
16508 (iwmmxt_wmiabtn): Likewise.
16509 (iwmmxt_wmiatbn): Likewise.
16510 (iwmmxt_wmiattn): Likewise.
16511 (iwmmxt_wmiawbb): Likewise.
16512 (iwmmxt_wmiawbt): Likewise.
16513 (iwmmxt_wmiawtb): Likewise.
16514 (iwmmxt_wmiawtt): Likewise.
16515 (iwmmxt_wmiawbbn): Likewise.
16516 (iwmmxt_wmiawbtn): Likewise.
16517 (iwmmxt_wmiawtbn): Likewise.
16518 (iwmmxt_wmiawttn): Likewise.
16519 (iwmmxt_wmerge): Likewise.
16520 (iwmmxt_tandc<mode>3): Likewise.
16521 (iwmmxt_torc<mode>3): Likewise.
16522 (iwmmxt_torvsc<mode>3): Likewise.
16523 (iwmmxt_textrc<mode>3): Likewise.
16524 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16525 (wmmxt_pack): Likewise.
16526 (wmmxt_mult_c1): Likewise.
16527 (wmmxt_mult_c2): Likewise.
16528 (wmmxt_alu_c1): Likewise.
16529 (wmmxt_alu_c2): Likewise.
16530 (wmmxt_alu_c3): Likewise.
16531 (wmmxt_transfer_c1): Likewise.
16532 (wmmxt_transfer_c2): Likewise.
16533 (wmmxt_transfer_c3): Likewise.
16534 (marvell_f_iwmmxt_wstr): Likewise.
16535 (marvell_f_iwmmxt_wldr): Likewise.
16537 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16539 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16541 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16544 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16545 * lra-constraints.c (need_for_split_p): Check call used hard regs
16546 living through calls.
16548 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16549 call used regs for call insn.
16551 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16554 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
16555 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
16557 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16559 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
16561 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16563 * lra-constraints.c (need_for_split_p): Check call used hard regs
16564 living through calls.
16566 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
16569 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
16570 to tie with any other modes. Eliminate Altivec vector mode tests,
16571 since these are a subset of ALTIVEC or VSX vector modes. Simplify
16572 code, to return 0 if testing MODE2 for a condition, if we've
16573 already tested MODE1 for the same condition.
16575 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16577 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
16580 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16582 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16583 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
16585 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16587 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
16589 (aarch64_symbolic_constant_p): Remove.
16590 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
16591 static. Fix line length and white space.
16592 (aarch64_symbolic_constant_p): Remove.
16593 * config/aarch64/predicates.md (aarch64_valid_symref):
16594 Use aarch64_classify_symbol_expression.
16596 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16598 * config/arm/constraints.md (Ts): New constraint.
16599 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
16601 (compare_scc): Use "Ts" constraint for operand 0.
16602 (ior_scc_scc): Likewise.
16603 (and_scc_scc): Likewise.
16604 (and_scc_scc_nodom): Likewise.
16605 (ior_scc_scc_cmp): Likewise for operand 7.
16606 (and_scc_scc_cmp): Likewise.
16607 * config/arm/thumb2.md (thumb2_movsi_insn):
16608 Add alternatives for 16-bit encodings.
16609 (thumb2_movhi_insn): Likewise.
16610 (thumb2_movsicc_insn): Likewise.
16611 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16612 (thumb2_negscc): Use "Ts" constraint.
16613 Move mvn instruction outside cond_exec block.
16614 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16615 for 16-bit encodings.
16617 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16619 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16621 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16622 (mulsi3subsi): Likewise.
16623 (mulsidi3adddi): Likewise.
16624 (mulsidi3_v6): Likewise.
16625 (umulsidi3_v6): Likewise.
16626 (umulsidi3adddi_v6): Likewise.
16627 (smulsi3_highpart_v6): Likewise.
16628 (umulsi3_highpart_v6): Likewise.
16629 (mulhisi3tb): Likewise.
16630 (mulhisi3bt): Likewise.
16631 (mulhisi3tt): Likewise.
16632 (maddhisi4): Likewise.
16633 (maddhisi4tb): Likewise.
16634 (maddhisi4tt): Likewise.
16635 (maddhidi4): Likewise.
16636 (maddhidi4tb): Likewise.
16637 (maddhidi4tt): Likewise.
16638 (zeroextractsi_compare0_scratch): Likewise.
16639 (insv_zero): Likewise.
16640 (insv_t2): Likewise.
16641 (anddi_notzesidi_di): Likewise.
16642 (anddi_notsesidi_di): Likewise.
16643 (andsi_notsi_si): Likewise.
16644 (iordi_zesidi_di): Likewise.
16645 (xordi_zesidi_di): Likewise.
16646 (andsi_iorsi3_notsi): Likewise.
16647 (smax_0): Likewise.
16648 (smax_m1): Likewise.
16649 (smin_0): Likewise.
16650 (not_shiftsi): Likewise.
16651 (unaligned_loadsi): Likewise.
16652 (unaligned_loadhis): Likewise.
16653 (unaligned_loadhiu): Likewise.
16654 (unaligned_storesi): Likewise.
16655 (unaligned_storehi): Likewise.
16656 (extv_reg): Likewise.
16657 (extzv_t2): Likewise.
16658 (divsi3): Likewise.
16659 (udivsi3): Likewise.
16660 (arm_zero_extendhisi2addsi): Likewise.
16661 (arm_zero_extendqisi2addsi): Likewise.
16662 (compareqi_eq0): Likewise.
16663 (arm_extendhisi2_v6): Likewise.
16664 (arm_extendqisi2addsi): Likewise.
16665 (arm_movt): Likewise.
16666 (thumb2_ldrd): Likewise.
16667 (thumb2_ldrd_base): Likewise.
16668 (thumb2_ldrd_base_neg): Likewise.
16669 (thumb2_strd): Likewise.
16670 (thumb2_strd_base): Likewise.
16671 (thumb2_strd_base_neg): Likewise.
16672 (arm_negsi2): Add alternative for 16-bit encoding.
16673 (arm_one_cmplsi2): Likewise.
16675 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16677 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16678 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16679 (movdfcc): Likewise.
16680 * config/arm/vfp.md (*thumb2_movsf_vfp):
16681 Disable predication for arm_restrict_it.
16682 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16683 (*thumb2_movdfcc_vfp): Likewise.
16684 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16685 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16686 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16687 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16688 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16689 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16690 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16691 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16692 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16693 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16694 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16695 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16696 Disable predication for arm_restrict_it.
16698 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16700 * config/i386/bmiintrin.h (_bextr_u32): New.
16701 (_bextr_u64): Ditto.
16703 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16705 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16706 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16707 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16708 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16709 * config/mips/n32-elf.h: ...this new file.
16711 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16714 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16715 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16717 2013-06-27 Catherine Moore <clm@codesourcery.com>
16719 * config/mips/mips-tables.opt: Regenerate.
16720 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16721 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16722 * doc/invoke.texi: Add -m14kc.
16724 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16727 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16728 constraints of operand 1 and 2.
16731 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16732 to match RTL canonicalization. Swap predicates and
16733 constraints of operand 1 and 2.
16735 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16737 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16738 Process OP_INOUT regs for splitting too.
16740 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16742 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16743 decl before the loop, initialize to NULL.
16744 (vectorizable_load): Initialize ptr_incr to NULL.
16746 2013-06-27 Martin Jambor <mjambor@suse.cz>
16749 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16750 * ipa-ref.c (ipa_maybe_record_reference): New function.
16751 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16752 * ipa-cp.c (create_specialized_node): Record potential references from
16754 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16756 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16758 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16759 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16760 'mode' to force_reg.
16761 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16762 (aarch64_expand_mov_immediate): Likewise.
16764 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16766 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16767 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16769 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16771 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16772 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16773 (struct machine_function): Add tbegin_p.
16774 (s390_canonicalize_comparison): Fold CC mode compares to
16775 conditional jump if possible.
16776 (s390_emit_jump): Return the emitted jump.
16777 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16778 Handle CCRAWmode compares.
16779 (s390_option_override): Default to -mhtm if available.
16780 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16781 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16782 FPRs instead of df_regs_ever_live_p.
16783 (s390_optimize_nonescaping_tx): New function.
16784 (s390_init_frame_layout): Extend clobbered_regs array to cover
16786 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16787 (s390_expand_tbegin): New function.
16788 (enum s390_builtin): New enum definition.
16789 (code_for_builtin): New array definition.
16790 (s390_init_builtins): New function.
16791 (s390_expand_builtin): New function.
16792 (TARGET_INIT_BUILTINS): Define.
16793 (TARGET_EXPAND_BUILTIN): Define.
16794 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16795 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16796 (s390_alc_comparison): Likewise.
16797 * config/s390/s390-modes.def: Add CCRAWmode.
16798 * config/s390/s390.h (processor_flags): Add PF_TX.
16799 (TARGET_CPU_HTM): Define macro.
16800 (TARGET_HTM): Define macro.
16801 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16802 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16803 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16804 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16805 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16806 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16807 ("*cc_to_int"): Move up.
16808 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16809 constants other than 0.
16810 ("*ccraw_to_int"): New insn and splitter definition.
16811 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16812 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16813 ("tx_assist"): New expander.
16814 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16815 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16816 * config/s390/s390.opt: Add -mhtm option.
16817 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16818 * config/s390/htmxlintrin.h: New file.
16819 * config/s390/htmintrin.h: New file.
16820 * config/s390/s390intrin.h: New file.
16821 * doc/extend.texi: Document htm builtins.
16822 * config.gcc: Add the new header files to extra_headers.
16824 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16826 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16827 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16829 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16830 Pat Haugen <pthaugen@us.ibm.com>
16831 Peter Bergner <bergner@vnet.ibm.com>
16833 * config/rs6000/power8.md: New.
16834 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16835 setting for power8 entry.
16836 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16837 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16838 test for Power4/Power5 only.
16839 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16841 (force_new_group): Adjust comment.
16842 * config/rs6000/rs6000.md: Include power8.md.
16844 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16846 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16847 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16849 (tune_params): Update comment.
16850 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16851 (arm_max_conditional_execute): New function.
16852 (thumb2_final_prescan_insn): Use max_insn_skipped and
16853 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16855 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16857 PR tree-optimization/57705
16858 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16859 SSA_NAME step, provided that it is not defined inside the loop.
16860 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16861 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16863 2013-06-25 Martin Jambor <mjambor@suse.cz>
16865 PR middle-end/57670
16866 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16867 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16869 (ipa_note_param_call): Initialize member_ptr flag.
16870 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16871 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16872 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16873 (ipa_read_indirect_edge_info): Likewise.
16875 2013-06-25 Richard Biener <rguenther@suse.de>
16877 PR middle-end/56977
16878 * passes.c (init_optimization_passes): Move pass_fold_builtins
16879 and pass_dce earlier with -Og.
16881 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16883 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16884 <BIT_FIELD_REF>: Remove trailing TAB.
16885 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16888 2013-06-24 Martin Jambor <mjambor@suse.cz>
16890 PR tree-optimization/57358
16891 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16892 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16893 (ipa_analyze_params_uses): Generate pessimistic info when true.
16895 2013-06-24 Martin Jambor <mjambor@suse.cz>
16897 PR tree-optimization/57539
16898 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16899 global.inlined_to of the new node to it. All callers changed.
16900 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16901 inlining_into, pass it to cgraph_clone_node.
16902 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16903 ipa_free_edge_args_substructures.
16904 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16905 rdesc linked list. Do not assert rdesc edges have inlined caller.
16906 Assert we have found an rdesc in the rdesc list.
16908 2013-06-24 Richard Biener <rguenther@suse.de>
16910 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16911 (pointer_set_lookup): Declare.
16912 (class pointer_map): New template class implementing a
16913 generic pointer to T map.
16914 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16915 pointer_map<T>::contains, pointer_map<T>::insert,
16916 pointer_map<T>::traverse): New functions.
16917 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16918 (pointer_set_lookup): New function.
16919 (pointer_set_contains): Use pointer_set_lookup.
16920 (pointer_set_insert): Likewise.
16921 (insert_aux): Remove.
16922 (struct pointer_map_t): Embed a pointer_set_t.
16923 (pointer_map_create): Adjust.
16924 (pointer_map_destroy): Likewise.
16925 (pointer_map_contains): Likewise.
16926 (pointer_map_insert): Likewise.
16927 (pointer_map_traverse): Likewise.
16928 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16929 pointer_map<unsigned> instead of a pointer_map_t.
16930 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16931 (streamer_tree_cache_lookup): Likewise.
16932 (streamer_tree_cache_create): Likewise.
16933 (streamer_tree_cache_delete): Likewise.
16934 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16935 pointer_map<unsigned> instead of a pointer_map_t.
16936 (lto_init_tree_ref_encoder): Adjust.
16937 (lto_destroy_tree_ref_encoder): Likewise.
16938 * lto-section-out.c (lto_output_decl_index): Likewise.
16939 (lto_record_function_out_decl_state): Likewise.
16940 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16942 2013-06-24 Richard Biener <rguenther@suse.de>
16944 PR tree-optimization/57488
16945 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16947 2013-06-24 Alan Modra <amodra@gmail.com>
16949 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16950 (gen_easy_altivec_constant): Likewise.
16951 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16952 easy_vector_constant_msb): Likewise.
16954 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16957 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16958 add missing return true.
16960 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16963 * config/sh/predicates.md (general_extend_operand): Invoke
16964 general_movsrc_operand for memory operands.
16965 (general_movsrc_operand): Allow reg+reg addressing, do not use
16966 general_operand for memory operands.
16968 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16970 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16971 when current target options does not apply.
16972 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16973 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16974 * config/i386/bmiintrin.h: Pass appropriate target
16975 attributes to header.
16976 * config/i386/mmintrin.h: Ditto.
16977 * config/i386/nmmintrin.h: Ditto.
16978 * config/i386/avx2intrin.h: Ditto.
16979 * config/i386/fxsrintrin.h: Ditto.
16980 * config/i386/tbmintrin.h: Ditto.
16981 * config/i386/xsaveintrin.h: Ditto.
16982 * config/i386/f16cintrin.h: Ditto.
16983 * config/i386/xtestintrin.h: Ditto.
16984 * config/i386/xsaveoptintrin.h: Ditto.
16985 * config/i386/bmi2intrin.h: Ditto.
16986 * config/i386/lzcntintrin.h: Ditto.
16987 * config/i386/smmintrin.h: Ditto.
16988 * config/i386/wmmintrin.h: Ditto.
16989 * config/i386/x86intrin.h: Remove all header include guards.
16990 * config/i386/prfchwintrin.h: Ditto.
16991 * config/i386/pmmintrin.h: Ditto.
16992 * config/i386/tmmintrin.h: Ditto.
16993 * config/i386/xmmintrin.h: Ditto.
16994 * config/i386/popcntintrin.h: Ditto.
16995 * config/i386/rdseedintrin.h: Ditto.
16996 * config/i386/ammintrin.h: Ditto.
16997 * config/i386/emmintrin.h: Ditto.
16998 * config/i386/immintrin.h: Remove all header include guards.
16999 * config/i386/fma4intrin.h: Ditto.
17000 * config/i386/lwpintrin.h: Ditto.
17001 * config/i386/xopintrin.h: Ditto.
17002 * config/i386/ia32intrin.h: Ditto.
17003 * config/i386/avxintrin.h: Ditto.
17004 * config/i386/rtmintrin.h: Ditto.
17005 * config/i386/fmaintrin.h: Ditto.
17006 * config/i386/mm3dnow.h: Ditto.
17008 2013-06-22 Sriraman Tallam <tmsriram@google.com>
17010 * common/config/i386/i386-common.c: Handle LZCNT.
17012 2013-06-22 Andi Kleen <ak@linux.intel.com>
17014 * doc/extend.texi: Use __atomic_store_n instead of
17015 __atomic_store in HLE example.
17017 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
17019 * config/sh/sh.c: Remove <cstdlib> workaround.
17021 2013-06-21 Andi Kleen <ak@linux.intel.com>
17023 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
17025 2013-06-21 Andi Kleen <ak@linux.intel.com>
17027 * doc/extend.texi: Document that __atomic_clear and
17028 __atomic_test_and_set should only be used with bool.
17030 2013-06-20 Jan Hubicka <jh@suse.cz>
17032 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
17033 types_same_for_odr.
17034 * tree.c (decls_same_for_odr): New function.
17035 (same_for_edr): New function.
17036 (types_same_for_odr): New function.
17037 (get_binfo_at_offset): Use it.
17038 * tree.h (types_same_for_odr): Declare.
17040 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
17041 Jason Merrill <jason@redhat.com>
17043 * system.h: Include <cstdlib> as well as <stdlib.h>.
17045 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
17048 * config/i386/i386.c (construct_container): Report error if
17049 long double is used with disabled x87 float returns.
17051 2013-06-20 Jan Hubicka <jh@suse.cz>
17053 * lto-cgraph.c (input_symtab): Do not set cgraph state.
17055 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
17057 PR rtl-optimization/57425
17058 PR rtl-optimization/57569
17059 * alias.c (write_dependence_p): Remove parameters mem_mode and
17060 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
17061 Changed all callers.
17062 (canon_anti_dependence): Get comments and semantics in sync.
17063 Add parameter mem_canonicalized. Changed all callers.
17064 * rtl.h (canon_anti_dependence): Update prototype.
17066 2013-06-20 Richard Biener <rguenther@suse.de>
17068 * data-streamer-in.c (streamer_read_uhwi): Optimize single
17069 byte case, inline streamer_read_uchar and defer section
17072 2013-06-20 Richard Biener <rguenther@suse.de>
17074 PR tree-optimization/57584
17075 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
17076 SSA names into the expanded expression that take part in
17077 abnormal coalescing.
17079 2013-06-19 Sharad Singhai <singhai@google.com>
17081 * gcov.c (print_usage): Handle new option.
17082 (process_args): Ditto.
17083 (get_gcov_intermediate_filename): New function.
17084 (output_intermediate_file): New function.
17085 (output_gcov_file): New function
17086 (generate_results): Handle new option.
17087 (release_function): Relase demangled name.
17088 (read_graph_file): Handle demangled name.
17089 (output_lines): Ditto.
17090 * doc/gcov.texi: Document gcov intermediate format.
17092 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
17095 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
17096 (lra_emit_add): Use the functions. Add comment about Y as an
17099 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
17102 * collect2.c (collect_atexit): New.
17103 (collect_exit): Delete.
17104 (main): Register collect_atexit with atexit.
17105 (collect_wait): Change collect_exit to exit.
17107 * collect2.h (collect_exit): Delete.
17108 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
17110 2013-06-19 Wei Mi <wmi@google.com>
17112 PR rtl-optimization/57518
17113 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
17114 if regno is used in paradoxical subreg.
17115 (update_equiv_regs): Check pdx_subregs[regno] before
17116 set a reg to be equivalent with a mem.
17118 2013-06-19 Matthias Klose <doko@ubuntu.com>
17121 * file-find.h (find_a_file): Add a mode parameter.
17122 * file-find.c (find_a_file): Likewise.
17123 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
17124 with X_OK for the executables.
17125 * collect2.c (main): Call find_a_file with X_OK.
17127 2013-06-19 Steve Ellcey <sellcey@mips.com>
17130 * config/mips/mips.md (casesi_internal_mips16_<mode>):
17131 Use NEXT_INSN instead of next_real_insn.
17133 2013-06-19 Jan Hubicka <jh@suse.cz>
17135 * cgraph.h (const_value_known_p): Replace by ...
17136 (ctor_for_folding): .. this one.
17137 * cgraphunit.c (process_function_and_variable_attributes): Use it.
17138 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
17139 * expr.c (expand_expr_real_1): Likewise.
17140 (string_constant): Likewise.
17141 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
17142 * ipa.c (process_references): Likewise.
17143 (symtab_remove_unreachable_nodes): Likewise.
17144 * ipa-inline-analysis.c (param_change_prob): Likewise.
17145 * gimple-fold.c (canonicalize_constructor_val): Likewise.
17146 (get_base_constructor): Likwise.
17147 * varpool.c (varpool_remove_node): Likewise.
17148 (varpool_remove_initializer): LIkewise.
17149 (dump_varpool_node): LIkwise.
17150 (const_value_known_p): Rewrite to ...
17151 (ctor_for_folding): ... this one.
17153 2013-06-19 Jakub Jelinek <jakub@redhat.com>
17156 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
17157 PERSONALITY in $PATH derived prefixes.
17159 2013-06-19 Jeff Law <law@redhat.com>
17161 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
17164 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
17165 (simplify_bitwise_binary): Use it to simpify certain binary ops on
17168 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
17170 * config/arm/vfp.md: Move VFP instruction classification documentation
17172 * config/arm/arm.md: ... here. Update instruction classification
17175 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
17177 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
17178 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
17179 pattern. Use more efficient sequences on ARMv5 and Thumb2.
17181 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
17184 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17185 with NEXT_INSN. Use tablejump_p to check for jump table data
17188 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
17191 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
17192 that now in C++ the value is correct per the C++ standards.
17194 2013-06-19 Richard Biener <rguenther@suse.de>
17196 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
17197 for global context.
17199 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17202 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17205 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17206 with next_active_insn.
17208 2013-06-18 Sriraman Tallam <tmsriram@google.com>
17210 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
17211 functions are inlined during failures to flag an error.
17212 * tree-inline.c (expand_call_inline): Allow the error to be flagged
17213 in early inline pass.
17215 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
17217 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
17220 2013-06-18 Julian Brown <julian@codesourcery.com>
17222 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
17223 Permit virtual register pre-reload if !strict.
17224 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
17226 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
17228 * config/arm/neon.md (movmisalign<mode>): Use
17229 neon_perm_struct_or_reg_operand instead of
17230 neon_struct_or_register_operand.
17231 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
17232 neon_permissive_struct_operand instead of neon_struct_operand.
17233 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
17234 neon_vector_mem_operand.
17235 * config/arm/predicates.md (neon_struct_operand): Adjust call to
17236 neon_vector_mem_operand.
17237 (neon_permissive_struct_operand): New.
17238 (neon_struct_or_register_operand): Rename to...
17239 (neon_perm_struct_or_reg_operand): This. Adjust call to
17240 neon_vector_mem_operand.
17242 2013-06-18 Richard Biener <rguenther@suse.de>
17244 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
17245 * lto-streamer.h: Include pointer-set.h.
17246 (struct lto_decl_slot): Remove.
17247 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
17248 Remove next_index entry.
17249 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17250 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17251 (lto_init_tree_ref_encoder): Adjust.
17252 (lto_destroy_tree_ref_encoder): Likewise.
17253 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17254 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17255 (lto_output_decl_index): Adjust.
17256 (lto_new_out_decl_state): Likewise.
17257 (lto_record_function_out_decl_state): Likewise.
17258 * lto-streamer-out.c (copy_function): Likewise.
17260 2013-06-18 Richard Biener <rguenther@suse.de>
17262 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
17263 * cgraphunit.c: Include cfgloop.h.
17264 (init_lowered_empty_function): Initialize the loop tree.
17265 (assemble_thunk): Insert new BBs into loops.
17267 2013-06-18 Richard Biener <rguenther@suse.de>
17269 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
17270 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
17271 the map from cache entry to cache index optional.
17272 (streamer_tree_cache_replace_tree): Adjust accordingly.
17273 (streamer_tree_cache_append): Likewise.
17274 (streamer_tree_cache_delete): Likewise.
17275 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
17276 streamer cache map from cache entry to cache index.
17277 * lto-streamer-out.c (create_output_block): Adjust.
17279 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
17281 * config/arm/arm.md (attribute "insn"): Move multiplication and
17282 division attributes to...
17283 (attribute "type"): ... here. Remove mult.
17284 (attribute "mul32"): New attribute.
17285 (attribute "mul64"): Add umaal.
17286 (*arm_mulsi3): Update attributes.
17287 (*arm_mulsi3_v6): Likewise.
17288 (*thumb_mulsi3): Likewise.
17289 (*thumb_mulsi3_v6): Likewise.
17290 (*mulsi3_compare0): Likewise.
17291 (*mulsi3_compare0_v6): Likewise.
17292 (*mulsi_compare0_scratch): Likewise.
17293 (*mulsi_compare0_scratch_v6): Likewise.
17294 (*mulsi3addsi): Likewise.
17295 (*mulsi3addsi_v6): Likewise.
17296 (*mulsi3addsi_compare0): Likewise.
17297 (*mulsi3addsi_compare0_v6): Likewise.
17298 (*mulsi3addsi_compare0_scratch): Likewise.
17299 (*mulsi3addsi_compare0_scratch_v6): Likewise.
17300 (*mulsi3subsi): Likewise.
17301 (*mulsidi3adddi): Likewise.
17302 (*mulsi3addsi_v6): Likewise.
17303 (*mulsidi3adddi_v6): Likewise.
17304 (*mulsidi3_nov6): Likewise.
17305 (*mulsidi3_v6): Likewise.
17306 (*umulsidi3_nov6): Likewise.
17307 (*umulsidi3_v6): Likewise.
17308 (*umulsidi3adddi): Likewise.
17309 (*umulsidi3adddi_v6): Likewise.
17310 (*smulsi3_highpart_nov6): Likewise.
17311 (*smulsi3_highpart_v6): Likewise.
17312 (*umulsi3_highpart_nov6): Likewise.
17313 (*umulsi3_highpart_v6): Likewise.
17314 (mulhisi3): Likewise.
17315 (*mulhisi3tb): Likewise.
17316 (*mulhisi3bt): Likewise.
17317 (*mulhisi3tt): Likewise.
17318 (maddhisi4): Likewise.
17319 (*maddhisi4tb): Likewise.
17320 (*maddhisi4tt): Likewise.
17321 (maddhidi4): Likewise.
17322 (*maddhidi4tb): Likewise.
17323 (*maddhidi4tt): Likewise.
17324 (divsi3): Likewise.
17325 (udivsi3): Likewise.
17326 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
17327 (thumb2_mulsi_short_compare0): Likewise.
17328 (thumb2_mulsi_short_compare0_scratch): Likewise.
17329 * config/arm/arm1020e.md (1020mult1): Update attribute change.
17330 (1020mult2): Likewise.
17331 (1020mult3): Likewise.
17332 (1020mult4): Likewise.
17333 (1020mult5): Likewise.
17334 (1020mult6): Likewise.
17335 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
17337 (cortex_a15_mult64): Likewise.
17338 (cortex_a15_sdiv): Likewise.
17339 (cortex_a15_udiv): Likewise.
17340 * config/arm/arm1026ejs.md (mult1): Update attribute change.
17346 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
17347 (pj4_ir_div): Likewise.
17348 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
17349 (11_mult2): Likewise.
17350 (11_mult3): Likewise.
17351 (11_mult4): Likewise.
17352 (11_mult5): Likewise.
17353 (11_mult6): Likewise.
17354 (11_mult7): Likewise.
17355 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
17356 (cortex_a8_mla): Likewise.
17357 (cortex_a8_mull): Likewise.
17358 (cortex_a8_smulwy): Likewise.
17359 (cortex_a8_smlald): Likewise.
17360 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
17361 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
17362 (cortex_r4_mul_3): Likewise.
17363 (cortex_r4_mla_4): Likewise.
17364 (cortex_r4_mla_3): Likewise.
17365 (cortex_r4_smlald): Likewise.
17366 (cortex_r4_mull): Likewise.
17367 (cortex_r4_sdiv): Likewise.
17368 (cortex_r4_udiv): Likewise.
17369 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
17370 (cortex_a7_idiv): Likewise.
17371 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
17372 (9_mult2): Likewise.
17373 (9_mult3): Likewise.
17374 (9_mult4): Likewise.
17375 (9_mult5): Likewise.
17376 (9_mult6): Likewise.
17377 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
17378 (cortex_a53_sdiv): Likewise.
17379 (cortex_a53_udiv): Likewise.
17380 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
17381 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
17382 (mp626_mult2): Likewise.
17383 (mp626_mult3): Likewise.
17384 (mp626_mult4): Likewise.
17385 * config/arm/fa526.md (526_mult1): Update attribute change.
17386 (526_mult2): Likewise.
17387 * config/arm/arm-generic.md (mult): Update attribute change.
17388 (mult_ldsched_strongarm): Likewise.
17389 (mult_ldsched): Likewise.
17390 (multi_cycle): Likewise.
17391 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
17392 * config/arm/fa606te.md (606te_mult1): Update attribute change.
17393 (606te_mult2): Likewise.
17394 (606te_mult3): Likewise.
17395 (606te_mult4): Likewise.
17396 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
17397 (cortex_a9_mac16): Likewise.
17398 (cortex_a9_multiply): Likewise.
17399 (cortex_a9_mac): Likewise.
17400 (cortex_a9_multiply_long): Likewise.
17401 * config/arm/fa626te.md (626te_mult1): Update attribute change.
17402 (626te_mult2): Likewise.
17403 (626te_mult3): Likewise.
17404 (626te_mult4): Likewise.
17406 2013-06-18 Richard Biener <rguenther@suse.de>
17409 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17411 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17414 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17415 with next_active_insn.
17417 2013-06-18 Alan Modra <amodra@gmail.com>
17419 * config/rs6000/rs6000.h (enum data_align): New.
17420 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17421 (DATA_ABI_ALIGNMENT): Define.
17422 (CONSTANT_ALIGNMENT): Correct comment.
17423 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17424 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17426 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17428 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17429 ATTRIBUTE_UNUSED marking.
17431 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17433 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17434 alternative and update.
17435 (aarch64_dup_lanedi): Delete.
17436 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17437 * config/aarch64/aarch64-simd-builtins.def: Update.
17439 2013-06-17 Richard Biener <rguenther@suse.de>
17441 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17442 (lto_input_scc): Declare.
17443 (lto_input_tree_1): Likewise.
17444 (struct lto_stats_d): Add num_tree_bodies_output and
17445 num_pickle_refs_output.
17446 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17447 (lto_read_tree_1): Split out from ...
17448 (lto_read_tree): ... this.
17449 (lto_input_scc): New function.
17450 (lto_input_tree_1): Split out from ...
17451 (lto_input_tree): ... this. Handle LTO_tree_scc.
17452 (lto_data_in_create): Create the streamer cache without hashes.
17453 * lto-streamer-out.c (create_output_block): Create the streamer
17454 cache with hashes when not doing WPA.
17455 (lto_write_tree_1): Split out from ...
17456 (lto_write_tree): ... this.
17457 (get_symbol_initial_value): New function.
17458 (lto_output_tree_1): Split out from ...
17459 (lto_output_tree): ... this. Write trees as series of SCCs
17460 using a DFS walk via DFS_write_tree.
17461 (struct sccs, struct scc_entry): New types.
17462 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17463 (DFS_write_tree_body): New function.
17464 (DFS_write_tree): Likewise.
17465 (hash_tree): Likewise.
17466 (scc_entry_compare): Likewise.
17467 (hash_scc): Likewise.
17468 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17469 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17470 TREE_CHAIN as regular reference.
17471 (streamer_read_integer_cst): Remove.
17472 (streamer_get_pickled_tree): Adjust.
17473 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17474 of DECL_EXTERNALs in BLOCK_VARS for now.
17475 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17477 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17478 Add hash value argument and record that if hashes are recorded
17480 (streamer_tree_cache_insert_1): Adjust.
17481 (streamer_tree_cache_insert): Likewise.
17482 (streamer_tree_cache_insert_at): Rename to ...
17483 (streamer_tree_cache_replace_tree): ... this and adjust.
17484 (streamer_tree_cache_append): Adjust.
17485 (record_common_node): Likewise.
17486 (streamer_tree_cache_create): Add argument whether to
17487 record hash values together with trees.
17488 (streamer_tree_cache_delete): Adjust.
17489 * tree-streamer.h (struct streamer_tree_cache_d): Add
17491 (streamer_read_integer_cst): Remove.
17492 (streamer_tree_cache_insert): Adjust.
17493 (streamer_tree_cache_append): Likewise.
17494 (streamer_tree_cache_insert_at): Rename to ...
17495 (streamer_tree_cache_replace_tree): ... this and adjust.
17496 (streamer_tree_cache_create): Add argument whether to record hashes.
17497 (streamer_tree_cache_get): Rename to ...
17498 (streamer_tree_cache_get_tree): ... this.
17499 (streamer_tree_cache_get_hash): New function.
17500 * tree.c (cache_integer_cst): New function.
17501 * tree.h (cache_integer_cst): Declare.
17502 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17503 * lto-symtab.c (lto_varpool_replace_node): Only release
17504 DECL_INITIAL of non-prevailing decls.
17505 * varpool.c (varpool_remove_initializer): Do not release
17506 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17508 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17510 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17511 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17512 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17513 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17514 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17515 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17516 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17517 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17518 instead of TARGET_64BIT.
17519 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17520 Require ISA_HAS_<D>DIV.
17522 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17524 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17525 (mips*-*-linux*): Move default with_llsc setting to where other
17527 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17529 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17530 Likewise. Remove default with_tune setting. Move default float
17531 setting to its own block. Handle with_llsc in the same block as above.
17533 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17535 PR rtl-optimization/57425
17536 PR rtl-optimization/57569
17537 * alias.c (write_dependence_p): Add new parameters mem_mode,
17538 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17539 Changed all callers.
17540 (canon_anti_dependence): New function.
17541 * cse.c (check_dependence): Use canon_anti_dependence.
17542 * cselib.c (cselib_invalidate_mem): Likewise.
17543 * rtl.h (canon_anti_dependence): Declare.
17545 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17547 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17548 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17549 ".set mips3" for 64-bit targets.
17551 2013-06-15 Dehao Chen <dehao@google.com>
17553 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17554 * gimple-low.c (gimple_check_call_matching_types): Likewise.
17555 (gimple_check_call_args): Likewise.
17556 * value-prof.c (check_ic_target): Likewise.
17557 * ipa-inline.c (early_inliner): Likewise.
17558 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
17559 * cgraph.c (cgraph_create_edge_1): Likewise.
17560 (cgraph_make_edge_direct): Likewise.
17562 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
17565 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
17566 rs6000_output_move_128bit to handle emitting quad memory
17567 operations. Set attribute length to 8 bytes.
17569 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
17571 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
17573 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
17574 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
17575 (aarch64_<su>mlsl<mode>): Likewise.
17577 2013-06-14 Mike Stump <mikestump@comcast.net>
17579 * Makefile.in (TARGET_H): Add insn-codes.h.
17581 2013-06-14 Alan Modra <amodra@gmail.com>
17583 PR middle-end/57134
17584 PR middle-end/57586
17585 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
17586 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
17587 bitfield expansion when EXPAND_MEMORY.
17588 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
17590 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
17592 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
17593 test for clearing quad memory on 32-bit later.
17595 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17597 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
17598 (fold_negate_expr): Likewise.
17599 (fold_real_zero_addition_p): Handle vectors.
17600 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
17602 2013-06-14 Alan Modra <amodra@gmail.com>
17604 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17606 2013-06-13 Jan Hubicka <jh@suse.cz>
17608 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17609 Local comdats are not externally visible.
17610 * symtab.c (dump_symtab_base): Dump externally visible.
17611 (verify_symtab_base): Verify back links in the symtab hash.
17613 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17615 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17616 CONVERT_EXPR as equal nodes.
17618 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17620 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17622 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17624 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17625 Generalize to complex and vector.
17626 * tree.c (build_all_ones_cst): New function.
17627 * tree.h (build_all_ones_cst): Declare it.
17629 2013-06-13 Alan Modra <amodra@gmail.com>
17631 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17632 * config/rs6000/rs6000.md (signbittf2): New insn.
17633 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17634 (abstf2_internal, cmptf_internal2): Likewise.
17635 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17637 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17638 Pat Haugen <pthaugen@us.ibm.com>
17639 Peter Bergner <bergner@vnet.ibm.com>
17641 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17642 power8 byte, half-word, and quad-word atomic instructions.
17643 (emit_store_conditional): Likewise.
17644 (rs6000_expand_atomic_compare_and_swap): Likewise.
17645 (rs6000_expand_atomic_op): Likewise.
17647 * config/rs6000/sync.md (larx): Add new modes for power8.
17649 (AINT): New mode iterator to include TImode as well as normal
17650 integer modes on power8.
17651 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17652 that VSX registers are not considered. Use AINT mode iterator
17653 instead of INT1 to allow inclusion of quad word atomic operations
17655 (load_locked<mode>): Likewise.
17656 (store_conditional<mode>): Likewise.
17657 (atomic_compare_and_swap<mode>): Likewise.
17658 (atomic_exchange<mode>): Likewise.
17659 (atomic_nand<mode>): Likewise.
17660 (atomic_fetch_<fetchop_name><mode>): Likewise.
17661 (atomic_nand_fetch<mode>): Likewise.
17662 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17664 (ATOMIC): On power8, add QImode, HImode modes.
17665 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17666 modes that promote to SImode.
17667 (load_lockedti): Convert TImode arguments to PTImode, so that we
17668 get a guaranteed even/odd register pair.
17669 (load_lockedpti): Likewise.
17670 (store_conditionalti): Likewise.
17671 (store_conditionalpti): Likewise.
17673 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17674 atomic load/store instructions.
17677 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17679 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17681 (insn_count): New attribute, with most cases extracted from...
17682 (length): ...here. Redefine most cases in terms of insn_count.
17683 (single_insn): Delete.
17684 (can_delay): Use insn_count to check for single instructions.
17685 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17686 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17687 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17688 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17689 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17690 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17691 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17692 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17693 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17694 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17695 rather than "length".
17696 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17697 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17698 Use "insn_count" rather than "length".
17699 * config/mips/mips-dsp.md
17700 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17701 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17704 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17706 PR tree-optimization/57361
17707 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17709 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17711 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17713 (aarch64_simd_combine<mode>): New instruction expansion.
17714 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17715 function prototype.
17716 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17717 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17719 2013-06-12 Jan Hubicka <jh@suse.cz>
17721 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17722 decl has when in streaming stage.
17723 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17724 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17726 2013-06-12 Roland Stigge <stigge@antcom.de>
17729 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17731 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17733 PR tree-optimization/57537
17734 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17735 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17737 2013-06-12 Richard Biener <rguenther@suse.de>
17739 * data-streamer.h (streamer_write_char_stream): CSE
17740 obs->current_pointer.
17741 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17742 streamer_write_char_stream manually and optimize the resulting loop.
17743 (streamer_write_hwi_stream): Likewise.
17745 2013-06-12 Jan Hubicka <jh@suse.cz>
17747 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17748 * cgraph.h (varpool_create_empty_node): Declare.
17749 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17751 * symtab.c (symtab_unregister_node): Be lax about missin entries
17753 (symtab_get_node): Update comment.
17754 * varpool.c (varpool_create_empty_node): Break out from ...
17755 (varpool_node_for_decl): ... here.
17756 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17758 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17760 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17761 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17762 part. Use straight-line flow at the end.
17763 <COMPONENT_REF>: Remove superfluous else.
17764 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17766 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17769 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17770 target hook even for !TREE_PUBLIC decls. If no resolution info
17771 is available, return false for common and external decls.
17773 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17775 * config/rl78/constraints.md (U): New constraint.
17776 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17779 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17782 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17783 to allow returning address to AT_PLATFORM name.
17785 2013-06-11 Jan Hubicka <jh@suse.cz>
17787 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17788 * cgraph.h (symtab_node_base): Add weakref flag.
17789 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17790 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17791 (output_weakrefs): Use weakref flag.
17792 * fold-const.c (simple_operand_p): Handle WEAK.
17793 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17794 * ipa.c (varpool_externally_visible_p): Drop weakref.
17795 (function_and_variable_visibility): Update comment; fix weakref
17796 sanity checks; do not clear DECL_WEAK on them.
17797 * lto-cgraph.c (lto_output_node): update.
17798 (lto_output_varpool_node): Update.
17799 (input_overwrite_node): Update.
17800 (input_node): Update.
17801 (input_varpool_node): Update.
17802 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17803 (lto_symtab_merge_symbols): Add sanity check.
17804 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17805 * passes.c (rest_of_decl_compilation): Set static flag, too.
17806 * symtab.c (dump_symtab_base): Dump weakref.
17807 (verify_symtab_base): Sanity check weakrefs.
17808 (symtab_make_decl_local): Remove duplicated code.
17809 (symtab_alias_ultimate_target): Simplify.
17810 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17812 2013-06-11 Tom de Vries <tom@codesourcery.com>
17814 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17815 sequence_vect == NULL.
17817 2013-06-11 DJ Delorie <dj@redhat.com>
17819 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17820 (rl78_unwind_word_mode): New.
17822 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17824 * final.c (debug_prefix_maps): Make static.
17826 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17828 * function.c (initial_trampoline): Remove stray copy.
17830 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17832 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17834 2013-06-11 Martin Jambor <mjambor@suse.cz>
17836 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17837 within bounds at the beginning of the function.
17839 2013-06-11 Alan Modra <amodra@gmail.com>
17841 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17843 (get_named_section): Don't NULL !DECL_P decl.
17845 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17847 * doc/invoke.texi (core-avx2): Document.
17849 (atom): Updated with MOVBE.
17851 2013-06-11 Richard Biener <rguenther@suse.de>
17853 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17855 2013-06-11 Anton Blanchard <anton@samba.org>
17857 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17858 correct shift value in little-endian mode.
17860 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17863 * varasm.c (get_variable_align): Move #endif to the right place.
17865 2013-06-10 Cary Coutant <ccoutant@google.com>
17867 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17868 for hash so that hash table traversal order is deterministic.
17870 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17871 Pat Haugen <pthaugen@us.ibm.com>
17872 Peter Bergner <bergner@vnet.ibm.com>
17874 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17875 of vectors in GPRS if they are direct moves or quad word load or
17878 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17880 (direct_move_p): Likewise.
17881 (quad_load_store_p): Likewise.
17883 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17884 classes into bins based on the physical register type.
17885 (reg_class_to_reg_type): Likewise.
17886 (IS_STD_REG_TYPE): Likewise.
17887 (IS_FP_VECT_REG_TYPE): Likewise.
17888 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17889 use direct move instructions.
17890 (reload_gpr_vsx): Likewise.
17891 (reload_vsx_gpr): Likewise.
17892 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17893 information that is a simplification of register classes. Also
17894 precalculate direct move reload helpers.
17895 (direct_move_p): New function to return true if the operation can
17896 be done as a direct move instruciton.
17897 (quad_load_store_p): New function to return true if the operation
17898 is a quad memory operation.
17899 (rs6000_legitimize_address): If quad memory, only allow register
17900 indirect for TImode addresses.
17901 (rs6000_legitimate_address_p): Likewise.
17902 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17903 (rs6000_reload_register_type): Likewise.
17904 (register_to_reg_type): Return register type.
17905 (rs6000_secondary_reload_simple_move): New helper function for
17906 secondary reload and secondary memory needed to identify anything
17907 that is a simple move, and does not need reloading.
17908 (rs6000_secondary_reload_direct_move): New helper function for
17909 secondary reload to identify cases that can be done with several
17910 instructions via the direct move instructions.
17911 (rs6000_secondary_reload_move): New helper function for secondary
17912 reload to identify moves between register types that can be done.
17913 (rs6000_secondary_reload): Add support for quad memory operations
17914 and for direct move.
17915 (rs6000_secondary_memory_needed): Likewise.
17916 (rs6000_debug_secondary_memory_needed): Change argument names.
17917 (rs6000_output_move_128bit): New function to return the move to
17918 use for 128-bit moves, including knowing about the various
17919 limitations of quad memory operations.
17921 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17922 memory operations. call rs6000_output_move_128bit for the actual
17923 instruciton(s) to generate.
17924 (vsx_movti_64bit): Likewise.
17926 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17927 (UNSPEC_P8V_MTVSRWZ): Likewise.
17928 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17929 (UNSPEC_P8V_MTVSRD): Likewise.
17930 (UNSPEC_P8V_XXPERMDI): Likewise.
17931 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17932 (UNSPEC_FUSION_GPR): Likewise.
17933 (FMOVE128_GPR): New iterator for direct move.
17934 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17935 (f32_sv): Likewise.
17936 (f32_dm): Likewise.
17937 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17938 loads and direct move instructions.
17939 (zero_extendsidi2_lfiwzx): Likewise.
17940 (extendsidi2_lfiwax): Likewise.
17941 (extendsidi2_nocell): Likewise.
17942 (floatsi<mode>2_lfiwax): Likewise.
17943 (lfiwax): Likewise.
17944 (floatunssi<mode>2_lfiwzx): Likewise.
17945 (lfiwzx): Likewise.
17946 (fix_trunc<mode>_stfiwx): Likewise.
17947 (fixuns_trunc<mode>_stfiwx): Likewise.
17948 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17949 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17950 (parity<mode>2_cmpb): Set length/type attr.
17951 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17952 for 'mr.' to fast_compare.
17953 (bpermd_<mode>): Change type attr to popcnt.
17954 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17955 (p8_mtvsrwz_1): Likewise.
17956 (p8_mtvsrwz_2): Likewise.
17957 (reload_fpr_from_gpr<mode>): Likewise.
17958 (p8_mtvsrd_1): Likewise.
17959 (p8_mtvsrd_2): Likewise.
17960 (p8_xxpermdi_<mode>): Likewise.
17961 (reload_vsx_from_gpr<mode>): Likewise.
17962 (reload_vsx_from_gprsf): Likewise.
17963 (p8_mfvsrd_3_<mode>): LIkewise.
17964 (reload_gpr_from_vsx<mode>): Likewise.
17965 (reload_gpr_from_vsxsf): Likewise.
17966 (p8_mfvsrd_4_disf): Likewise.
17967 (multi-word GPR splits): Do not split direct moves or quad memory
17970 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17972 * tree-into-ssa.c (interesting_blocks): Make static.
17974 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17977 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17978 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17979 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17980 (get_variable_align): New function.
17981 (get_variable_section, emit_bss, emit_common,
17982 assemble_variable_contents, place_block_symbol): Use
17983 get_variable_align instead of DECL_ALIGN.
17984 (assemble_noswitch_variable): Add align argument, use it
17985 instead of DECL_ALIGN.
17986 (assemble_variable): Adjust caller. Use get_variable_align
17987 instead of DECL_ALIGN.
17988 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17990 (DATA_ABI_ALIGNMENT): Define.
17991 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17992 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17993 opt is false, only return the psABI mandated alignment increase.
17994 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17995 (DATA_ABI_ALIGNMENT): ... this.
17996 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17997 (DATA_ABI_ALIGNMENT): ... this.
17998 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17999 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
18000 (DATA_ABI_ALIGNMENT): ... this.
18001 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
18002 * doc/tm.texi: Regenerated.
18004 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
18006 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
18007 cmp_code to construct REG_EQUAL note.
18009 2013-06-09 Jakub Jelinek <jakub@redhat.com>
18012 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
18013 that operands[2] doesn't overlap with operands[0].
18015 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
18016 Jan Hubicka <jh@suse.cz>
18018 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
18019 hack to mark symbols as used.
18021 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
18023 PR rtl-optimization/57559
18024 * lra-constraints.c (process_alt_operands): Don't discourage
18025 memory with known offset for offsetable memory constraint.
18026 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
18028 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
18030 * varasm.c (struct oc_local_state): Reorder fields.
18031 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
18032 and adjust accordingly.
18033 (output_constructor): Reorder initialization code and adjust call to
18034 output_constructor_bitfield.
18036 2013-06-07 Jan Hubicka <jh@suse.cz>
18038 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
18040 2013-06-07 David Malcolm <dmalcolm@redhat.com>
18042 * tree-object-size.c (unknown): Make const.
18044 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18046 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
18047 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
18048 for last alternative in the cpu_facility attribute.
18050 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18053 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
18054 (xordi3): Change operand 2 constraint to arm_xordi_operand.
18055 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
18056 * config/arm/constraints.md (Dg): New constraint.
18057 * config/arm/neon.md (xordi3_neon): Remove.
18058 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
18059 * config/arm/predicates.md (arm_xordi_operand): New predicate.
18061 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18063 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
18064 Clean up alternatives.
18066 2013-06-07 Alan Modra <amodra@gmail.com>
18068 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
18071 2013-06-07 Alan Modra <amodra@gmail.com>
18073 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
18075 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18077 * config/arm/constraints.md (Df): New constraint.
18078 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
18079 Correct length attribute for last two alternatives.
18081 2013-06-07 Alan Modra <amodra@gmail.com>
18083 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
18084 override user -mfp-in-toc.
18085 (offsettable_ok_by_alignment): Consider just the current access
18086 rather than the whole object, unless BLKmode. Handle
18087 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
18088 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
18089 for -mcmodel=medium.
18090 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
18091 override user -mfp-in-toc or -msum-in-toc. Default to
18092 -mno-fp-in-toc for -mcmodel=medium.
18094 2013-06-06 DJ Delorie <dj@redhat.com>
18096 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
18097 TARGET_VALID_POINTER_MODE.
18099 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
18100 Pat Haugen <pthaugen@us.ibm.com>
18101 Peter Bergner <bergner@vnet.ibm.com>
18103 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
18104 Document new power8 builtins.
18106 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
18107 condition code register, to allow 128-bit logical operations to be
18108 done in the VSX or GPR registers.
18109 (nor<mode>3): Use the canonical form for nor.
18110 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
18111 vclz*, and vpopcnt* vector instructions.
18112 (nand<mode>3): Likewise.
18113 (orc<mode>3): Likewise.
18114 (clz<mode>2): LIkewise.
18115 (popcount<mode>2): Likewise.
18117 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
18118 that only the GPRs are recognized.
18120 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18121 support for new power8 builtins.
18123 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
18125 (xscvdpspn): Likewise.
18131 (vpopcnt): Likewise.
18132 (vpopcntb): Likewise.
18133 (vpopcnth): Likewise.
18134 (vpopcntw): Likewise.
18135 (vpopcntd): Likewise.
18137 (vmrgew): Likewise.
18138 (vmrgow): Likewise.
18140 (eqv_v16qi3): Likewise.
18141 (eqv_v8hi3): Likewise.
18142 (eqv_v4si3): Likewise.
18143 (eqv_v2di3): Likewise.
18144 (eqv_v4sf3): Likewise.
18145 (eqv_v2df3): Likewise.
18147 (nand_v16qi3): Likewise.
18148 (nand_v8hi3): Likewise.
18149 (nand_v4si3): Likewise.
18150 (nand_v2di3): Likewise.
18151 (nand_v4sf3): Likewise.
18152 (nand_v2df3): Likewise.
18154 (orc_v16qi3): Likewise.
18155 (orc_v8hi3): Likewise.
18156 (orc_v4si3): Likewise.
18157 (orc_v2di3): Likewise.
18158 (orc_v4sf3): Likewise.
18159 (orc_v2df3): Likewise.
18161 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
18162 allow power8 quad mode in 64-bit.
18163 (rs6000_builtin_vectorized_function): Add support to vectorize
18164 ISA 2.07 count leading zeros, population count builtins.
18165 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
18166 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
18167 (builtin_function_type): Add vgbbd builtin function which takes an
18169 (altivec_expand_vec_perm_const): Add support for new power8 merge
18172 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
18173 that does not include TImdoe for use with 32-bit.
18174 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
18176 (UNSPEC_VSX_CVDPSPN): Likewise.
18177 (vsx_xscvdpspn): Likewise.
18178 (vsx_xscvspdpn): Likewise.
18179 (vsx_xscvdpspn_scalar): Likewise.
18180 (vsx_xscvspdpn_directmove): Likewise.
18181 (vsx_and<mode>3): Split logical operations into 32-bit and
18182 64-bit. Add support to do logical operations on TImode as well as
18183 VSX vector types. Allow logical operations to be done in either
18184 VSX registers or in general purpose registers in 64-bit mode. Add
18185 splitters if GPRs were used. For AND, add clobber of CCmode to
18186 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
18188 (vsx_and<mode>3_32bit): Likewise.
18189 (vsx_and<mode>3_64bit): Likewise.
18190 (vsx_ior<mode>3): Likewise.
18191 (vsx_ior<mode>3_32bit): Likewise.
18192 (vsx_ior<mode>3_64bit): Likewise.
18193 (vsx_xor<mode>3): Likewise.
18194 (vsx_xor<mode>3_32bit): Likewise.
18195 (vsx_xor<mode>3_64bit): Likewise.
18196 (vsx_one_cmpl<mode>2): Likewise.
18197 (vsx_one_cmpl<mode>2_32bit): Likewise.
18198 (vsx_one_cmpl<mode>2_64bit): Likewise.
18199 (vsx_nor<mode>3): Likewise.
18200 (vsx_nor<mode>3_32bit): Likewise.
18201 (vsx_nor<mode>3_64bit): Likewise.
18202 (vsx_andc<mode>3): Likewise.
18203 (vsx_andc<mode>3_32bit): Likewise.
18204 (vsx_andc<mode>3_64bit): Likewise.
18205 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
18206 and xxlorc instructions.
18207 (vsx_eqv<mode>3_64bit): Likewise.
18208 (vsx_nand<mode>3_32bit): Likewise.
18209 (vsx_nand<mode>3_64bit): Likewise.
18210 (vsx_orc<mode>3_32bit): Likewise.
18211 (vsx_orc<mode>3_64bit): Likewise.
18213 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
18215 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
18217 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
18218 (p8_vmrgow): Likewise.
18219 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
18220 GPRs to be split under VSX.
18221 (p8v_clz<mode>2): Add power8 count leading zero support.
18222 (p8v_popcount<mode>2): Add power8 population count support.
18223 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
18226 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
18229 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
18231 (vec_nand): Likewise.
18232 (vec_vclz): Likewise.
18233 (vec_vclzb): Likewise.
18234 (vec_vclzd): Likewise.
18235 (vec_vclzh): Likewise.
18236 (vec_vclzw): Likewise.
18237 (vec_vgbbd): Likewise.
18238 (vec_vmrgew): Likewise.
18239 (vec_vmrgow): Likewise.
18240 (vec_vpopcnt): Likewise.
18241 (vec_vpopcntb): Likewise.
18242 (vec_vpopcntd): Likewise.
18243 (vec_vpopcnth): Likewise.
18244 (vec_vpopcntw): Likewise.
18246 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18248 PR rtl-optimization/57468
18249 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
18252 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18254 PR rtl-optimization/57459
18255 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
18256 type when setting live regs.
18258 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18260 * config/s390/s390.opt (mlra): New option.
18261 * config/s390/s390.c (s390_decompose_address): Check displacement
18262 for all registers for LRA.
18263 (s390_secondary_reload): Don't used secondary reloads for LRA.
18264 (s390_lra_p): New function.
18265 (TARGET_LRA_P): Define.
18266 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
18267 of attribute cpu_facility to zarch for the last alternative.
18268 (*cmpmem_short): Ditto.
18270 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
18272 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
18273 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
18274 (arm_expand_prologue): Likewise.
18276 2013-06-06 Teresa Johnson <tejohnson@google.com>
18279 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
18280 as this is now done by redirect_edge_and_branch_force.
18281 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
18282 barriers, and fix interaction with splitting.
18283 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
18284 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
18285 reflect changes made in the routine.
18286 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
18287 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
18288 since this is called in cfglayout mode, and replace partition fixup
18289 with assert as that is now done by force_nonfallthru_and_redirect.
18290 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
18291 already be marked with region crossing note.
18292 (insert_section_boundary_note): Make non-static, gate on flag
18293 has_bb_partition, rewrite to also check for multiple partitions.
18294 (rest_of_handle_reorder_blocks): Remove call to
18295 insert_section_boundary_note, now done later during free_cfg.
18296 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
18297 * bb-reorder.h (insert_section_boundary_note): Declare.
18298 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
18299 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
18300 invoke insert_section_boundary_note.
18301 (try_redirect_by_replacing_jump): Remove unnecessary
18302 check for region crossing note.
18303 (fixup_partition_crossing): New function.
18304 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
18305 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
18306 in non-cfglayout mode.
18307 (force_nonfallthru_and_redirect): Fixup partition boundaries,
18308 remove old code that tried to do this. Emit barrier correctly
18309 when we are in cfglayout mode.
18310 (last_bb_in_partition): New function.
18311 (rtl_split_edge): Correctly fixup partition boundaries.
18312 (commit_one_edge_insertion): Remove old code that tried to
18313 fixup region crossing edge since this is now handled in
18314 split_block, and set up insertion point correctly since
18315 block may now end in a jump.
18316 (verify_hot_cold_block_grouping): Guard against checking when not in
18317 linearized RTL mode.
18318 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
18320 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
18321 rtl_verify_flow_info, so not called in cfglayout mode.
18322 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
18323 (fixup_reorder_chain): Remove old code that attempted to fixup region
18324 crossing note as this is now handled in force_nonfallthru_and_redirect.
18325 (duplicate_insn_chain): Don't duplicate switch section notes.
18326 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
18328 * basic-block.h (emit_barrier_after_bb): Declare.
18330 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18332 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
18333 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
18334 arm_usatsihi): Adjust alternatives for arm_restrict_it.
18336 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18338 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
18340 * config/arm/ldmstm.md: Regenerate.
18342 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18344 * config/arm/sync.md (atomic_loaddi_1):
18345 Disable predication for arm_restrict_it.
18346 (arm_load_exclusive<mode>): Likewise.
18347 (arm_load_exclusivesi): Likewise.
18348 (arm_load_exclusivedi): Likewise.
18349 (arm_load_acquire_exclusive<mode>): Likewise.
18350 (arm_load_acquire_exclusivesi): Likewise.
18351 (arm_load_acquire_exclusivedi): Likewise.
18352 (arm_store_exclusive<mode>): Likewise.
18353 (arm_store_exclusive<mode>): Likewise.
18354 (arm_store_release_exclusivedi): Likewise.
18355 (arm_store_release_exclusive<mode>): Likewise.
18357 2013-06-06 Richard Biener <rguenther@suse.de>
18359 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
18361 (lto_tag_is_tree_code_p): Adjust.
18362 (lto_tag_is_gimple_code_p): Likewise.
18363 (lto_gimple_code_to_tag): Likewise.
18364 (lto_tag_to_gimple_code): Likewise.
18365 (lto_tree_code_to_tag): Likewise.
18366 (lto_tag_to_tree_code): Likewise.
18367 * data-streamer.h (streamer_write_hwi_in_range): Use
18368 uhwi streaming to stream the normalized range.
18369 (streamer_read_hwi_in_range): Likewise.
18371 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18373 * config/arm/arm.md (enabled_for_depr_it): New attribute.
18374 (predicable_short_it): Likewise.
18375 (predicated): Likewise.
18376 (enabled): Handle above.
18377 (define_cond_exec): Set predicated attribute to yes.
18379 2013-06-05 Mike Stump <mikestump@comcast.net>
18381 * gdbinit.in (__FUNCTION__): Add.
18383 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
18385 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
18386 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
18388 2013-06-05 Jan Hubicka <jh@suse.cz>
18390 * varasm.c (mark_decl_referenced): Revert the removal until targets
18393 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
18395 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
18396 instead of mark_decl_referenced.
18398 2013-06-05 Jan Hubicka <jh@suse.cz>
18400 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
18401 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
18402 and symtab_used_from_object_file_p.
18403 (cgraph_make_node_local_1): Clear forced_by_abi.
18404 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
18405 * cgraph.h (symtab_node_base): Add forced_by_abi.
18406 (decide_is_variable_needed): Remove.
18407 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
18408 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
18409 (decide_is_symbol_needed): ... this one; handle symbols in general;
18410 always analyze virtuals; honnor forced_by_abi.
18411 (cgraph_finalize_function): Update.
18412 (varpool_finalize_decl): Update.
18413 (symbol_defined_and_needed): Remove.
18414 (analyze_functions): Update.
18415 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18416 output_refs, input_overwrite_node): Handle forced_by_abi.
18417 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18418 (address_taken_from_non_vtable_p): ... this one.
18419 (comdat_can_be_unshared_p_1): New function.
18420 (cgraph_comdat_can_be_unshared_p): Rename to ...
18421 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18422 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18423 (function_and_variable_visibility): Clear forced_by_abi as needed.
18424 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18425 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18426 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18427 * varpool.c (decide_is_variable_needed): Remove.
18429 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18431 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18432 (arm_option_override): Override arm_restrict_it where appropriate.
18433 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18434 * config/arm/arm.opt (mrestrict-it): New command-line option.
18435 * doc/invoke.texi: Document -mrestrict-it.
18437 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18439 * tsan.c (tsan_atomic_table): Make const.
18441 2013-06-05 Richard Biener <rguenther@suse.de>
18443 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18444 index associated with the tree we are supposed to replace.
18445 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18446 TREE_ASM_WRITTEN as zero for everything but SSA names.
18448 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18450 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18452 2013-06-04 Jan Hubicka <jh@suse.cz>
18454 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18455 (input_node, input_varpool_node): Handle correctly external same
18457 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18458 nodes at ltrans stage.
18460 2013-06-04 Jan Hubicka <jh@suse.cz>
18462 * ipa-inline.c (update_caller_keys): Fix availability test.
18463 (update_callee_keys): Likewise.
18464 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18465 to follow ELF standard.
18467 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18469 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18470 (mips64r5900el-*-elf*): New configurations.
18471 * config/mips/mips-cpus.def (r5900): New processor.
18472 * config/mips/mips-tables.opt: Regenerate.
18473 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18474 (mips_issue_rate): Handle PROCESSOR_R5900.
18475 (mips_reorg_process_insns): Force reorder mode for the R5900.
18476 * config/mips/mips.h (TARGET_MIPS5900): Define.
18477 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18479 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18481 * config/mips/mips.md (processor): Add r5900.
18482 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18484 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18486 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18487 into function to generate MOVI instruction.
18488 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18489 (aarch64_preferred_simd_mode): Turn into wrapper.
18490 (aarch64_output_scalar_simd_mov_immediate): New function.
18491 * config/aarch64/aarch64-protos.h: Add prototype for above.
18493 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18495 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18496 element_char member.
18497 (sizetochar): Return signed char.
18498 (aarch64_simd_valid_immediate): Remove elchar and other
18499 unnecessary variables.
18500 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18501 Calculate element_char as required.
18502 * config/aarch64/aarch64-protos.h: Update and move prototype
18503 for aarch64_output_simd_mov_immediate.
18504 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18507 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18509 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18510 information completed by aarch64_simd_valid_immediate.
18511 (aarch64_legitimate_constant_p): Update arguments.
18512 (aarch64_simd_valid_immediate): Work with struct rather than many
18514 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18515 (aarch64_simd_make_constant): Update arguments.
18516 (aarch64_output_simd_mov_immediate): Work with struct rather than
18517 many pointers. Output immediate directly rather than as operand.
18518 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18520 * config/aarch64/constraints.md (Dn): Update arguments.
18522 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18524 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18526 (aarch64_simd_immediate_valid_for_move): Remove.
18527 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18528 (aarch64_simd_make_constant): Update call.
18529 (aarch64_output_simd_mov_immediate): Update call.
18530 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18532 * config/aarch64/constraints.md (Dn): Update call.
18534 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18536 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18537 return type to bool for prototype.
18538 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18539 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18540 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18542 2013-06-04 Catherine Moore <clm@codesourcery.com>
18544 * config/mips/mips.opt (meva): New.
18545 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18546 (ASM_SPEC): Handle -meva.
18547 * doc/invoke.texi (meva): Document.
18549 2013-06-04 Alan Modra <amodra@gmail.com>
18551 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18554 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18556 * rtl.def: Add extra fourth optional field to define_cond_exec.
18557 * gensupport.c (process_one_cond_exec): Process attributes from
18559 * doc/md.texi: Document fourth field in define_cond_exec.
18561 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
18563 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
18564 out the processing order as in store_bit_field_1.
18566 2013-06-04 Jan Hubicka <jh@suse.cz>
18568 PR middle-end/57500
18569 * cgraphunit.c (cgraph_process_same_body_aliases): Create
18570 non-VAR_DECL node if it does not exist yet.
18572 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
18574 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
18575 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
18576 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
18577 target_cpu_default setting.
18579 2013-06-03 Teresa Johnson <tejohnson@google.com>
18581 * dumpfile.c (opt_info_switch_p): Change -fopt-info
18582 default to -fopt-info=optimized instead of all.
18583 * doc/invoke.texi: Ditto.
18584 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
18585 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
18586 (execute_vect_slp): Emit BB vectorization success under
18587 MSG_OPTIMIZED_LOCATIONS.
18588 * tree-vect-slp.c (vect_slp_transform_bb): Change
18589 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
18590 * tree-vect-loop.c (vect_transform_loop): Ditto.
18592 2013-06-03 Jason Merrill <jason@redhat.com>
18595 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
18596 Use TARGET_EXPR for C++.
18598 2013-06-03 Jakub Jelinek <jakub@redhat.com>
18600 PR rtl-optimization/57268
18601 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
18602 if DEBUG_INSN_P (insn).
18605 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18607 PR rtl-optimization/57268
18608 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18609 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18611 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18613 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18614 (ix86_avoid_lea_for_addr): Likewise.
18615 (exact_dependency_1): Likewise.
18616 (ix86_adjust_cost): Likewise.
18617 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18619 (do_reorder_for_imul): Fix typo, formatting and
18620 !reload_completed check removed.
18621 (ix86_sched_reorder): Fix typo and formatting.
18622 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18625 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18627 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18629 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18631 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18632 <STRING_CST>: Likewise.
18633 <VECTOR_CST>: Likewise.
18635 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18636 Mikael Morin <mikael@gcc.gnu.org>
18638 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18639 * config.in: Regenerated.
18640 * configure: Regenerated.
18642 2013-06-01 Jan Hubicka <jh@suse.cz>
18644 PR middle-end/57366
18645 * cgraphunit.c (compile): When weakref is not supported,
18646 set up transparent aliases before final output pass.
18647 * varasm.c (assemble_alias): Do not try to do it here.
18649 2013-06-01 Jan Hubicka <jh@suse.cz>
18651 PR middle-end/57467
18652 * passes.c (for_per_function): Skip unanalyzed functions.
18654 2013-06-01 Jan Hubicka <jh@suse.cz>
18656 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18657 (lto_symtab_merge_symbols_1): ... this one.
18658 (lto_symtab_merge_cgraph_nodes): Rename to ...
18659 (lto_symtab_merge_symbols): ... this one; simplify.
18660 * cgraph.c (same_body_aliases_done): Rename to ...
18661 (cpp_implicit_aliases_done): ... this one.
18662 (cgraph_create_function_alias): Update.
18663 (cgraph_same_body_alias): Update.
18664 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18665 (verify_edge_corresponds_to_fndecl): Simplify.
18666 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18667 (cgraph_node): Remove same_body_alias.
18668 (varpool_node): Remove alias_of and extra_name_alias.
18669 (same_body_aliases_done): Rename to ..
18670 (cpp_implicit_aliases_done): ... this one.
18671 (symtab_alias_ultimate_target): Add default parameter.
18672 (symtab_resolve_alias): New function.
18673 (fixup_same_cpp_alias_visibility): Declare.
18674 (cgraph_function_node): Add default parameter.
18675 (cgraph_node_asm_name): Likewise.
18676 (cgraph_function_or_thunk_node): Add default parameter; do
18677 not ICE when it is NULL.
18678 (varpool_variable_node): Likewise.
18679 * tree-emutls.c (create_emultls_var): Update.
18680 (ipa_lower_emutls): Update.
18681 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18682 (cgraph_reset_node): Reset alias info.
18683 (cgraph_finalize_function): Update.
18684 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18685 (analyze_function): Simplify.
18686 (cgraph_process_same_body_aliases): Simplify.
18687 (analyze_functions): Fixup same body aliases.
18688 (handle_alias_pairs): Simplify.
18689 (assemble_thunk): Update.
18690 (assemble_thunks_and_aliases): Update.
18691 (output_weakrefs): Rewrite.
18692 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18693 (lto_output_varpool_node): Likewise.
18694 (compute_ltrans_boundary): Remve assert.
18695 (get_alias_symbol): New functoin.
18696 (input_node): Rewrite alias handling.
18697 (input_varpool_node): Likewise.
18698 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18699 * ipa.c (process_references): Handle weakrefs correctly.
18700 (symtab_remove_unreachable_nodes): Likewise.
18701 * trans-mem.c (get_cg_data): Update.
18702 (ipa_tm_create_version_alias): Update.
18703 (ipa_tm_execute): Update.
18704 * symtab.c (dump_symtab_base): Dump aliases.
18705 (verify_symtab_base): Verify aliases.
18706 (symtab_node_availability): New function.
18707 (symtab_alias_ultimate_target): Simplify.
18708 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18709 handle all the fixup cases.
18710 (symtab_resolve_alias): New function.
18711 * passes.c (ipa_write_summaries): Handle weakrefs.
18712 * varpool.c (varpool_analyze_node): Simplify.
18713 (assemble_aliases): Update.
18714 (varpool_create_variable_alias): Simplify.
18715 (varpool_extra_name_alias): Simplify.
18716 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18717 (lto_symtab_merge_symbols): ... this one.
18719 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18722 PR rtl-optimization/57268
18723 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18724 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18726 2013-06-01 Tobias Burnus <burnus@net-b.de>
18728 Partially reverted:
18729 2013-05-31 Tobias Burnus <burnus@net-b.de>
18731 PR middle-end/57073
18732 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18735 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18737 PR rtl-optimization/57268
18738 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18739 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18741 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18743 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18744 unordered comparison operators when -fno-trapping-math is in effect
18746 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18747 and implement unordered comparison operators properly on the e500.
18749 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18751 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18752 for constant scalar integers.
18753 (simplify_relational_operation_1): Likewise.
18755 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18757 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18758 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18761 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18762 Igor Zamyatin <igor.zamyatin@intel.com>
18764 Silvermont (SLM) architecture performance tuning.
18765 * config/i386/i386.h (enum ix86_tune_indices): Add
18766 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18767 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18769 * config/i386/i386.c (initial_ix86_tune_features)
18770 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18771 (ix86_lea_outperforms): Handle Silvermont tuning.
18772 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18774 (ix86_use_lea_for_mov): Likewise.
18775 (ix86_avoid_lea_for_addr): Likewise.
18776 (ix86_lea_for_add_ok): Likewise.
18777 (exact_dependency_1): New function.
18778 (exact_store_load_dependency): Likewise.
18779 (ix86_adjust_cost): Handle Silvermont tuning.
18780 (do_reoder_for_imul): Likewise.
18781 (swap_top_of_ready_list): New function.
18782 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18784 * config/i386/i386.md (peepholes that split memory operand in fp
18787 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18789 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18790 Remove un-necessary braces.
18792 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18794 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18795 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18797 2013-05-31 Tobias Burnus <burnus@net-b.de>
18799 PR middle-end/57073
18800 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18802 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18805 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18806 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18807 * config/arm/neon.md (iordi3_neon): Remove.
18808 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18809 * config/arm/predicates.md (imm_for_neon_logic_operand):
18810 Move to earlier in the file.
18811 (neon_logic_op2): Likewise.
18812 (arm_iordi_operand_neon): New predicate.
18814 2013-05-31 Richard Biener <rguenther@suse.de>
18816 PR tree-optimization/57478
18817 PR tree-optimization/57453
18818 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18821 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18823 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18824 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18826 2013-05-30 Tobias Burnus <burnus@net-b.de>
18827 Thomas Koenig <tkoenig@gcc.gnu.org>
18829 PR middle-end/57073
18830 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18831 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18833 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18835 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18837 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18839 * target.def (register_usage_leveling_p): New hook.
18840 * targhooks.c (default_register_usage_leveling_p): New.
18841 * targhooks.h (default_register_usage_leveling_p): New prototype.
18842 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18843 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18844 * doc/tm.texi: Update.
18845 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18847 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18849 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18850 (*insv_reg<mode>): New define_insn.
18852 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18854 PR rtl-optimization/57439
18855 * postreload.c (move2add_valid_value_p): Check that we have
18856 a zero subreg_regno_offset when accessing the register in
18857 the requested mode.
18859 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18860 Igor Zamyatin <igor.zamyatin@intel.com>
18862 Silvermont (SLM) architecture pipeline model, tuning and
18864 * config.gcc: Add slm config options and target.
18866 * config/i386/slm.md: New.
18868 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18870 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18872 (ix86_target_macros_internal): Likewise.
18874 * config/i386/i386.c (slm_cost): New cost.
18875 (m_SLM): New macro flag.
18876 (initial_ix86_tune_features): Set m_SLM.
18877 (x86_accumulate_outgoing_args): Likewise.
18878 (x86_arch_always_fancy_math_387): Likewise.
18879 (processor_target_table): Add slm cost.
18880 (cpu_names): Add slm cpu name.
18881 (x86_option_override_internal): Set SLM ISA.
18882 (ix86_issue_rate): New case PROCESSOR_SLM.
18883 (ia32_multipass_dfa_lookahead): Likewise.
18884 (fold_builtin_cpu): Add slm.
18886 * config/i386/i386.h (TARGET_SLM): New target macro.
18887 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18888 (processor_type): Add PROCESSOR_SLM.
18890 * config/i386/i386.md (cpu): Add new value "slm".
18891 (slm.md): Include slm.md.
18893 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18894 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18896 * config/arm/arm-protos.h: Add and update function protos.
18897 * config/arm/arm.c (use_simple_return_p): New added.
18898 (thumb2_expand_return): Check simple_return flag.
18899 * config/arm/arm.md: Add simple_return and conditional simple_return.
18900 * config/arm/iterators.md: Add iterator for return and simple_return.
18902 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18904 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18905 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18906 (arm_emit_vfp_multi_reg_pop): Likewise.
18907 (thumb2_emit_ldrd_pop): Likewise.
18908 (arm_expand_epilogue): Add misc REG_CFA notes.
18909 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18911 2013-05-29 Lawrence Crowl <crowl@google.com>
18913 * config/arm/t-arm: Update for below.
18915 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18916 Change type to hash_table. Update dependent calls and types.
18918 * config/i386/t-cygming: Update for below.
18920 * config/i386/t-interix: Update for below.
18922 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18923 Change type to hash_table. Update dependent calls and types.
18924 (i386_find_on_wrapper_list::wrappers): Likewise.
18926 * config/ia64/t-ia64: Update for below.
18928 * config/ia64/ia64.c (bundle_state_table):
18929 Change type to hash_table. Update dependent calls and types.
18931 * config/mips/mips.c (mips_reorg_process_insns::htab):
18932 Change type to hash_table. Update dependent calls and types.
18934 * config/sol2.c (solaris_comdat_htab):
18935 Change type to hash_table. Update dependent calls and types.
18937 * config/t-sol2: Update for above.
18939 2013-05-29 Teresa Johnson <tejohnson@google.com>
18941 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18942 functions are not yet marked as defined.
18944 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18945 Pat Haugen <pthaugen@us.ibm.com>
18946 Peter Bergner <bergner@vnet.ibm.com>
18948 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18952 (vrotl<mode>3): Likewise.
18953 (vashl<mode>3): Likewise.
18954 (vlshr<mode>3): Likewise.
18955 (vashr<mode>3): Likewise.
18957 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18958 support for power8 V2DI builtins.
18960 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18961 power8 V2DI builtins.
18962 (vupkhsw): Likewise.
18963 (vupklsw): Likewise.
18964 (vaddudm): Likewise.
18965 (vminsd): Likewise.
18966 (vmaxsd): Likewise.
18967 (vminud): Likewise.
18968 (vmaxud): Likewise.
18969 (vpkudum): Likewise.
18970 (vpksdss): Likewise.
18971 (vpkudus): Likewise.
18972 (vpksdus): Likewise.
18977 (vsubudm): Likewise.
18978 (vcmpequd): Likewise.
18979 (vcmpgtsd): Likewise.
18980 (vcmpgtud): Likewise.
18981 (vcmpequd_p): Likewise.
18982 (vcmpgtsd_p): Likewise.
18983 (vcmpgtud_p): Likewise.
18984 (vupkhsw): Likewise.
18985 (vupklsw): Likewise.
18986 (vaddudm): Likewise.
18987 (vmaxsd): Likewise.
18988 (vmaxud): Likewise.
18989 (vminsd): Likewise.
18990 (vminud): Likewise.
18991 (vpksdss): Likewise.
18992 (vpksdus): Likewise.
18993 (vpkudum): Likewise.
18994 (vpkudus): Likewise.
18999 (vsubudm): Likewise.
19001 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
19002 support for power8 V2DI instructions.
19004 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
19005 power8 V2DI instructions. Combine pack and unpack insns to use an
19006 iterator for each mode. Check whether a particular mode supports
19007 Altivec instructions instead of just checking TARGET_ALTIVEC.
19008 (UNSPEC_VPKUWUM): Likewise.
19009 (UNSPEC_VPKSHSS): Likewise.
19010 (UNSPEC_VPKSWSS): Likewise.
19011 (UNSPEC_VPKUHUS): Likewise.
19012 (UNSPEC_VPKSHUS): Likewise.
19013 (UNSPEC_VPKUWUS): Likewise.
19014 (UNSPEC_VPKSWUS): Likewise.
19015 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
19016 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
19017 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
19018 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
19019 (UNSPEC_VUPKHSB): Likewise.
19020 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
19021 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
19022 (UNSPEC_VUPKHSH): Likewise.
19023 (UNSPEC_VUPKLSB): Likewise.
19024 (UNSPEC_VUPKLSH): Likewise.
19026 (VI_char): Likewise.
19027 (VI_scalar): Likewise.
19028 (VI_unit): Likewise.
19030 (VP_small): Likewise.
19031 (VP_small_lc): Likewise.
19032 (VU_char): Likewise.
19033 (add<mode>3): Likewise.
19034 (altivec_vaddcuw): Likewise.
19035 (altivec_vaddu<VI_char>s): Likewise.
19036 (altivec_vadds<VI_char>s): Likewise.
19037 (sub<mode>3): Likewise.
19038 (altivec_vsubcuw): Likewise.
19039 (altivec_vsubu<VI_char>s): Likewise.
19040 (altivec_vsubs<VI_char>s): Likewise.
19041 (altivec_vavgs<VI_char>): Likewise.
19042 (altivec_vcmpbfp): Likewise.
19043 (altivec_eq<mode>): Likewise.
19044 (altivec_gt<mode>): Likewise.
19045 (altivec_gtu<mode>): Likewise.
19046 (umax<mode>3): Likewise.
19047 (smax<mode>3): Likewise.
19048 (umin<mode>3): Likewise.
19049 (smin<mode>3): Likewise.
19050 (altivec_vpkuhum): Likewise.
19051 (altivec_vpkuwum): Likewise.
19052 (altivec_vpkshss): Likewise.
19053 (altivec_vpkswss): Likewise.
19054 (altivec_vpkuhus): Likewise.
19055 (altivec_vpkshus): Likewise.
19056 (altivec_vpkuwus): Likewise.
19057 (altivec_vpkswus): Likewise.
19058 (altivec_vpks<VI_char>ss): Likewise.
19059 (altivec_vpks<VI_char>us): Likewise.
19060 (altivec_vpku<VI_char>us): Likewise.
19061 (altivec_vpku<VI_char>um): Likewise.
19062 (altivec_vrl<VI_char>): Likewise.
19063 (altivec_vsl<VI_char>): Likewise.
19064 (altivec_vsr<VI_char>): Likewise.
19065 (altivec_vsra<VI_char>): Likewise.
19066 (altivec_vsldoi_<mode>): Likewise.
19067 (altivec_vupkhsb): Likewise.
19068 (altivec_vupkhs<VU_char>): Likewise.
19069 (altivec_vupkls<VU_char>): Likewise.
19070 (altivec_vupkhsh): Likewise.
19071 (altivec_vupklsb): Likewise.
19072 (altivec_vupklsh): Likewise.
19073 (altivec_vcmpequ<VI_char>_p): Likewise.
19074 (altivec_vcmpgts<VI_char>_p): Likewise.
19075 (altivec_vcmpgtu<VI_char>_p): Likewise.
19076 (abs<mode>2): Likewise.
19077 (vec_unpacks_hi_v16qi): Likewise.
19078 (vec_unpacks_hi_v8hi): Likewise.
19079 (vec_unpacks_lo_v16qi): Likewise.
19080 (vec_unpacks_hi_<VP_small_lc>): Likewise.
19081 (vec_unpacks_lo_v8hi): Likewise.
19082 (vec_unpacks_lo_<VP_small_lc>): Likewise.
19083 (vec_pack_trunc_v8h): Likewise.
19084 (vec_pack_trunc_v4si): Likewise.
19085 (vec_pack_trunc_<mode>): Likewise.
19087 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
19089 (vec_vmaxsd): Likewise.
19090 (vec_vmaxud): Likewise.
19091 (vec_vminsd): Likewise.
19092 (vec_vminud): Likewise.
19093 (vec_vpksdss): Likewise.
19094 (vec_vpksdus): Likewise.
19095 (vec_vpkudum): Likewise.
19096 (vec_vpkudus): Likewise.
19097 (vec_vrld): Likewise.
19098 (vec_vsld): Likewise.
19099 (vec_vsrad): Likewise.
19100 (vec_vsrd): Likewise.
19101 (vec_vsubudm): Likewise.
19102 (vec_vupkhsw): Likewise.
19103 (vec_vupklsw): Likewise.
19105 2013-05-29 Jan Hubicka <jh@suse.cz>
19107 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
19108 flags; reorder rest of fields in more consistent way.
19109 (varpool_node): Remove analyzed, finalized and alias.
19110 (cgraph_ndoe): Likewise.
19111 (symtab_alias_ultimate_target): New function.
19112 (cgraph_function_node): Move offline.
19113 (cgraph_reset_node): Declare.
19114 (cgraph_comdat_can_be_unshared_p): Remove.
19115 (varpool_remove_initializer): Declare.
19116 (varpool_first_defined_variable, varpool_next_defined_variable
19117 cgraph_first_defined_function, cgraph_next_defined_function): Update.
19118 (cgraph_function_with_gimple_body_p): Update.
19119 (varpool_all_refs_explicit_p): Update.
19120 (symtab_alias_target): New function.
19121 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
19122 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
19123 (cgraph_function_or_thunk_node): Simplify using
19124 symtab_alias_ultimate_target.
19125 (varpool_variable_node): Likewise.
19126 * cgraph.c (cgraph_create_function_alias): Update.
19127 (cgraph_add_thunk): Update.
19128 (cgraph_remove_node): Update.
19129 (dump_cgraph_node): Do not dump removed flags.
19130 (cgraph_function_body_availability): Update.
19131 (cgraph_propagate_frequency): Update.
19132 (verify_cgraph_node): Check sanity of local flag.
19133 (cgraph_function_node): Move here from cgraph.h; revamp for
19134 cgraph_function_or_thunk_node.
19135 * lto-symtab.c (lto_varpool_replace_node): Update.
19136 (lto_symtab_resolve_can_prevail_p): Update.
19137 (lto_symtab_merge_cgraph_nodes): Update.
19138 * ipa-cp.c (determine_versionability, initialize_node_lattices,
19139 propagate_constants_accross_call, devirtualization_time_bonus,
19140 ipcp_propagate_stage): Update.
19141 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
19142 * ipa-inline-transform.c (clone_inlined_nodes,
19143 preserve_function_body_p): Update.
19144 * ipa-reference.c (propagate): Update.
19145 (write_node_summary_p): Update.
19146 * toplev.c (wrapup_global_declaration_2): Update.
19147 * cgraphunit.c (cgraph_analyze_function): Rename to ...
19148 (analyze_function) ... this one.
19149 (cgraph_process_new_functions): Update.
19150 (cgraph_reset_node): Export.
19151 (cgraph_finalize_function): Update.
19152 (cgraph_add_new_function): Update.
19153 (process_function_and_variable_attributes): Update.
19154 (varpool_finalize_decl): Update.
19155 (symbol_finalized): Remove.
19156 (symbol_finalized_and_needed): Rename to ...
19157 (symbol_defined_and_needed): ... update.
19158 (cgraph_analyze_functions): Update.
19159 (handle_alias_pairs): Update.
19160 (mark_functions_to_output): Update.
19161 (assemble_thunk): Update.
19162 (output_in_order): Update.
19163 (output_weakrefs): Update.
19164 (finalize_compilation_unit): Update.
19165 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
19166 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
19167 input_node, input_varpool_node): Update.
19168 * dbxout.c (dbxout_expand_expr): Update.
19169 * cgraphclones.c (cgraph_clone_node): Update.
19170 (cgraph_copy_node_for_versioning): Update.
19171 (cgraph_materialize_clone): Update.
19172 (cgraph_materialize_all_clones): Update.
19173 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
19174 propagate_pure_const, propagate_nothrow): Update.
19175 * lto-streamer-out.c (lto_output, write_symbol): Update.
19176 * ipa-utils.c (ipa_reverse_postorder): Update.
19177 * ipa-inline.c (can_inline_edge_p): Update.
19178 (update_caller_keys, ipa_inline): Update.
19179 * dwarf2out.c (reference_to_unused,
19180 premark_types_used_by_global_vars_helper): Update.
19181 * tree-eh.c (tree_could_trap_p): Update.
19182 * ipa-split.c (consider_split, execute_split_functions): Update.
19183 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
19184 has_addr_references_p): Update; move ahead in file for better
19186 (process_references): Simplify.
19187 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
19188 bodies are removed.
19189 (cgraph_comdat_can_be_unshared_p): Make static.
19190 (cgraph_externally_visible_p): Update.
19191 (varpool_externally_visible_p): Update.
19192 (function_and_variable_visibility): Update.
19193 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
19194 ipa_tm_mark_force_output_node): Update.
19195 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
19196 estimate_edge_devirt_benefit, inline_generate_summary,
19197 inline_write_summary): Update.
19198 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
19199 * ipa-prop.c (ipa_compute_jump_functions): Update.
19200 (ipa_print_node_params, ipa_prop_read_section,
19201 ipa_update_after_lto_read, read_replacements_section): Update.
19202 * varasm.c (mark_decl_referenced): Update.
19203 (assemble_alias, dump_tm_clone_pairs): Update.
19204 * tree-inline.c (copy_bb): Update.
19205 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
19207 * symtab.c (dump_symtab_base): Print new flags.
19208 (verify_symtab_base): Verify new flags.
19209 (symtab_alias_ultimate_target): New function.
19210 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
19211 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
19213 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
19215 * i386.c (ix86_get_function_versions_dispatcher,
19216 ix86_generate_version_dispatcher_body): Update.
19217 (fold_builtin_cpu): Use varpool_add_new_variable.
19218 * varpool.c (varpool_remove_initializer): Break out from ...
19219 (varpool_remove_node): ... this one.
19220 (dump_varpool_node, varpool_node_for_asm,
19221 cgraph_variable_initializer_availability, varpool_analyze_node,
19222 varpool_assemble_decl, varpool_remove_unreferenced_decls,
19223 varpool_finalize_named_section_flags, varpool_create_variable_alias):
19226 2013-05-29 Jan Hubicka <jh@suse.cz>
19228 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
19230 2013-05-29 Easwaran Raman <eraman@google.com>
19232 PR tree-optimization/57442
19233 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
19234 when control exits the main loop.
19236 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
19238 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
19240 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
19241 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
19242 * rx/t-rx: Add rx100 under multi library matches option for nofpu
19245 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19247 PR tree-optimization/57441
19248 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
19249 Don't limit size of incr_vec to number of candidates.
19251 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
19253 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
19254 and mips16 directories.
19255 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
19256 (MULTILIB_DIRNAMES): Ditto.
19257 (MULTILIB_EXCEPTIONS): Add new exceptions.
19258 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
19259 (MULTILIB_DIRNAMES): Ditto.
19260 (MULTILIB_EXCEPTIONS): Add new exceptions.
19262 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19263 Marcus Shawcroft <marcus.shawcroft@arm.com>
19265 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
19266 SYMBOL_TINY_ABSOLUTE.
19267 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
19268 SYMBOL_TINY_ABSOLUTE.
19269 (aarch64_expand_mov_immediate): Likewise.
19270 (aarch64_classify_symbol): Likewise.
19271 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
19272 Permit SYMBOL_TINY_ABSOLUTE.
19273 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
19275 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19276 Marcus Shawcroft <marcus.shawcroft@arm.com>
19278 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
19279 Refactor if/switch. Replace gcc_assert with if.
19281 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19283 * config/i386/i386.c (initial_ix86_tune_features): Enable
19284 FP Reassociation for AMD bdver1 and bdver2.
19286 2013-05-29 Martin Jambor <mjambor@suse.cz>
19288 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
19289 and IMAGPART_EXPR do not occur within other handled_components.
19291 2013-05-29 Richard Biener <rguenther@suse.de>
19293 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
19294 access on whether the use is in the BB we currently try to
19296 (vect_bb_vectorization_profitable_p): Pass the BB we currently
19297 vectorize to vect_bb_slp_scalar_cost.
19299 2013-05-29 Richard Biener <rguenther@suse.de>
19301 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
19302 computing scalar cost offsetted by stmts that are kept live
19304 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
19305 for computation of scalar cost.
19307 2013-05-28 Steve Ellcey <sellcey@mips.com>
19309 * config/mips/mips-cpus.def (mips32r2): Change processor type.
19311 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
19313 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
19314 array notation built-in reduction functions.
19315 * doc/passes.texi (Passes): Added documentation about changes done
19317 * doc/invoke.texi (C Dialect Options): Added documentation about
19318 the -fcilkplus flag.
19319 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
19320 (BUILTINS_DEF): Depend on cilkplus.def.
19321 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
19322 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
19323 * cilkplus.def: New file.
19325 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
19327 PR rtl-optimization/57439
19328 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
19330 2013-05-28 Easwaran Raman <eraman@google.com>
19332 PR tree-optimization/57337
19333 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
19334 (find_insert_point): Correctly identify the insertion point
19335 when two statements with the same UID is compared.
19337 2013-05-28 Richard Biener <rguenther@suse.de>
19339 PR tree-optimization/56787
19340 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
19341 from the list of data references.
19342 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
19344 (vect_analyze_loop_operations): Likewise.
19345 (vect_transform_loop): Remove clobbers.
19347 2013-05-28 Martin Jambor <mjambor@suse.cz>
19349 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
19350 and REALPART_EXPRs have scalar type.
19352 2013-05-28 Richard Biener <rguenther@suse.de>
19354 PR tree-optimization/57411
19355 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
19357 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
19358 virtual operand propagation.
19360 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19362 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
19363 destination register for bmasksi_vis.
19364 (vector_init_bshuffle): Likewise.
19365 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
19367 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19369 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
19370 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
19371 mode if the instruction isn't available in the original mode.
19372 * config/sparc/sparc.opt (mfix-ut699): New option.
19373 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
19374 (divdf3): Turn into expander.
19375 (divdf3_nofix): New insn.
19376 (divdf3_fix): Likewise.
19377 (divsf3): Disable if -mfix-ut699.
19378 (sqrtdf2): Turn into expander.
19379 (sqrtdf2_nofix): New insn.
19380 (sqrtdf2_fix): Likewise.
19381 (sqrtsf2): Disable if -mfix-ut699.
19383 2013-05-27 Richard Biener <rguenther@suse.de>
19385 PR middle-end/57412
19386 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
19387 block for the new loop.
19389 2013-05-27 Richard Biener <rguenther@suse.de>
19391 PR tree-optimization/57343
19392 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
19393 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
19394 (number_of_iterations_cond): Do not build the folded tree.
19396 2013-05-27 Richard Biener <rguenther@suse.de>
19399 PR middle-end/57381
19400 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19401 OEP_CONSTANT_ADDRESS_OF retained.
19403 PR tree-optimization/57417
19404 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
19405 for unchanged base.
19406 (set_ssa_val_to): Compare addresses using
19407 get_addr_base_and_unit_offset.
19409 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19411 PR rtl-optimization/56833
19412 * postreload.c (move2add_record_mode): New function.
19413 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19414 (move2add_use_add2_insn): Use move2add_record_sym_value.
19415 (move2add_use_add3_insn): Likewise.
19416 (reload_cse_move2add): Use move2add_valid_value_p and
19417 move2add_record_mode. Invalidate call-clobbered and REG_INC
19418 affected regs by setting reg_mode to VOIDmode.
19419 (move2add_note_store): Don't pretend the inside of a SUBREG is
19420 the actual destination. Invalidate single/leading registers by
19421 setting reg_mode to VOIDmode.
19422 Use move2add_record_sym_value, move2add_valid_value_p and
19423 move2add_record_mode.
19425 2013-05-27 Richard Biener <rguenther@suse.de>
19427 PR tree-optimization/57396
19428 * tree-affine.c (double_int_constant_multiple_p): Properly
19429 return false for val == 0 and div != 0.
19431 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19433 * config/mips/mips.h: Use #elif in preprocessor conditions.
19435 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19438 * config/mips/constraints.md (kl): New constraint.
19439 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19440 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19441 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19442 from LO for MIPS16.
19443 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19445 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19448 * config/mips/mips.c (mips_can_inline_p): New function.
19449 (TARGET_CAN_INLINE_P): Define.
19451 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19453 * sched-int.h (ds_t, dw_t): Make unsigned int.
19454 Fix documentation that describes how all the ds_t bits are used.
19455 Reserve the last bit for delayed-branch scheduling.
19456 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19457 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19458 (gen_dep_weak_1): Remove prototype.
19459 * sched-deps.c (get_dep_weak_1): Make static.
19460 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19461 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19462 * doc/tm.texi: Regenerate.
19463 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19465 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19468 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19470 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19471 Sandra Loosemore <sandra@codesourcery.com>
19473 * config.gcc (powerpc-*): Allow native for with-cpu.
19475 2013-05-24 Jeff Law <law@redhat.com>
19477 PR tree-optimization/57124
19478 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19479 conversion feeding a condition if the range has an overflow
19480 if -fstrict-overflow. Add warnings for when we do make the
19483 2013-05-24 Dehao Chen <dehao@google.com>
19485 * tree-cfg.c (locus_discrim_map): Fix the typo.
19486 (locus_discrim_hasher): Likewise.
19487 (locus_discrim_hasher::hash): Likewise.
19488 (locus_discrim_hasher::equal): Likewise.
19490 2013-05-24 Martin Jambor <mjambor@suse.cz>
19492 PR tree-optimization/57294
19493 * cgraph.h (ipa_record_stmt_references): Declare.
19494 * cgraphbuild.c (ipa_record_stmt_references): New function.
19495 (build_cgraph_edges): Use ipa_record_stmt_references.
19496 (rebuild_cgraph_edges): Likewise.
19497 (cgraph_rebuild_references): Likewise.
19498 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19499 associated with the old statement and build references from the
19500 newly built statements.
19501 * ipa-ref.c (ipa_remove_stmt_references): New function.
19502 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19504 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19506 * lra-constraints.c (emit_spill_move): Use smaller mode for
19508 (check_and_process_move): Consider mem-reg moves for secondary
19510 (curr_insn_transform): Don't lose insns emitted before for
19511 secondary memory moves.
19512 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19513 reg set up in the current insn.
19515 2013-05-24 Dehao Chen <dehao@google.com>
19517 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19519 (locus_descrim_hasher::equal): Likewise.
19520 (build_gimple_cfg): New discriminator assignment algorithm.
19521 (make_edges): Likewise.
19522 (next_discriminator_for_locus): Likewise.
19523 (same_line_p): Likewise.
19524 (assign_discriminators): Likewise.
19525 (make_cond_expr_edges): Likewise.
19526 (make_gimple_switch_edges): Likewise.
19527 (make_goto_expr_edges): Likewise.
19528 (make_gimple_asm_edges): Likewise.
19530 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19532 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19533 X format specifier to only display bottom 16 bits.
19534 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19535 immediate to match for operand 2, since it will be masked.
19537 2013-05-24 Richard Biener <rguenther@suse.de>
19539 PR tree-optimization/57287
19540 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19541 all SSA names that occur in abnormal PHIs.
19543 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19546 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19547 that index is not negative.
19549 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19551 PR rtl-optimization/55177
19552 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19553 (simplify_byte_swapping_operation): New.
19554 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
19555 (simplify_relational_operation_1): Deal with BSWAP.
19557 2013-05-23 Richard Henderson <rth@redhat.com>
19560 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
19561 (ix86_reorg): Call it.
19563 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
19566 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
19567 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
19568 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
19570 2013-05-23 Christian Bruel <christian.bruel@st.com>
19573 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
19575 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19576 Marcus Shawcroft <marcus.shawcroft@arm.com>
19578 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
19579 * config/aarch64/constraints.md (Usa): Remove.
19580 * doc/md.texi (AArch64 Usa): Remove.
19582 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19583 Marcus Shawcroft <marcus.shawcroft@arm.com>
19585 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
19586 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
19587 * config/aarch64/predicates.md (aarch64_const_address): Remove.
19588 (aarch64_mov_operand): Use aarch64_mov_operand_p.
19590 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
19592 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
19593 instruction (AdvSIMD).
19594 * config/aarch64/aarch64-builtins.c
19595 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
19596 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
19598 2013-05-23 Martin Jambor <mjambor@suse.cz>
19600 PR middle-end/57347
19601 * tree.h (contains_bitfld_component_ref_p): Declare.
19602 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19603 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19605 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19606 not access a bit-field. Assert all final offsets are byte-aligned.
19608 2013-05-23 Richard Biener <rguenther@suse.de>
19610 PR tree-optimization/57380
19611 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19612 least one invariant or re-used load.
19613 * passes.c (init_optimization_passes): Move pass_phiprop before
19616 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19618 * config/aarch64/aarch64-simd.md
19619 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19621 2013-05-23 Richard Biener <rguenther@suse.de>
19623 PR middle-end/57381
19624 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19625 OEP_CONSTANT_ADDRESS_OF retained.
19627 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19629 PR middle-end/57344
19630 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19631 don't lower unit. Handle unit not being always BITS_PER_WORD.
19633 2013-05-23 Richard Biener <rguenther@suse.de>
19635 PR rtl-optimization/57341
19636 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19637 instead of true_dependence.
19639 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19641 * bb-reorder.c (branch_threshold): Make const.
19642 (exec_threshold): Ditto.
19644 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19645 Pat Haugen <pthaugen@us.ibm.com>
19646 Peter Bergner <bergner@vnet.ibm.com>
19648 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19649 documentation for the power8 crypto builtins.
19651 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19653 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19654 macros for defining power8 builtin functions.
19655 (BU_P8V_AV_2): Likewise.
19656 (BU_P8V_AV_P): Likewise.
19657 (BU_P8V_VSX_1): Likewise.
19658 (BU_P8V_OVERLOAD_1): Likewise.
19659 (BU_P8V_OVERLOAD_2): Likewise.
19660 (BU_CRYPTO_1): Likewise.
19661 (BU_CRYPTO_2): Likewise.
19662 (BU_CRYPTO_3): Likewise.
19663 (BU_CRYPTO_OVERLOAD_1): Likewise.
19664 (BU_CRYPTO_OVERLOAD_2): Likewise.
19665 (XSCVSPDP): Fix typo, point to the correct instruction.
19666 (VCIPHER): Add power8 crypto builtins.
19667 (VCIPHERLAST): Likewise.
19668 (VNCIPHER): Likewise.
19669 (VNCIPHERLAST): Likewise.
19670 (VPMSUMB): Likewise.
19671 (VPMSUMH): Likewise.
19672 (VPMSUMW): Likewise.
19673 (VPERMXOR_V2DI): Likewise.
19674 (VPERMXOR_V4SI: Likewise.
19675 (VPERMXOR_V8HI: Likewise.
19676 (VPERMXOR_V16QI: Likewise.
19677 (VSHASIGMAW): Likewise.
19678 (VSHASIGMAD): Likewise.
19679 (VPMSUM): Likewise.
19680 (VPERMXOR): Likewise.
19681 (VSHASIGMA): Likewise.
19683 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19684 __CRYPTO__ if the crypto instructions are available.
19685 (altivec_overloaded_builtins): Add support for overloaded power8
19688 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19689 support for power8 crypto builtins.
19690 (builtin_function_type): Likewise.
19691 (altivec_init_builtins): Add support for builtins that take vector
19692 long long (V2DI) arguments.
19694 * config/rs6000/crypto.md: New file, define power8 crypto
19697 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19698 Pat Haugen <pthaugen@us.ibm.com>
19699 Peter Bergner <bergner@vnet.ibm.com>
19701 * doc/invoke.texi (Option Summary): Add power8 options.
19702 (RS/6000 and PowerPC Options): Likewise.
19704 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19705 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19706 wm, wn, wr documentation.
19708 * config/rs6000/constraints.md (wm): New constraint for VSX
19709 registers if direct move instructions are enabled.
19710 (wn): New constraint for no registers.
19711 (wq): New constraint for quad word even GPR registers.
19712 (wr): New constraint if 64-bit instructions are enabled.
19713 (wv): New constraint if power8 vector instructions are enabled.
19714 (wQ): New constraint for quad word memory locations.
19716 * config/rs6000/predicates.md (const_0_to_15_operand): New
19717 constraint for 0..15 for crypto instructions.
19718 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19719 as GPR and floating point registers.
19720 (int_reg_operand): New predicate to match only GPR registers.
19721 (base_reg_operand): New predicate to match base registers.
19722 (quad_int_reg_operand): New predicate to match even GPR registers
19723 for quad memory operations.
19724 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19725 operations in both GPR and VSX registers.
19726 (quad_memory_operand): New predicate for quad memory operations.
19727 (reg_or_indexed_operand): New predicate for direct move support.
19729 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19730 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19731 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19732 (POWERPC_MASKS): Add power8 options.
19733 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19736 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19737 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19739 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19740 (-mpower8-fusion): New power8 options.
19741 (-mpower8-fusion-sign): Likewise.
19742 (-mpower8-vector): Likewise.
19743 (-mcrypto): Likewise.
19744 (-mdirect-move): Likewise.
19745 (-mquad-memory): Likewise.
19747 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19748 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19750 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19751 (rs6000_debug_vector_unit): Add p8_vector.
19752 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19753 definitions. Also print fusion state.
19754 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19755 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19756 (rs6000_option_override_internal): Add support for power8.
19757 (rs6000_common_init_builtins): Add debugging for skipped builtins
19758 if -mdebug=builtin.
19759 (rs6000_adjust_cost): Add power8 support.
19760 (rs6000_issue_rate): Likewise.
19761 (insn_must_be_first_in_group): Likewise.
19762 (insn_must_be_last_in_group): Likewise.
19763 (force_new_group): Likewise.
19764 (rs6000_register_move_cost): Likewise.
19765 (rs6000_opt_masks): Likewise.
19767 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19768 power8 capable assembler, default to power7 options.
19769 (TARGET_DIRECT_MOVE): Likewise.
19770 (TARGET_CRYPTO): Likewise.
19771 (TARGET_P8_VECTOR): Likewise.
19772 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19773 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19774 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19775 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19776 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19777 (TARGET_XSCVDPSPN): Likewise.
19778 (TARGET_XSCVSPDPN): Likewsie.
19779 (TARGET_SYNC_HI_QI): Likewise.
19780 (TARGET_SYNC_TI): Likewise.
19781 (MASK_CRYPTO): Likewise.
19782 (MASK_DIRECT_MOVE): Likewise.
19783 (MASK_P8_FUSION): Likewise.
19784 (MASK_P8_VECTOR): Likewise.
19785 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19786 temporary used by some of the direct move instructions to get two FP
19787 temporary registers does not force creation of a stack frame.
19788 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19789 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19790 that any VSX registers are tieable, even if they are also an
19791 Altivec vector mode.
19792 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19793 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19794 (RS6000_BTM_CRYPTO): Likewise.
19795 (RS6000_BTM_COMMON): Likewise.
19797 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19798 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19799 (enum rs6000_vector): Add power8 vector support.
19801 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19805 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19806 (any_sibcall_could_use_r3): this and handle indirect calls.
19807 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19809 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19811 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19813 2013-05-22 Richard Biener <rguenther@suse.de>
19815 PR middle-end/57349
19816 * profile.c (branch_prob): Do not split blocks that are
19817 abnormally receiving from ECF_RETURNS_TWICE functions.
19819 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19821 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19822 address mode. Move pointer mode initialization to the same place.
19824 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19826 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19827 while it has any effect.
19829 2013-05-21 Easwaran Raman <eraman@google.com>
19831 PR tree-optimization/57322
19832 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19833 UID of the statement added to the BB to be 1.
19835 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19837 PR tree-optimization/57331
19838 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19839 of conversion from pointer type to integral type with integer.
19841 2013-05-21 Martin Jambor <mjambor@suse.cz>
19844 * ipa-prop.c (ipa_read_node_info): Process param_used and
19845 controlled_uses in the same order as when writing.
19847 2013-05-21 Magnus Granberg <baldrick@free.fr>
19850 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19852 2013-05-21 Richard Biener <rguenther@suse.de>
19854 PR tree-optimization/57318
19855 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19856 estimate stmts with side-effects as likely eliminated.
19858 2013-05-21 Richard Biener <rguenther@suse.de>
19860 PR tree-optimization/57330
19861 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19862 preserve the call stmts fntype.
19864 2013-05-21 Richard Biener <rguenther@suse.de>
19866 PR tree-optimization/57303
19867 * tree-ssa-sink.c (statement_sink_location): Improve killing
19868 stmt detection and properly handle self-assignments.
19870 2013-05-21 Christian Bruel <christian.bruel@st.com>
19872 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19873 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19874 registers. Set register size out of the PARALLEL loop.
19876 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19879 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19880 (*fmasf4, *fmasf4_media): New insns.
19882 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19884 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19885 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19886 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19887 (mips_idiv_insns): Update the comments to say that the returned
19888 instruction counts are in units of BASE_INSN_LENGTH.
19889 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19890 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19891 using 2 rather than 4 as the length of indirect MIPS16 and
19892 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19893 length of a NOP. Don't divide MIPS16 lengths by 2.
19894 (mips16_split_long_branches): Assume a branch is long if the
19895 length is greater than 4 rather than 8.
19896 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19897 rather than multiplying them by 2. Multiply instruction counts
19898 by BASE_INSN_LENGTH rather than 4.
19899 (*jump_mips16, tls_get_tp_mips16_<mode>)
19900 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19902 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19904 * config/mips/mips.md (extended_mips16): Remove branch case.
19905 (length): Remove duplicated extended_mips16 test.
19907 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19909 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19911 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19913 * recog.h (Recog_data): Rename to...
19914 (recog_data_d): ...this.
19915 (recog_data): Update accordingly.
19916 * recog.c (recog_data): Likewise.
19917 * reload.c (save_recog_data): Likewise.
19918 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19919 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19921 2013-05-17 Julian Brown <julian@codesourcery.com>
19923 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19924 found in a REG_EQUAL note, invalidate it.
19926 2013-05-17 Easwaran Raman <eraman@google.com>
19928 * tree-ssa-reassoc.c (find_insert_point): New function.
19929 (insert_stmt_after): Likewise.
19930 (get_def_stmt): Likewise.
19931 (ensure_ops_are_available): Likewise.
19932 (not_dominated_by): Likewise.
19933 (rewrite_expr_tree): Do not move statements beyond what is
19934 necessary. Remove call to swap_ops_for_binary_stmt...
19935 (reassociate_bb): ... and move it here.
19936 (build_and_add_sum): Assign UIDs for new statements.
19937 (linearize_expr): Likewise.
19938 (do_reassoc): Renumber gimple statement UIDs.
19940 2013-05-17 Jan Hubicka <jh@suse.cz>
19942 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19944 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19945 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19947 (output_weakrefs): Update.
19949 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19950 Martin Jambor <mjambor@suse.cz>
19952 PR middle-end/57276
19953 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19954 value that corresponds to the given aggval is found in values vector.
19956 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19958 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19959 sse, sse2, sse3, ssse3 and sse4a flags to options.
19961 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19963 * gengtype-state.c: (s_expr_writer): New class, to handle
19964 prettifying of output layout of s-expressions.
19965 (state_writer): New class, to write out gtype.state.
19966 (state_written_type_count): Move this variable into member data of
19968 (s_expr_writer::s_expr_writer): New code: constructor for new class
19969 (state_writer::state_writer(): ditto
19970 (s_expr_writer::write_new_line): New function
19971 (s_expr_writer::write_any_indent): ditto
19972 (s_expr_writer::begin_s_expr): ditto
19973 (s_expr_writer::end_s_expr): ditto
19974 (write_state_fileloc): convert to method of state_writer...
19975 (state_writer:: write_state_fileloc): ...and use methods of
19976 s_expr_writer to write indentation into the gtype.state output file
19977 to visually represent the hierarchical structure of the list
19979 (write_state_fields): ditto, renaming to...
19980 (state_writer::write_state_fields)
19981 (write_state_a_string): ditto, renaming to...
19982 (state_writer::write_state_a_string)
19983 (write_state_string_option): ditto, renaming to...
19984 (state_writer::write_state_string_option)
19985 (write_state_type_option): ditto, renaming to...
19986 (state_writer::write_state_type_option)
19987 (write_state_nested_option): ditto, renaming to...
19988 (state_writer::write_state_nested_option)
19989 (write_state_option): ditto, renaming to...
19990 (state_writer::write_state_option)
19991 (write_state_options): ditto, renaming to...
19992 (state_writer::write_state_options)
19993 (write_state_lang_bitmap): ditto, renaming to...
19994 (state_writer::write_state_lang_bitmap)
19995 (write_state_version): ditto, renaming to...
19996 (state_writer::write_state_version)
19997 (write_state_scalar_type): ditto, renaming to...
19998 (state_writer::write_state_scalar_type)
19999 (write_state_string_type): ditto, renaming to...
20000 (state_writer::write_state_string_type)
20001 (write_state_undefined_type): ditto, renaming to...
20002 (state_writer::write_state_undefined_type)
20003 (write_state_struct_union_type): ditto, renaming to...
20004 (state_writer::write_state_struct_union_type)
20005 (write_state_struct_type): ditto, renaming to...
20006 (state_writer::write_state_struct_type)
20007 (write_state_user_struct_type): ditto, renaming to...
20008 (state_writer::write_state_user_struct_type)
20009 (write_state_lang_struct_type): ditto, renaming to...
20010 (state_writer::write_state_lang_struct_type)
20011 (write_state_param_struct_type): ditto, renaming to...
20012 (state_writer::write_state_param_struct_type)
20013 (write_state_pointer_type): ditto, renaming to...
20014 (state_writer::write_state_pointer_type)
20015 (write_state_array_type): ditto, renaming to...
20016 (state_writer::write_state_array_type)
20017 (write_state_gc_used): ditto, renaming to...
20018 (state_writer::write_state_gc_used)
20019 (write_state_common_type_content): ditto, renaming to...
20020 (state_writer::write_state_common_type_content)
20021 (write_state_type): ditto, renaming to...
20022 (state_writer::write_state_type)
20023 (write_state_pair_list): ditto, renaming to...
20024 (state_writer::write_state_pair_list)
20025 (write_state_pair): ditto, renaming to...
20026 (state_writer::write_state_pair)
20027 (write_state_typedefs): ditto, renaming to...
20028 (state_writer::write_state_typedefs)
20029 (write_state_structures): ditto, renaming to...
20030 (state_writer::write_state_structures)
20031 (write_state_param_structs): ditto, renaming to...
20032 (state_writer::write_state_param_structs)
20033 (write_state_variables): ditto, renaming to...
20034 (state_writer::write_state_variables)
20035 (write_state_srcdir): ditto, renaming to...
20036 (state_writer::write_state_srcdir)
20037 (write_state_files_list): ditto, renaming to...
20038 (state_writer::write_state_files_list)
20039 (write_state_languages): ditto, renaming to...
20040 (state_writer::write_state_languages)
20041 (write_state): create a state_writer instance and use it when
20042 writing out the state file
20044 2013-05-17 Mike Stump <mikestump@comcast.net>
20046 PR rtl-optimization/57304
20047 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
20048 accessing DF_REF_REAL_LOC.
20050 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20052 PR rtl-optimization/57281
20053 PR rtl-optimization/57300
20054 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
20055 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
20056 what the other splitter did if the registers are dead.
20058 2013-05-17 Richard Biener <rguenther@suse.de>
20060 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
20063 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20065 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
20068 2013-05-17 Marek Polacek <polacek@redhat.com>
20070 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
20071 length when doing non-zero store of storing '\0' to '\0'.
20073 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20075 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
20076 vect_external_def oprnd1 with loop_vinfo, try to emit
20077 optional cast, negation and and stmts on the loop preheader
20078 edge instead of into the pattern def seq.
20080 PR tree-optimization/57051
20081 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
20082 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
20084 2013-05-16 Nick Clifton <nickc@redhat.com>
20086 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
20087 (rl78_is_naked_func): New function.
20088 (rl78_expand_prologue): Skip prologue generation for naked functions.
20089 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
20090 * doc/extend.texi (naked): Add RL78 to the list of processors
20091 that supports this attribute.
20093 2013-05-16 Jeff Law <law@redhat.com>
20095 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
20097 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20099 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
20100 cache parameters using detect_caches_amd also for CYRIX,
20101 NSC and TM2 signatures.
20103 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20104 Dzianis Kahanovich <mahatma@eu.by>
20108 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
20109 VIA/Centaur processors and determine their cache parameters
20110 using detect_caches_amd.
20112 2013-05-16 Teresa Johnson <tejohnson@google.com>
20114 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
20115 (rtl_verify_edges): New function.
20116 (rtl_verify_bb_insns): Ditto.
20117 (rtl_verify_bb_pointers): Ditto.
20118 (rtl_verify_bb_insn_chain): Ditto.
20119 (rtl_verify_fallthru): Ditto.
20120 (rtl_verify_bb_layout): Ditto.
20121 (rtl_verify_flow_info_1): Outline checks into new functions.
20122 (rtl_verify_flow_info): Ditto.
20124 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
20126 * cfghooks.c (copy_bbs): Add update_dominance argument.
20127 * cfghooks.h (copy_bbs): Update prototype.
20128 * tree-cfg.c (gimple_duplicate_sese_region):
20129 Add update_dominance argument.
20130 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
20131 * tree-ssa-loop-ch.c (copy_loop_headers): Update
20132 gimple_duplicate_sese_region call.
20133 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
20134 Update copy_bbs call.
20135 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
20136 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
20138 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20140 * tree-vectorizer.h (NUM_PATTERNS): Increment.
20141 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
20142 vect_recog_rotate_pattern.
20143 (vect_recog_rotate_pattern): New function.
20145 2013-05-16 Jason Merrill <jason@redhat.com>
20147 * Makefile.in (LLINKER): New variable.
20148 (mostlyclean): Remove link mutex.
20149 * configure.ac: Handle --enable-link-mutex.
20150 * lock-and-run.sh: New script.
20152 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20155 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
20158 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20160 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
20162 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
20164 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
20165 * config/arm/arm.c (next_consecutive_mem): New function.
20166 (gen_movmem_ldrd_strd): Likewise.
20167 * config/arm/arm.md (movmemqi): Update condition and code.
20168 (unaligned_loaddi, unaligned_storedi): New patterns.
20170 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20172 * config.gcc: Obsolete *-*-solaris2.9*.
20173 * doc/install.texi (Specific, *-*-solaris2*): Document it.
20175 2013-05-16 Richard Biener <rguenther@suse.de>
20177 * passes.c (init_optimization_passes): Move pass_parallelize_loops
20178 earlier, after GRAPHITE transforms and IV canonicalization.
20180 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20182 * omp-low.c (extract_omp_for_data): For collapsed loops,
20183 if at least one of the loops is known at compile time to
20184 iterate zero times, set count to 0.
20185 (expand_omp_regimplify_p): New function.
20186 (expand_omp_for_generic): For collapsed loops, if at least
20187 one of the loops isn't known to iterate at least once,
20188 add runtime check with setting count to 0.
20189 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
20190 For unsigned types if it isn't known at compile time that
20191 the loop will iterate at least once, add runtime check to bypass
20192 the whole loop if initial condition isn't true.
20194 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
20196 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
20198 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
20200 PR middle-end/57286
20201 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
20202 transformations to avoid an infinite loop.
20204 2013-05-16 Marek Polacek <polacek@redhat.com>
20206 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
20208 2013-05-15 Leif Ekblad <leif@rdos.net>
20210 * config/i386/i386.c (ix86_decompose_address): Use
20211 DEFAULT_TLS_SEG_REG to access TLS segment register.
20212 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
20213 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
20214 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
20216 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
20219 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
20220 sibling calls to functions that would normally be lazily bound,
20221 unless $gp is call-clobbered.
20223 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
20225 * config/i386/i386.c (ix86_option_override_internal): Update
20226 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
20227 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
20228 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
20229 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
20231 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
20233 2013-05-15 Andreas Schwab <schwab@suse.de>
20235 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
20236 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
20239 2013-05-15 Teresa Johnson <tejohnson@google.com>
20241 * loop-unroll.c (report_unroll_peel): Check decision before
20242 emitting unroll/peel message.
20244 2013-05-15 Teresa Johnson <tejohnson@google.com>
20246 * function.h (has_bb_partition): New rtl_data flag.
20247 (bb_reorder_complete): Ditto.
20248 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
20249 instead of flag_reorder_blocks_and_partition.
20250 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
20251 with some enhancements.
20252 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
20253 * bb-reorder.c (connect_traces): Check for has_bb_partition
20254 instead of flag_reorder_blocks_and_partition.
20255 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
20256 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
20257 verify_hot_cold_block_grouping.
20258 (partition_hot_cold_basic_blocks): Set has_bb_partition.
20260 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20263 * config/arm/predicates.md (call_insn_operand): New predicate.
20264 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
20265 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
20266 if insn is not a tail call.
20267 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
20269 * config/arm/arm.h (enum reg_class): New caller save register class.
20270 (REG_CLASS_NAMES): Likewise.
20271 (REG_CLASS_CONTENTS): Likewise.
20272 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
20275 2013-05-15 Richard Biener <rguenther@suse.de>
20277 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
20278 of MSG_OPTIMIZED_LOCATIONS.
20279 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
20280 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
20282 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
20283 of MSG_OPTIMIZED_LOCATIONS.
20284 (execute_vect_slp): Likewise.
20285 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
20286 (vect_create_cond_for_alias_checks): Likewise.
20287 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
20288 (vect_recog_widen_mult_pattern): Likewise.
20289 (vect_recog_widen_sum_pattern): Likewise.
20290 (vect_recog_over_widening_pattern): Likewise.
20291 (vect_recog_widen_shift_pattern): Likewise.
20292 (vect_recog_vector_vector_shift_pattern): Likewise.
20293 (vect_recog_divmod_pattern): Likewise.
20294 (vect_recog_mixed_size_cond_pattern): Likewise.
20295 (vect_recog_bool_pattern): Likewise.
20296 (vect_pattern_recog_1): Likewise.
20298 2013-05-15 Martin Jambor <mjambor@suse.cz>
20300 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
20301 non-functions to builtin_unreachable.
20302 * ipa-inline-transform.c (inline_call): Do not assert estimates were
20303 correct when new direct edges were discovered.
20305 2013-05-15 Martin Jambor <mjambor@suse.cz>
20307 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
20308 header, print symbol order instead of node uid, print more information
20309 about indirect edge targets.
20310 (ipa_make_edge_direct_to_target): Print symbol order instead of node
20312 (ipa_make_edge_direct_to_target): Likewise.
20313 (remove_described_reference): Likewise.
20314 (propagate_controlled_uses): Likewise.
20315 (ipa_print_node_params): Also print symbol order.
20316 (ipcp_transform_function): Print symbol order instead of node uids.
20317 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
20318 (cgraph_get_create_real_symbol_node): Likewise.
20319 * ipa-cp.c (print_lattice): Likewise.
20320 (print_all_lattices): Likewise.
20321 (determine_versionability): Likewise.
20322 (initialize_node_lattices): Likewise.
20323 (estimate_local_effects): Likewise.
20324 (update_profiling_info): Likewise.
20325 (create_specialized_node): Likewise.
20326 (perhaps_add_new_callers): Likewise.
20327 (decide_about_value): Likewise.
20328 (decide_whether_version_node): Likewise.
20329 (identify_dead_nodes): Likewise.
20330 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
20331 (dump_inline_summary): Likewise.
20332 (estimate_node_size_and_time): Likewise.
20333 (inline_analyze_function): Likewise.
20334 * ipa-inline.c (report_inline_failed_reason): Likewise.
20335 (want_early_inline_function_p): Likewise.
20336 (edge_badness): Likewise.
20337 (update_edge_key): Likewise.
20338 (inline_small_functions): Likewise. Add dumping of order to two other
20340 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
20341 instead of node uids.
20342 (propagate_pure_const): Likewise.
20343 (propagate_pure_const): Likewise.
20344 * ipa-utils.c (dump_cgraph_node_set): Likewise.
20345 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
20346 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
20348 * tree-pretty-print.c (dump_function_header): Likewise.
20349 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
20350 Print symbol order instead of node uids.
20352 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20354 * config/s390/s390.c (s390_register_move_cost): Don't impose the
20355 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
20357 2013-05-15 Richard Biener <rguenther@suse.de>
20359 PR tree-optimization/57275
20360 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
20361 return value for fail to do runtime alias checks for gather loads.
20363 2013-05-15 Jan Hubicka <jh@suse.cz>
20367 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
20368 weakrefs are not external.
20369 (lto_symtab_merge_decls): Fix thinko when dealing with
20370 non-lto_symtab decls.
20371 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
20372 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
20373 * varpool.c (dump_varpool_node): Dump more flags.
20375 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20377 * config/i386/i386.c (processor_alias_table): Add instruction
20378 FSGSBASE for AMD bdver3 architecture.
20380 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20382 * tree.c (warn_deprecated_use): Print file:line using locus color.
20383 * diagnostic.c (diagnostic_report_current_module): Print file:line
20384 and file:line:column using locus color.
20386 2013-05-14 Mike Stump <mikestump@comcast.net>
20388 * gdbinit.in: Add __null.
20390 2013-05-14 Mike Stump <mikestump@comcast.net>
20392 * recog.h: Rename struct recog_data to Recog_data.
20393 * recog.c: Likewise.
20394 * reload.c (can_reload_into): Likewise.
20395 * config/picochip/picochip.c: Likewise.
20397 2013-05-14 Mike Stump <mikestump@comcast.net>
20399 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
20401 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
20403 * resource.h (struct resources): Remove unch_memory member.
20404 (CLEAR_RESOURCE): Don't clear unch_memory.
20405 * resource.c (mark_referenced_resources): Don't set it.
20406 (mark_set_resources): Likewise.
20407 (mark_target_live_regs): Don't clear it.
20408 (init_resource_info): Likewise.
20409 * reorg.c (resource_conflicts_p): Don't compare it.
20410 (redundant_insn): Don't set it.
20412 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20414 * emit-rtl.c (next_label): Remove unused function.
20415 (skip_consecutive_labels, link_cc0_insns): Move to ...
20416 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20417 only place where these functions are used, and make them static.
20419 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20421 * fold-const.c (fold_negate_expr): Handle vectors.
20422 (fold_truth_not_expr): Make it static.
20423 (fold_invert_truthvalue): New static function.
20424 (invert_truthvalue_loc): Handle vectors. Do not call
20425 fold_truth_not_expr directly.
20426 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20427 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20428 (fold_comparison): Handle vector constants.
20429 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20430 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20431 * tree.h (fold_truth_not_expr): Remove declaration.
20433 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20435 * config/aarch64/aarch64-simd.md
20436 (aarch64_vcond_internal<mode>): Rename to...
20437 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20438 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20439 float modes. Clarify all iterator modes.
20440 (vcond<mode><mode>): Use new name for vcond expanders.
20441 (vcond<v_cmp_result><mode>): Likewise.
20442 (vcondu<mode><mode>: Likewise.
20443 * config/aarch64/iterators.md (VDQF_COND): New.
20445 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20448 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20449 variable for the shift amount. Check that we shift by non-negative
20452 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20455 * config/arm/arm.h (EPILOGUE_USES): Only return true
20456 for LR_REGNUM after epilogue_completed.
20458 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20460 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20461 is error_mark_node.
20463 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20466 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20467 and Solaris 11+/x86 with gld.
20468 * configure: Regenerate.
20470 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20472 * expmed.c (expand_shift_1): Canonicalize rotates by
20473 constant bitsize / 2 to bitsize - 1.
20474 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20475 case ROTATERT>: Likewise.
20478 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20480 * config/i386/i386.md (rotateinv): New code attr.
20481 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20482 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20483 roll $31, %eax, etc.
20485 2013-05-14 Richard Biener <rguenther@suse.de>
20487 PR middle-end/57235
20488 * tree-eh.c (sink_clobbers): Give up for successors with
20489 multiple predecessors and no virtual uses.
20491 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20493 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20494 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20496 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20498 PR middle-end/57251
20499 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20500 the case when both op0 and op1 have VOIDmode.
20502 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20504 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20505 in multiply-accumulate mode.
20507 2013-05-13 Guozhi Wei <carrot@google.com>
20509 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20511 2013-05-13 Kai Tietz <ktietz@redhat.com>
20514 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20515 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20516 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20517 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20518 (ix86_expand_prologue): Likewise.
20519 (ix86_expand_split_stack_prologue): Likewise.
20520 (legitimate_pic_address_disp_p): Likewise.
20521 (legitimize_pic_address): Likewise.
20522 (legitimize_tls_address): Likewise.
20523 (legitimize_pe_coff_symbol): Likewise.
20524 (output_pic_addr_const): Likewise.
20525 (construct_plt_address): Likewise.
20526 (ix86_expand_call): Likewise.
20527 (x86_output_mi_thunk): Likewise.
20528 (x86_function_profiler): Likewise.
20530 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20532 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20533 similar switch cases.
20534 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20535 (aarch64_simd_mov_to_<mode>low): Delete.
20536 (aarch64_simd_mov_to_<mode>high): Delete.
20537 (move_lo_quad_<mode>): Add w<-r alternative.
20538 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20539 (aarch64_simd_mov_from_*): Update type attribute.
20540 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20543 2013-05-13 Jan Hubicka <jh@suse.cz>
20545 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20546 * config/i386/i386.c (ix86_compute_frame_layout,
20547 ix86_expand_epilogue, emit_i387_cw_initialization,
20548 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20549 ix86_local_alignment): Fix use of size/speed predicates.
20551 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20553 PR tree-optimization/45216
20554 PR tree-optimization/57157
20555 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
20556 the (-Y) & (B - 1) variant if OP is |.
20557 * expmed.c (expand_shift_1): For rotations by const0_rtx just
20558 return shifted. Use (-op1) & (prec - 1) as other_amount
20559 instead of prec - op1.
20561 2013-05-13 Martin Jambor <mjambor@suse.cz>
20563 PR middle-end/42371
20564 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
20565 (ipa_constant_data): New type.
20566 (ipa_jump_func): Use ipa_constant_data to hold information about
20567 constant jump functions.
20568 (ipa_get_jf_constant): Adjust to jump function type changes.
20569 (ipa_get_jf_constant_rdesc): New function.
20570 (ipa_param_descriptor): New field controlled_uses.
20571 (ipa_get_controlled_uses): New function.
20572 (ipa_set_controlled_uses): Likewise.
20573 * ipa-ref.h (ipa_find_reference): Declare.
20574 * ipa-prop.c (ipa_cst_ref_desc): New type.
20575 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
20577 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
20578 New parameter cs. Adjust all callers.
20579 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
20580 (remove_described_reference): New function.
20581 (jfunc_rdesc_usable): Likewise.
20582 (try_make_edge_direct_simple_call): Decrement controlled use count,
20583 attempt to remove reference if it hits zero.
20584 (combine_controlled_uses_counters): New function.
20585 (propagate_controlled_uses): Likewise.
20586 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
20587 (ipa_edge_duplication_hook): Duplicate reference descriptions.
20588 (ipa_print_node_params): Print described use counter.
20589 (ipa_write_jump_function): Adjust to jump function type changes.
20590 (ipa_read_jump_function): New parameter CS, pass it to
20591 ipa_set_jf_constant. Adjust caller.
20592 (ipa_write_node_info): Stream controlled use count
20593 (ipa_read_node_info): Likewise.
20594 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
20596 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
20597 count. Remove cloning-added reference if it reaches zero.
20598 * ipa-ref.c (ipa_find_reference): New function.
20600 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20602 * config/i386/i386.c (processor_target_table): Modified default
20603 alignment values for AMD BD and BT architectures.
20605 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20607 * tree-vect-generic.c (uniform_vector_p): Move ...
20608 * tree.c (uniform_vector_p): ... here.
20609 * tree.h (uniform_vector_p): Declare it.
20610 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20613 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20615 PR tree-optimization/57230
20616 * tree-ssa-strlen.c (handle_char_store): Record length for
20617 array store from STRING_CST.
20619 PR tree-optimization/57230
20620 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20623 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20625 * config/epiphany/epiphany.c (epiphany_init): Check size of
20626 NUM_MODES_FOR_MODE_SWITCHING.
20627 (epiphany_expand_prologue):
20628 Remove CONFIG_REGNUM initial value handling code.
20629 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20630 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20631 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20632 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20633 Don't return 1 for FP_MODE_NONE.
20634 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20635 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20636 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20637 * config/epiphany/epiphany.md (save_config): New pattern.
20639 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20641 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20643 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20645 * config/i386/i386.md (memory): Handle sseishft1.
20646 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20647 (*vec_extractv2di_1): Ditto.
20649 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20651 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20654 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20656 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20657 Add mthumb/march=armv7-a multilib.
20658 Add mthumb/march=armv7-r multilib.
20659 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20661 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20663 * config/v850/t-rtems: Add more multilibs.
20665 2013-05-10 Richard Biener <rguenther@suse.de>
20667 PR tree-optimization/57214
20668 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20669 not propagate from SSA names that occur in abnormal PHI nodes.
20671 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20673 * stor-layout.c (element_precision): New function.
20674 * machmode.h (element_precision): Declare it.
20675 * tree.c (build_minus_one_cst): New function.
20676 (element_precision): Likewise.
20677 * tree.h (build_minus_one_cst): Declare new function.
20678 (element_precision): Likewise.
20679 * fold-const.c (operand_equal_p): Use element_precision.
20680 (fold_binary_loc): Handle vector types.
20681 * convert.c (convert_to_integer): Use element_precision.
20682 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20685 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20687 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20688 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20689 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20690 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20691 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20692 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20693 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20694 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20695 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20696 (Uuw8): New constraints.
20697 (Usb4): Move into alphabetical order.
20698 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20699 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20700 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20701 previously unnamed patterns.
20702 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20703 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20704 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20705 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20706 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20707 of specific lengths.
20709 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20711 * config/i386/i386.md (rotateinv): New code attr.
20712 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20713 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20714 roll $31, %eax, etc.
20716 PR tree-optimization/45216
20717 PR tree-optimization/57157
20718 * tree-ssa-forwprop.c (simplify_rotate): New function.
20719 (ssa_forward_propagate_and_combine): Call it.
20721 2013-05-10 Richard Biener <rguenther@suse.de>
20723 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20724 disable peeling when we version for aliasing.
20725 (vector_alignment_reachable_p): Honor explicit user alignment.
20726 (vect_supportable_dr_alignment): Likewise.
20727 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20728 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20729 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20730 then peeling to arrange for the cost-model check to come first.
20732 2013-05-10 Alan Modra <amodra@gmail.com>
20734 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20735 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20736 * configure: Regenerate.
20738 2013-05-10 Alan Modra <amodra@gmail.com>
20741 * varasm.c (default_elf_select_section): Move !DECL_P check..
20742 (get_named_section): ..to here before calling get_section_name.
20744 (default_section_type_flags): Add DECL_P check.
20745 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20746 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20748 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20750 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20751 When using gen_stack_adjust_str with a register offset, add a
20752 REG_FRAME_RELATED_EXPR note.
20754 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20756 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20757 (*vec_extractv4si_zext_mem): Ditto.
20758 (*vec_extractv2di): Add 0->x and x->x alternatives.
20759 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20760 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20762 2013-05-09 Jason Merrill <jason@redhat.com>
20764 N3639 C++1y VLA support
20765 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20768 * tree.c (build_constructor_va): New.
20769 * tree.h: Declare it.
20771 2013-05-09 Martin Jambor <mjambor@suse.cz>
20774 * gimple-fold.c (canonicalize_constructor_val): Call
20775 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20777 2013-05-09 Jan Hubicka <jh@suse.cz>
20778 Richard Biener <rguenther@suse.de>
20781 * symtab.c (symtab_make_decl_local): Do not add private names.
20783 2013-05-09 Jan Hubicka <jh@suse.cz>
20786 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20787 (unlink_from_assembler_name_hash): Likewise.
20788 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20789 symtab_unregister_node, symtab_initialize_asm_name_hash,
20790 change_decl_assembler_name): Update.
20792 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20794 * config/aarch64/aarch64.md: New movtf split.
20795 (*movtf_aarch64): Update.
20796 (aarch64_movdi_tilow): Handle TF modes and rename to
20797 aarch64_movdi_<mode>low.
20798 (aarch64_movdi_tihigh): Handle TF modes and rename to
20799 aarch64_movdi_<mode>high
20800 (aarch64_movtihigh_di): Handle TF modes and rename to
20801 aarch64_mov<mode>high_di
20802 (aarch64_movtilow_di): Handle TF modes and rename to
20803 aarch64_mov<mode>low_di
20804 (aarch64_movtilow_tilow): Remove spurious whitespace.
20805 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20807 (aarch64_print_operand): Update.
20809 2013-05-09 Alan Modra <amodra@gmail.com>
20811 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20813 * configure: Regenerate.
20815 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20817 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20818 splitter preparation statements.
20819 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20820 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20823 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20825 * gimple-ssa-strength-reduction.c (count_candidates): Change
20826 return value to int.
20827 (analyze_candidates_and_replace): Change type of length to int.
20829 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20831 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20832 (*vec_extract<mode>): Use VI12_128 mode iterator.
20833 (*vec_extract<mode>_mem): Ditto.
20834 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20837 2013-05-08 Diego Novillo <dnovillo@google.com>
20842 2012-08-17 Diego Novillo <dnovillo@google.com>
20845 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20846 * config.in: Regenerate.
20847 * configure: Regenerate.
20848 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20850 2013-05-08 Jan Hubicka <jh@suse.cz>
20853 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20854 * cgraph.h (symtab_node_base): Add unique_name.
20855 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20856 input_overwrite_node, input_varpool_node): Stream unique_name.
20857 * cgraphclones.c (cgraph_create_virtual_clone,
20858 cgraph_function_versioning): Set unique_name.
20859 * ipa.c (function_and_variable_visibility): Set unique_name.
20861 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20863 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20864 (alloc_cand_and_find_basis): Restrict conditional candidate
20865 processing to CAND_MULTs.
20867 2013-05-08 Jan Hubicka <jh@suse.cz>
20870 lto-symtab.c (lto_symtab_symbol_p): New function.
20871 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20872 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20873 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20874 Skip static symbols.
20876 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20878 PR tree-optimization/57200
20879 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20880 Only call inform if the preceding warning_at returns true.
20882 2013-05-07 Han Shen <shenhan@google.com>
20884 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20885 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20886 * common.opt (fstack-protector-strong): New option.
20887 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20888 * doc/invoke.texi (Optimization Options): Document
20889 "-fstack-protector-strong".
20890 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20892 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20894 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20896 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20898 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20899 it is not lhs of the insn.
20900 (setup_reg_equiv): Remove insn having equiv memory which it is not
20902 * lra-constraints.c (process_address): Try to improve generation
20903 code for address base + disp.
20904 (lra_constraints): Make correct the code for checking insn setting
20905 up backward equivalence. Remove insn only if it is in the init
20907 * lra-eliminations.c (update_reg_eliminate): Change return value.
20908 (lra_eliminate): Use the result.
20910 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20912 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20913 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20914 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20915 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20916 register target operands.
20917 (*vec_extractv8hi_sse2): New pattern.
20918 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20919 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20920 (*vec_extract<mode>_mem): New insn and split pattern.
20922 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20924 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20925 (TARGET_ASAN_SHADOW_OFFSET): Define.
20926 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20927 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20929 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20931 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20932 (incr_vec_index): Return -1 if increment not found.
20933 (create_add_on_incoming_edge): Assert if increment not found.
20934 (record_increment): Limit number of increments recorded.
20935 (all_phi_incrs_profitable): Return false if an increment not found.
20936 (replace_profitable_candidates): Don't process increments that were
20938 (analyze_candidates_and_replace): Limit size of incr_vec.
20940 2013-05-07 Richard Biener <rguenther@suse.de>
20942 * calls.c (special_function_p): setjmp-like functions are leaf.
20943 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20944 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20946 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20948 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20949 (aarch64_simd_mov<mode>): New expander.
20950 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20951 (aarch64_simd_mov_to_<mode>high): Likewise.
20952 (aarch64_simd_mov_from_<mode>low): Likewise.
20953 (aarch64_simd_mov_from_<mode>high): Likewise.
20954 (aarch64_dup_lane<mode>): Update.
20955 (aarch64_dup_lanedi): New instruction pattern.
20956 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20957 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20959 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20961 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20962 (replace_mult_candidate): Remove unnecessary argument; remove
20963 unnecessary parameter from call to introduce_cast_before_cand.
20964 (replace_unconditional_candidate): Remove unnecessary parameter
20965 from call to replace_mult_candidate.
20966 (replace_conditional_candidate): Likewise.
20967 (insert_initializers): Use make_temp_ssa_name.
20968 (introduce_cast_before_cand): Remove unnecessary argument; use
20969 make_temp_ssa_name.
20970 (replace_one_candidate): Remove unnecessary argument; remove
20971 unnecessary parameter from calls to introduce_cast_before_cand.
20972 (replace_profitable_candidates): Remove unnecessary parameters
20973 from calls to replace_one_candidate.
20975 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20977 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20978 phi def as possibly hiding a basis for a CAND_ADD whose operands
20979 have been commuted in the analysis.
20980 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20982 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20984 * config/aarch64/aarch64.md
20985 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20986 shift value between 0-4.
20988 2013-05-07 Richard Biener <rguenther@suse.de>
20990 * double-int.h (rshift): New overload.
20991 * double-int.c (rshift): New function.
20992 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20993 (create_reference_ops_from_ref): Remove.
20994 (vn_reference_insert): Use shared ops for constructing the
20995 reference and copy it.
20997 2013-05-07 Richard Biener <rguenther@suse.de>
20999 PR middle-end/57190
21000 * tree-eh.c (sink_clobbers): Properly propagate
21001 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
21003 2013-05-07 Jakub Jelinek <jakub@redhat.com>
21005 PR tree-optimization/57149
21006 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
21007 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
21008 collect_phi_def_edges, execute_late_warn_uninitialized): Use
21009 uninit_undefined_value_p instead of ssa_undefined_value_p.
21012 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
21013 for modifier == EXPAND_INITIALIZER.
21015 2013-05-07 Anton Blanchard <anton@samba.org>
21017 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
21018 for powerpc64 little endian.
21019 * configure: Regenerate.
21021 2013-05-06 Graham Stott <grahams@btinternet.com>
21023 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
21024 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
21025 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
21026 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
21028 2013-05-06 Graham Stott <grahams@btinternet.com>
21030 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
21031 codes which allow non-lvalues.
21033 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
21035 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
21036 components are all 1s.
21037 (integer_minus_onep): New function.
21038 * tree.h (integer_minus_onep): Declare it.
21039 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
21040 integer_minus_onep instead of integer_all_onesp.
21042 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21045 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
21046 variations of these patterns.
21048 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21050 * config/i386/i386.md (isa): Add x64_sse4 member.
21051 (enabled): Handle x64_sse4.
21052 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
21053 instruction for 64bit SSE4_1 targets. Update insn attributes.
21054 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
21055 instruction for SSE4_1 targets. Update insn attributes.
21056 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21057 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
21058 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
21060 (*vec_extractv4si): Rename from *sse4_1_pextrd.
21061 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
21062 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
21064 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21067 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
21069 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
21071 * final.c (do_assembler_dialects): Don't handle curly braces and
21072 vertical bar escaped by % as dialect delimiters.
21073 (output_asm_insn): Print curly braces and vertical bar if escaped
21074 by % and ASSEMBLER_DIALECT defined.
21075 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
21076 * doc/tm.texi: Regenerated.
21078 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
21080 * config/mips/mips.c: Include tree-pass.h.
21081 (mips_reorg): Split in pre- and post-dbr_schedule parts.
21082 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
21083 (pass_mips_machine_reorg2): New machine specific pass.
21084 (insert_pass_mips_machine_reorg2): New pass plugin definition.
21085 (mips_option_override): Register the new pass.
21086 * rtl.h (cleanup_barriers): Remove prototype.
21087 (dbr_schedule): Likewise.
21088 * jump.c (cleanup_barriers): Make static.
21089 * reorg.c (dbr_schedule): Likewise.
21091 2013-05-06 Richard Biener <rguenther@suse.de>
21093 PR tree-optimization/57185
21094 * tree-parloops.c (add_field_for_reduction): Handle anonymous
21095 SSA names properly.
21097 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21100 * config/i386/i386.c (add_parameter_dependencies): Add dependence
21101 between "first_arg" and "insn", not "last" and "insn".
21103 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21105 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
21106 (find_candidates_in_block): Re-enable slsr_process_phi.
21107 (create_phi_basis): Fix double counting of candidate adjustment.
21109 2013-05-06 Richard Biener <rguenther@suse.de>
21111 PR middle-end/57147
21112 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
21113 the edge is also fallthru, preserve it and just clear the
21115 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
21116 also complex, preserve that and just clear the fallthru flag.
21117 * tree-inline.c (update_ssa_across_abnormal_edges): Also
21118 update virtual operands.
21120 2013-05-06 Alan Modra <amodra@gmail.com>
21122 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
21123 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
21124 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
21125 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
21126 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21127 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
21129 2013-05-06 Alan Modra <amodra@gmail.com>
21131 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
21132 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
21133 (DEFAULT_ASM_ENDIAN): Define.
21134 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21135 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
21136 Update -K PIC clause from sysv4.h.
21137 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
21138 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
21140 2013-05-06 Alan Modra <amodra@gmail.com>
21142 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
21143 twice for little-endian.
21144 (ashrdi3_no_power, ashrdi3): Support little-endian.
21146 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21149 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
21150 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
21152 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
21153 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
21154 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
21157 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
21158 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21160 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
21161 * config/pa/pa.opt: Make mbig-switch a no-op.
21162 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
21163 (CASE_VECTOR_MODE): Always return SImode.
21164 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
21165 for the !TARGET_BIG_SWITCH case.
21166 * config/pa/pa-linux.h: Likewise.
21167 * config/pa/pa-openbsd.h: Likewise.
21168 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
21169 * config/pa/pa.md (short_jump): Remove define_insn.
21170 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
21171 (casesi0): Remove define_insn.
21172 (type): Remove btable_branch.
21173 (pa_combine_type): Likewise.
21174 (in_nullified_branch_delay): Likewise.
21175 (in_call_delay): Likewise.
21176 (define_delay): Likewise.
21177 (define_insn_reservation "Z3"): Likewise.
21178 (define_insn_reservation "Z4"): Likewise.
21179 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
21180 (pa_adjust_insn_length): Remove adjustment for btable branches.
21181 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
21184 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
21186 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21187 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
21188 Add m->r,x alternatives.
21189 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
21190 splitters using SWI48x mode iterator.
21191 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
21192 TARGET_64BIT. Add m->x alternative.
21193 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
21194 Add o->x alternative. Enable for TARGET_SSE.
21195 (sse_storeq): Remove expander.
21196 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
21197 with memory input operand.
21198 (*vec_extractv2di_1 splitter): New.
21199 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
21200 * config/i386/i386.md (ssevecmodelower): New mode attribute.
21202 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21204 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
21205 (INT_LOWPART): Delete.
21206 (extract_MB): Adjust.
21207 (extract_ME): Adjust.
21208 (print_operand): Adjust.
21210 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21212 * config/rs6000/predicates.md (reg_or_add_cint_operand,
21213 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21214 (reg_or_logical_cint_operand, easy_fp_constant,
21215 logical_const_operand): Delete "CONST_DOUBLE" case.
21216 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
21217 "HOST_BITS_PER_WIDE_INT == 64" test.
21218 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
21219 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
21220 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
21221 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
21223 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
21224 CONST_DOUBLE DImode/VOIDmode case.
21225 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
21226 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21227 CONST_DOUBLE VOIDmode case.
21228 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21229 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
21230 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
21231 Delete CONST_DOUBLE case.
21232 (splitters for mov FMOVE64 const_double): Delete
21233 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21234 "HOST_BITS_PER_WIDE_INT >= 64" test.
21235 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
21237 (mov DI const_double): Delete.
21239 2013-05-04 Jakub Jelinek <jakub@redhat.com>
21241 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
21242 on op shows all bits zero in mode of a lowpart subreg, return zero.
21244 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
21247 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
21248 to save TFmode registers and DImode to save TImode registers for
21249 caller save operations.
21250 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
21251 mark being partially clobbered since they only use the first
21254 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
21255 and TDmode only use the upper 64-bits of each VSX register.
21257 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21259 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
21260 (find_candidates_in_block): Disable slsr_process_phi.
21262 2013-05-03 Guozhi Wei <carrot@google.com>
21264 * coverage.c (coverage_obj_init): Move the construction of gcov
21266 (build_init_ctor): ... here.
21268 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21270 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
21271 (slsr_cand_d): Redefine def_phi.
21272 (stride_status, phi_adjust_status, count_phis_status): New enums.
21273 (find_phi_def): New.
21274 (find_basis_for_base_expr): New.
21275 (find_basis_for_candidate): Handle hidden bases.
21276 (alloc_cand_and_find_basis): Handle phi candidates.
21277 (slsr_process_phi): New.
21278 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
21279 (create_mul_imm_cand): Likewise.
21280 (create_add_ssa_cand): Exclude phi base candidates.
21281 (create_add_imm_cand): Likewise.
21282 (slsr_process_cast): Likewise.
21283 (slsr_process_copy): Likewise.
21284 (find_candidates_in_block): Handle phi candidates.
21285 (dump_candidate): Likewise.
21286 (unconditional_cands): Delete.
21287 (unconditional_cands_with_known_stride_p): Delete.
21288 (phi_dependent_cand_p): New.
21289 (cand_increment): Handle phi-dependent candidates.
21290 (replace_dependent): Delete.
21291 (replace_mult_candidate): New.
21292 (replace_unconditional_candidate): New.
21293 (incr_vec_index): Move to avoid forward reference.
21294 (create_add_on_incoming_edge): New.
21295 (create_phi_basis): New.
21296 (replace_dependents): Delete.
21297 (replace_conditional_candidate): New.
21298 (phi_add_costs): New.
21299 (replace_uncond_cands_and_profitable_phis): New.
21300 (record_increment): Handle phi adjustments.
21301 (record_phi_increments): New.
21302 (record_increments): Handle phi adjustments.
21303 (phi_incr_cost): New.
21304 (lowest_cost_path): Handle phis.
21305 (total_savings): Likewise.
21306 (analyze_increments): Likewise.
21307 (ncd_with_phi): New.
21308 (ncd_of_cand_and_phis): New.
21309 (nearest_common_dominator_for_cands): Handle phi increments.
21310 (all_phi_incrs_profitable): New.
21311 (replace_profitable_candidates): Handle phi-dependent candidates.
21312 (analyze_candidates_and_replace): Likewise.
21314 2013-05-03 Teresa Johnson <tejohnson@google.com>
21317 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
21318 do not exceed REG_BR_PROB_BASE.
21320 2013-05-03 Jeff Law <law@redhat.com>
21322 PR tree-optimization/57144
21323 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
21324 operand of the condition will bit into the new type when eliminating
21325 a cast feeding a condition.
21327 2013-05-03 Jakub Jelinek <jakub@redhat.com>
21329 PR rtl-optimization/57130
21330 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
21331 of COMPARE as in_code to the recursive call if needed.
21333 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
21335 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
21336 (enabled): Handle new members.
21337 * config/i386/sse.md (*vec_concatv2si): Merge from
21338 *vec_concatv2si_sse2 and vec_concatv2si_sse.
21339 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
21341 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
21343 PR tree-optimization/57027
21344 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
21345 for fnms opportunity, check we got the prerequisite kind
21346 of tree / gimple before using accessor functions.
21348 2013-05-03 Richard Biener <rguenther@suse.de>
21350 * double-int.h (lshift): New overload without precision
21351 and arith argument.
21352 (operator *=, operator +=, operator -=): Move ...
21353 * double-int.c (operator *=, operator +=, operator -=): ... here
21354 and implement more efficiently.
21355 (mul_double_with_sign): Remove.
21356 (lshift_double): Adjust to take unsinged shift argument, push
21357 dispatching code to callers.
21358 (mul_double_wide_with_sign): Add early out for callers that
21359 are not interested in high parts or overflow.
21360 (lshift): New function.
21361 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
21362 dispatch code here.
21363 (lrotate, rrotate): Use logical shifts.
21364 * expr.c (get_inner_reference): Use lshift.
21365 * fixed-value.c (do_fixed_divide): Likewise.
21366 * tree-dfa.c (get_ref_base_and_extent): Likewise.
21367 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
21368 (indirect_refs_may_alias_p): Likewise.
21369 (stmt_kills_ref_p_1): Likewise.
21371 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21373 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
21375 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21377 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
21378 scalar form of FABD instruction.
21380 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
21382 * lra-constraints.c (process_alt_operands): Add checking alt
21383 number to choose the best alternative.
21385 2013-05-02 Richard Biener <rguenther@suse.de>
21387 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
21388 bitmap and its handling.
21389 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
21391 2013-05-02 Richard Biener <rguenther@suse.de>
21393 PR middle-end/57140
21394 * tree-inline.c (copy_loops): Properly handle removed loops.
21395 (copy_cfg_body): Mark destination loops for fixup if source
21396 loops needed fixup.
21398 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
21401 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
21402 generating simple_return for naked functions.
21404 2013-05-02 Martin Jambor <mjambor@suse.cz>
21406 PR middle-end/56988
21407 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
21408 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
21410 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21411 ipa_agg_replacement_value structures.
21412 (known_aggs_to_agg_replacement_list): Likewise.
21413 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21414 (read_agg_replacement_chain): Likewise.
21415 (ipcp_transform_function): Also check that by_ref flags match.
21417 2013-05-02 Richard Biener <rguenther@suse.de>
21419 * graphds.h (struct graph): Add obstack member.
21420 * graphds.c (new_graph): Initialize obstack and allocate
21422 (add_edge): Allocate edge from the obstack.
21423 (free_graph): Free the obstack instead of all edges and vertices.
21425 2013-05-02 Teresa Johnson <tejohnson@google.com>
21427 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21429 * cfg.c (update_bb_profile_for_threading): Ditto.
21430 * tree-inline.c (copy_bb): Ditto.
21431 (copy_edges_for_bb): Ditto.
21432 (initialize_cfun): Ditto.
21433 (copy_cfg_body): Ditto.
21434 (expand_call_inline): Ditto.
21435 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21436 (estimate_node_size_and_time): Ditto.
21437 (inline_merge_summary): Ditto.
21438 * cgraphclones.c (cgraph_clone_edge): Ditto.
21439 (cgraph_clone_node): Ditto.
21440 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21441 (compute_trg_info): Ditto.
21443 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21445 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21446 S reg when fp attribute set.
21447 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21449 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21451 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21453 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21454 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21455 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21457 2013-05-02 Richard Biener <rguenther@suse.de>
21459 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21460 (struct instantiate_cache_entry): New type.
21461 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21462 (struct instantiate_cache_type): New type.
21463 (set_instantiated_value, get_instantiated_value): Remove.
21464 (get_instantiated_value_entry): New function.
21465 (instantiate_scev_name): Use the new cache and adjust.
21466 (instantiate_scev_poly): Adjust.
21467 (instantiate_scev_binary): Likewise.
21468 (instantiate_array_ref): Likewise.
21469 (instantiate_scev_convert): Likewise.
21470 (instantiate_scev_not): Likewise.
21471 (instantiate_scev_3): Likewise.
21472 (instantiate_scev_2): Likewise.
21473 (instantiate_scev_r): Likewise.
21474 (instantiate_scev): Likewise.
21475 (resolve_mixers): Likewise.
21477 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21480 * lra-constraints.c (best_small_class_operands_num): Remove.
21481 (process_alt_operands): Remove small_class_operands_num. Take
21482 small classes operands into losers and only if the operand is not
21483 matched. Modify debugging output.
21484 (curr_insn_transform): Remove best_small_class_operands_num.
21487 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21489 * config/aarch64/aarch64-builtins.c
21490 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21491 * config/aarch64/aarch64-simd-builtins.def
21492 (reduc_splus_): Add new modes.
21493 (reduc_uplus_): New.
21494 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21495 (reduc_uplus_v4sf): Likewise.
21496 (reduc_splus_v4sf): Likewise.
21497 (aarch64_addv<mode>): Likewise.
21498 (reduc_uplus_<mode>): Likewise.
21499 (reduc_splus_<mode>): Likewise.
21500 (aarch64_addvv2di): Likewise.
21501 (reduc_uplus_v2di): Likewise.
21502 (reduc_splus_v2di): Likewise.
21503 (aarch64_addvv2si): Likewise.
21504 (reduc_uplus_v2si): Likewise.
21505 (reduc_splus_v2si): Likewise.
21506 (reduc_<sur>plus_<mode>): New.
21507 (reduc_<sur>plus_v2di): Likewise.
21508 (reduc_<sur>plus_v2si): Likewise.
21509 (reduc_<sur>plus_v4sf): Likewise.
21510 (aarch64_addpv4sf): Likewise.
21511 * config/aarch64/arm_neon.h
21512 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21513 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21514 add UNSPEC_SADDV, UNSPEC_UADDV.
21516 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21518 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21520 * config/aarch64/arm_neon.h
21521 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21523 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21525 * config/aarch64/aarch64-builtins
21526 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21528 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21530 * config/aarch64/aarch64-simd-builtins.def
21531 (reduc_smax_): New.
21532 (reduc_smin_): Likewise.
21533 (reduc_umax_): Likewise.
21534 (reduc_umin_): Likewise.
21535 (reduc_smax_nan_): Likewise.
21536 (reduc_smin_nan_): Likewise.
21539 (smax): Update for V2SF, V4SF and V2DF modes.
21542 (smin_nan): Likewise.
21543 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21544 (<su><maxmin><mode>3): ...This, refactor.
21545 (s<maxmin><mode>3): New.
21546 (<maxmin_uns><mode>3): Likewise.
21547 (reduc_<maxmin_uns>_<mode>): Refactor.
21548 (reduc_<maxmin_uns>_v4sf): Likewise.
21549 (reduc_<maxmin_uns>_v2si): Likewise.
21550 (aarch64_<fmaxmin><mode>: Remove.
21551 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21553 (vmin<q>_f<32,64>): Likewise.
21554 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21556 (su): Add mappings for smax, smin, umax, umin.
21558 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21559 (FMAXMIN): Rename as...
21560 (FMAXMIN_UNS): ...This.
21562 (fmaxminv): Likewise.
21563 (fmaxmin): Likewise.
21565 (maxmin_uns_op): Likewise.
21567 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21569 * config/aarch64/arm_neon.h
21570 (vac<ge, gt><sd>_f<32, 64>): Rename to...
21571 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
21572 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
21574 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21576 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
21577 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
21579 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21581 * config/aarch64/aarch64-simd.md
21582 (vcond<mode>_internal): Handle special cases for constant masks.
21583 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
21584 (vcondu<mode><mode>): Likewise.
21585 (vcond<v_cmp_result><mode>): New.
21587 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21589 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
21590 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
21591 * config/aarch64/aarch64-simd-builtins.def
21592 (cmeq): Update to BUILTIN_VALLDI.
21597 * config/aarch64/arm_neon.h
21598 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
21599 to builtins or C as appropriate.
21601 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21603 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21605 (cmhi): Rename to...
21607 * config/aarch64/aarch64-simd.md
21608 (simd_mode): Add SF.
21609 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21610 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21611 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21612 (cstore<mode>_neg): ...This.
21613 * config/aarch64/iterators.md
21615 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21616 (COMPARISONS): New.
21617 (UCOMPARISONS): Likewise.
21618 (optab): Add missing comparisons.
21624 (VCMP_S): Likewise.
21625 (VCMP_U): Likewise.
21626 (V_cmp_result): Add DF, SF modes.
21627 (v_cmp_result): Likewise.
21629 (vmtype): Likewise.
21630 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21632 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21634 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21635 define_insn to define_insn_and_split.
21636 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21637 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21638 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21639 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21640 (thumb2_negscc): Likewise.
21642 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21644 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21646 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21648 * config/arm/thumb2.md: Remove trailing whitespaces.
21650 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21652 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21653 Use gen_int_mode rather than GEN_INT.
21655 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21657 * value-prof.c (stream_in_histogram_value): Remove the strayed
21660 2013-04-30 Richard Biener <rguenther@suse.de>
21662 PR middle-end/57122
21663 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21665 2013-04-30 Richard Biener <rguenther@suse.de>
21667 PR middle-end/57107
21668 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21670 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21672 PR rtl-optimization/56957
21673 PR rtl-optimization/57105
21674 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21675 variable. Use just INSN_UID for determining whether an insn
21676 should be only disconnected from the insn stream.
21677 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21679 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21681 PR tree-optimization/57104
21682 * tsan.c (instrument_expr): Don't instrument accesses to
21683 DECL_HARD_REGISTER VAR_DECLs.
21685 2013-04-30 Richard Biener <rguenther@suse.de>
21687 * function.h (loops_for_fn): New inline function.
21688 (set_loops_for_fn): Likewise.
21689 * cfgloop.h (place_new_loop): Add struct function parameter.
21690 (get_loop): Likewise.
21691 (get_loops): Likewise.
21692 (number_of_loops): Likewise.
21693 (fel_next): Adjust.
21694 (fel_init): Likewise.
21695 * cfg.c (get_loop_copy): Adjust.
21696 * cfgloop.c (flow_loops_dump): Likewise.
21697 (record_loop_exits): Likewise.
21698 (verify_loop_structure): Likewise.
21699 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21700 (estimate_reg_pressure_cost): Likewise.
21701 (mark_loop_exit_edges): Likewise.
21702 * cfgloopmanip.c (place_new_loop): Likewise.
21703 (add_loop): Likewise.
21704 (duplicate_loop): Likewise.
21705 * graph.c (draw_cfg_nodes): Likewise.
21706 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21707 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21708 (extract_affine_chrec): Likewise.
21709 (build_scop_iteration_domain): Likewise.
21710 * graphite.c (graphite_initialize): Likewise.
21711 * ira-build.c (create_loop_tree_nodes): Likewise.
21712 (more_one_region_p): Likewise.
21713 (rebuild_regno_allocno_maps): Likewise.
21714 (mark_loops_for_removal): Likewise.
21715 (mark_all_loops_for_removal): Likewise.
21716 (remove_unnecessary_regions): Likewise.
21717 (ira_build): Likewise.
21718 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21719 * loop-init.c (fix_loop_structure): Likewise.
21720 (gate_rtl_move_loop_invariants): Likewise.
21721 (gate_rtl_unswitch): Likewise.
21722 (gate_rtl_unroll_and_peel_loops): Likewise.
21723 (rtl_doloop): Likewise.
21724 * lto-streamer-in.c (input_cfg): Likewise.
21725 * lto-streamer-out.c (output_cfg): Likewise.
21726 * modulo-sched.c (sms_schedule): Likewise.
21727 * predict.c (tree_estimate_probability): Likewise.
21728 (tree_estimate_probability_driver): Likewise.
21729 (estimate_loops): Likewise.
21730 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21731 (move_sese_region_to_fn): Likewise.
21732 (debug_loop_num): Likewise.
21733 * tree-chrec.c (chrec_evaluate): Likewise.
21734 (hide_evolution_in_other_loops_than_loop): Likewise.
21735 (chrec_component_in_loop_num): Likewise.
21736 (reset_evolution_in_loop): Likewise.
21737 (evolution_function_is_invariant_rec_p): Likewise.
21738 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21739 * tree-inline.c (copy_loops): Likewise.
21740 (copy_cfg_body): Likewise.
21741 (tree_function_versioning): Likewise.
21742 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21743 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21745 (add_to_evolution_1): Likewise.
21746 (scev_const_prop): Likewise.
21747 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21748 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21749 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21750 (tree_ssa_lim_initialize): Likewise.
21751 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21752 (verify_loop_closed_ssa): Likewise.
21753 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21754 (tree_ssa_loop_im): Likewise.
21755 (tree_ssa_loop_unswitch): Likewise.
21756 (tree_vectorize): Likewise.
21757 (check_data_deps): Likewise.
21758 (tree_ssa_loop_ivcanon): Likewise.
21759 (tree_ssa_loop_bounds): Likewise.
21760 (tree_complete_unroll): Likewise.
21761 (tree_complete_unroll_inner): Likewise.
21762 (tree_parallelize_loops): Likewise.
21763 (tree_ssa_loop_prefetch): Likewise.
21764 (tree_ssa_loop_ivopts): Likewise.
21765 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21766 * tree-vectorizer.c (vectorize_loops): Likewise.
21768 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21770 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21771 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21772 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21773 with EABI_LINK_SPEC.
21775 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21778 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21781 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21784 * lra-constraints.c (process_alt_operands): Discourage a bit more
21785 using memory for pseudos. Print cost dump for alternatives.
21786 Modify cost values for conflicts with early clobbers.
21787 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21789 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21792 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21794 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21796 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21797 from/to S register.
21798 (movdi_aarch64): Support LDR/STR from/to D register.
21800 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21802 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21803 or higher by default.
21805 2013-04-29 Richard Biener <rguenther@suse.de>
21807 PR middle-end/57075
21808 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21809 even if not adding abnormal edges for calls that can make
21812 2013-04-29 Richard Biener <rguenther@suse.de>
21814 PR middle-end/57103
21815 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21817 (move_stmt_r): Remove redundant checking.
21819 2013-04-29 Teresa Johnson <tejohnson@google.com>
21822 * basic-block.h (apply_scale): New function.
21823 (apply_probability): Use apply_scale.
21824 * gimple-streamer-in.c (input_bb): Ditto.
21825 * lto-streamer-in.c (input_cfg): Ditto.
21826 * lto-cgraph.c (merge_profile_summaries): Ditto.
21827 * tree-optimize.c (execute_fixup_cfg): Ditto.
21828 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21829 (copy_edges_for_bb): Ditto.
21830 (copy_cfg_body): Ditto.
21832 2013-04-29 Tom de Vries <tom@codesourcery.com>
21834 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21835 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21836 (tail_merge_optimize): Handle current_loops == NULL.
21838 2013-04-26 Jeff Law <law@redhat.com>
21840 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21841 (simplify_cond_using_ranges): Generalize code to simplify
21842 COND_EXPRs where one argument is a constant and the other
21843 is an SSA_NAME created by an integral type conversion.
21845 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21847 * config/arm/arm.md (store_minmaxsi): Use only when
21848 optimize_insn_for_size_p.
21850 2013-04-29 Christian Bruel <christian.bruel@st.com>
21853 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21855 2013-04-29 Richard Biener <rguenther@suse.de>
21857 PR middle-end/57089
21858 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21859 loop tree make sure to schedule a fixup for the child as well.
21860 (expand_omp_for_generic): Properly add loops.
21861 (expand_omp_for_static_nochunk): Likewise.
21862 (expand_omp_for_static_chunk): Likewise.
21863 (expand_omp_for): For the degenerate case fixup loops.
21864 (expand_omp_sections): Fix default bb placement in loops.
21865 (expand_omp_atomic_pipeline): Properly add loops.
21867 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21869 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21871 2013-04-29 Tom de Vries <tom@codesourcery.com>
21873 * tree-ssa-tail-merge.c: Update header comment.
21875 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21877 * config/aarch64/arm_neon.h
21878 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21879 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21880 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21881 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21882 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21883 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21884 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21885 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21887 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21889 * config/aarch64/aarch64-simd.md
21890 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21891 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21892 fix_trunc, fixuns_trunc.
21893 (ftrunc<VDQF:mode>2): New.
21894 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21895 (fix_trunc_optab): New.
21897 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21899 * config/aarch64/aarch64-builtins.c
21900 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21901 iceilf, lround, iroundf.
21903 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21906 * config/i386/i386.h (enum ix86_tune_indices)
21907 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21908 New, split from X86_TUNE_INTER_UNIT_MOVES.
21909 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21910 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21911 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21912 (TARGET_INTER_UNIT_MOVES): Remove.
21913 * config/i386/i386.c (initial_ix86_tune_features): Update.
21914 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21915 (ix86_expand_convert_uns_didf_sse): Use
21916 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21917 (ix86_expand_vector_init_one_nonzero): Ditto.
21918 (ix86_expand_vector_init_interleave): Ditto.
21919 (inline_secondary_memory_needed): Return true for moves from SSE class
21920 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21921 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21922 * config/i386/constraints.md (Yi, Ym): Depend on
21923 TARGET_INTER_UNIT_MOVES_TO_VEC.
21924 (Yj, Yn): New constraints.
21925 * config/i386/i386.md (*movdi_internal): Change constraints of
21926 operand 1 from Yi to Yj and from Ym to Yn.
21927 (*movsi_internal): Ditto.
21928 (*movdf_internal): Ditto.
21929 (*movsf_internal): Ditto.
21930 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21931 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21932 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21933 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21934 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21935 * config/i386/sse.md (movdi_to_sse): Ditto.
21936 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21937 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21938 TARGET_INTER_UNIT_MOVES.
21939 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21940 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21941 instead of TARGET_INTER_UNIT_MOVES.
21942 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21943 operand 1 from Yi to Yj and from Ym to Yn.
21945 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21947 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21948 (float_truncate_hi_): Likewise.
21949 (float_extend_lo_): Likewise.
21950 (float_truncate_lo_): Likewise.
21951 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21952 (aarch64_float_extend_lo_v2df): Likewise.
21953 (vec_unpacks_hi_v4sf): Likewise.
21954 (aarch64_float_truncate_lo_v2sf): Likewise.
21955 (aarch64_float_truncate_hi_v4sf): Likewise.
21956 (vec_pack_trunc_v2df): Likewise.
21957 (vec_pack_trunc_df): Likewise.
21959 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21961 * config/aarch64/aarch64-builtins.c
21962 (aarch64_fold_builtin): Fold float conversions.
21963 * config/aarch64/aarch64-simd-builtins.def
21964 (floatv2si, floatv4si, floatv2di): New.
21965 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21966 * config/aarch64/aarch64-simd.md
21967 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21968 * config/aarch64/iterators.md (FLOATUORS): New.
21969 (optab): Add float, floatuns.
21970 (su_optab): Likewise.
21972 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21974 * config/aarch64/aarch64-builtins.c
21975 (aarch64_builtin_vectorized_function): Use new names for
21977 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21978 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21979 (fcvtzu): Split as...
21980 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21981 (fcvtas): Split as...
21982 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21983 (fcvtau): Split as...
21984 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21985 (fcvtps): Split as...
21986 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21987 (fcvtpu): Split as...
21988 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21989 (fcvtms): Split as...
21990 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21991 (fcvtmu): Split as...
21992 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21993 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21994 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21995 (lfrintnusf, lfrintnudf): Likewise.
21996 * config/aarch64/aarch64-simd.md
21997 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21999 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
22000 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
22001 (fcvt_pattern): Likewise.
22003 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22005 * config/aarch64/aarch64-simd.md
22006 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
22007 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
22009 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22011 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
22012 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
22013 (vrnd<a,m,n,p>_f32): Implement using builtins.
22014 (vrnd<i,x><q>_f<32, 64>): New.
22016 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22018 * config/aarch64/aarch64-builtins.c
22019 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
22020 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
22021 (frintz): Rename to...
22023 (frintp): Rename to...
22025 (frintm): Rename to...
22027 (frinti): Rename to...
22028 (nearbyint): ...this.
22029 (frintx): Rename to...
22031 (frinta): Rename to...
22033 * config/aarch64/aarch64-simd.md
22034 (aarch64_frint<frint_suffix><mode>): Delete.
22035 (<frint_pattern><mode>2): Convert to insn.
22036 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
22037 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
22038 (frint_pattern): Likewise.
22039 (frint_suffix): Likewise.
22041 2013-04-29 Richard Biener <rguenther@suse.de>
22043 PR tree-optimization/57081
22044 * loop-init.c: Include tree-flow.h.
22045 (loop_optimizer_finalize): Free number of iteration estimates.
22046 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
22048 2013-04-29 Jakub Jelinek <jakub@redhat.com>
22050 PR tree-optimization/57083
22051 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
22052 non-singleton shift count range, zero extend low_bound for uns case.
22054 * config/i386/predicates.md (general_vector_operand): New predicate.
22055 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
22056 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
22057 if they aren't nonimmediate operands. If their original values
22058 satisfy const_vector_equal_evenodd_p, don't shift them.
22059 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
22060 predicates. For the SSE4.1 case force operands[{1,2}] into registers
22061 if not nonimmediate_operand.
22062 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
22063 instead of register_operand.
22064 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
22066 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22068 * stor-layout.c (finalize_size_functions): Allocate a structure and
22069 reset cfun before dumping the functions.
22071 2013-04-27 Jakub Jelinek <jakub@redhat.com>
22073 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
22076 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
22077 use xop_pmacsdqh if uns_p.
22078 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
22079 the immediate rotate count.
22081 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
22083 * rtl.h (struct rtx_def): Add comment for field jump.
22084 (LRA_SUBREG_P): New macro.
22085 * recog.c (register_operand): Check LRA_SUBREG_P.
22086 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22088 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22089 (lra_final_code_change): Skip subreg change for operators.
22090 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22091 if there are no operand changes.
22092 * lra-constraints.c (curr_insn_set): New.
22093 (match_reload): Set LRA_SUBREG_P.
22094 (emit_spill_move): Ditto.
22095 (check_and_process_move): Use curr_insn_set. Process only single
22096 set insns. Don't initialize sec_mem_p and change_p.
22097 (simplify_operand_subreg): Use LRA_SUBREG_P.
22098 (reg_in_class_p): New function.
22099 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22100 of #ifdef. Add code to remove cycling.
22101 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22102 non-null disp. Reload inner instead of disp when base and index
22103 are null. Try to put lo_sum into register.
22104 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22105 (check_and_process_move): Move code for move cost check to
22106 simple_move_p. Remove equiv_substitution.
22107 (simple_move_p): New function.
22108 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22109 curr_insn_set. Call check_and_process_move only for single set
22110 insns. Use the new function. Move call of check_and_process_move
22111 after operand equiv substitution and address process.
22113 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22116 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
22117 with nonlocal goto receivers or returns twice calls, ignore
22118 unininitialized values from abnormal edges to nl goto receiver
22119 or returns twice call.
22121 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22123 PR tree-optimization/57051
22124 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
22125 and VEC_RSHIFT_EXPR if shift count is a multiple of element
22128 2013-04-26 Richard Biener <rguenther@suse.de>
22130 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
22131 (expand_omp_taskreg): Likewise. Mark loops for fixup.
22132 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
22133 (fixup_loop_arrays_after_move): New function.
22134 (move_sese_region_to_fn): Properly outline the loop tree parts
22135 of the SESE region.
22137 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
22139 * config/i386/i386.md (type, unit): Fix long lines.
22141 2013-04-26 Richard Biener <rguenther@suse.de>
22143 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
22144 (lto-streamer-out.o): Likewise.
22145 * cfgloop.c (init_loops_structure): Export, add struct function
22146 argument and adjust.
22147 (flow_loops_find): Adjust.
22148 * cfgloop.h (enum loop_estimation): Add EST_LAST.
22149 (init_loops_structure): Declare.
22150 * lto-streamer-in.c: Include cfgloop.h.
22151 (input_cfg): Input the loop tree.
22152 * lto-streamer-out.c: Include cfgloop.h.
22153 (output_cfg): Output the loop tree.
22154 (output_struct_function_base): Do not drop PROP_loops.
22156 2013-03-26 Richard Biener <rguenther@suse.de>
22158 * tree-cfg.c (execute_build_cfg): Build the loop tree.
22159 (pass_build_cfg): Provide PROP_loops.
22160 (move_sese_region_to_fn): Remove loops that are outlined into fn
22162 * tree-inline.c: Include cfgloop.h.
22163 (initialize_cfun): Do not drop PROP_loops.
22164 (copy_loops): New function.
22165 (copy_cfg_body): Copy loop structure.
22166 (tree_function_versioning): Initialize destination loop tree.
22167 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
22168 (pass_parallelize_loops): Do IL verification.
22169 * loop-init.c (loop_optimizer_init): Fixup loops if required.
22170 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
22171 the CFG make sure we fixup loops as well.
22172 * tree-ssa-tail-merge.c: Include cfgloop.h.
22173 (replace_block_by): When merging loop latches mark loops for fixup.
22174 * lto-streamer-out.c (output_struct_function_base): Drop
22175 PROP_loops for now.
22176 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
22177 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
22178 * ipa-split.c: Include cfgloop.h.
22179 (split_function): Add the new return block to the loop tree root.
22180 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
22181 whether we have removed the forwarder block.
22182 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
22183 * cfgloop.h (place_new_loop): Declare.
22184 * cfgloopmanip.c (place_new_loop): Export.
22185 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
22186 (tree-switch-conversion.o): Likewise.
22187 (tree-complex.o): Likewise.
22188 (tree-inline.o): Likewise.
22189 (tree-ssa-tailmerge.o): Likewise.
22190 (ipa-split.o): Likewise.
22191 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
22192 (tree-ssa-copy.o): Likewise.
22193 * tree-switch-conversion.c: Include cfgloop.h
22194 (process_switch): If we emit a bit-test cascade, schedule loops
22196 * tree-complex.c: Include cfgloop.h.
22197 (expand_complex_div_wide): Properly add new basic-blocks to loops.
22198 * asan.c: Include cfgloop.h.
22199 (create_cond_insert_point): Properly add new basic-blocks to
22200 loops, schedule loop fixup.
22201 * cfgloop.c (verify_loop_structure): Check that looks are not
22203 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
22205 (expand_omp_for_generic): Likewise.
22206 (expand_omp_sections): Likewise.
22207 (expand_omp_atomic_pipeline): Schedule loops for fixup.
22208 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
22209 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
22210 is initialized, not when loops are present.
22211 * tree-parloops.c (parallelize_loops): Remove checking here.
22212 * passes.c (init_optimization_passes): Schedule a copy-propagation
22213 pass before complete unrolling of inner loops.
22215 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22217 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
22218 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
22219 (colorize_init): Add argument to _WIN32 version.
22220 * toplev.c: Include diagnostic-color.h.
22221 (process_options): Default to -fdiagnostics-color=auto if
22222 GCC_COLORS env var is in the environment.
22223 * common.opt (fdiagnostics-color=): Add Var and Init.
22224 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
22225 env var is in the environment, the default is auto rather than never.
22227 * diagnostic.h (file_name_as_prefix): Add context argument.
22228 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
22229 the string as locus.
22230 * langhooks.c (lhd_print_error_function): Adjust caller.
22232 2013-04-25 Lawrence Crowl <crowl@google.com>
22234 * var-tracking.c (shared_hash_def::htab):
22235 Change type to hash_table. Update dependent calls and types.
22237 2013-04-25 Lawrence Crowl <crowl@google.com>
22239 * Makefile.in: Update as needed below.
22241 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
22242 Move declaration to after the type's method definitons.
22244 * attribs.c (htab_t scoped_attributes::attribute_hash):
22245 Change type to hash_table. Update dependent calls and types.
22247 * bitmap.c (htab_t bitmap_desc_hash):
22248 Change type to hash_table. Update dependent calls and types.
22250 * cselib.c (htab_t cselib_hash_table):
22251 Change type to hash_table. Update dependent calls and types.
22253 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
22254 (hash_string_slot_node): Move implementation into lto-streamer.h
22255 struct string_slot_hasher.
22256 (eq_string_slot_node): Likewise.
22258 * data-streamer-out.c: Update output_block::string_hash_table
22259 dependent calls and types.
22261 * dwarf2cfi.c (htab_t trace_index):
22262 Change type to hash_table. Update dependent calls and types.
22264 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
22265 Change type to hash_table. Update dependent calls and types.
22266 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
22267 (htab_t optimize_external_refs::map): Likewise.
22268 (htab_t output_comp_unit::extern_map): Likewise.
22269 (htab_t output_comdat_type_unit::extern_map): Likewise.
22270 (htab_t output_macinfo::macinfo_htab): Likewise.
22271 (htab_t optimize_location_lists::htab): Likewise.
22272 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
22274 * except.c (htab_t ehspec_hash_type):
22275 Change type to hash_table. Update dependent calls and types.
22276 (assign_filter_values::ttypes): Likewise.
22277 (assign_filter_values::ehspec): Likewise.
22278 (sjlj_assign_call_site_values::ar_hash): Likewise.
22279 (convert_to_eh_region_ranges::ar_hash): Likewise.
22281 * gcse.c (htab_t pre_ldst_table):
22282 Change type to hash_table. Update dependent calls and types.
22284 * ggc-common.c (htab_t saving_htab):
22285 Change type to hash_table. Update dependent calls and types.
22286 (htab_t loc_hash): Likewise.
22287 (htab_t ptr_hash): Likewise.
22288 (call_count): Rename ggc_call_count.
22289 (call_alloc): Rename ggc_call_alloc.
22290 (loc_descriptor): Rename make_loc_descriptor.
22291 (add_statistics): Rename ggc_add_statistics.
22293 * ggc-common.c (saving_htab):
22294 Change type to hash_table. Update dependent calls and types.
22296 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
22297 (push_gimplify_context): Likewise.
22298 (pop_gimplify_context): Likewise.
22299 (struct gimple_temp_hash_elt): Added.
22300 (struct gimplify_hasher): Likewise.
22301 (struct gimplify_ctx.temp_htab):
22302 Change type to hash_table. Update dependent calls and types.
22304 * gimple-fold.c: Include gimplify-ctx.h.
22306 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
22307 Change type to hash_table. Update dependent calls and types.
22308 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
22309 avoid potential global name collision.
22311 * gimplify.c: Include gimplify-ctx.h.
22312 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
22313 (htab_t gimplify_ctx::temp_htab):
22314 Update dependent calls and types for new type hash_table.
22315 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
22316 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
22318 * gimplify-ctx.h: New.
22319 (struct gimple_temp_hash_elt): Move from gimplify.c.
22320 (class gimplify_hasher): New.
22321 (struct gimplify_ctx): Move from gimple.h.
22322 (htab_t gimplify_ctx::temp_htab):
22323 Change type to hash_table. Update dependent calls and types.
22325 * graphite-clast-to-gimple.c: Include graphite-htab.h.
22326 (htab_t ivs_params::newivs_index):
22327 Change type to hash_table. Update dependent calls and types.
22328 (htab_t ivs_params::params_index): Likewise.
22329 (htab_t print_generated_program::params_index): Likewise.
22330 (htab_t gloog::newivs_index): Likewise.
22331 (htab_t gloog::params_index): Likewise.
22333 * graphite.c: Include graphite-htab.h.
22334 4htab_t graphite_transform_loops::bb_pbb_mapping):
22335 Change type to hash_table. Update dependent calls and types.
22337 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
22338 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
22339 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
22341 * graphite-dependences.c: Include graphite-htab.h.
22342 (loop_is_parallel_p): Change hash table type of parameter.
22344 * graphite-htab.h: New.
22345 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
22346 (extern find_pbb_via_hash): Move from graphite-poly.h.
22347 (extern loop_is_parallel_p): Move from graphite-poly.h.
22348 (extern get_loop_body_pbbs): Move from graphite-poly.h.
22350 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
22351 (extern loop_is_parallel_p): Move to graphite-htab.h.
22352 (extern get_loop_body_pbbs): Move to graphite-htab.h.
22354 * haifa-sched.c (htab_t delay_htab):
22355 Change type to hash_table. Update dependent calls and types.
22356 (htab_t delay_htab_i2): Likewise.
22358 * ira-color.c (htab_t allocno_hard_regs_htab):
22359 Change type to hash_table. Update dependent calls and types.
22361 * ira-costs.c (htab_t cost_classes_htab):
22362 Change type to hash_table. Update dependent calls and types.
22364 * loop-invariant.c (htab_t merge_identical_invariants::eq):
22365 Change type to hash_table. Update dependent calls and types.
22367 * loop-iv.c (htab_t bivs):
22368 Change type to hash_table. Update dependent calls and types.
22370 * loop-unroll.c (htab_t opt_info::insns_to_split):
22371 Change type to hash_table. Update dependent calls and types.
22372 (htab_t opt_info::insns_with_var_to_expand): Likewise.
22374 * lto-streamer.h (struct string_slot): Move from data-streamer.h
22375 (struct string_slot_hasher): New.
22376 (htab_t output_block::string_hash_table):
22377 Change type to hash_table. Update dependent calls and types.
22379 * lto-streamer-in.c (freeing_string_slot_hasher): New.
22380 (htab_t file_name_hash_table):
22381 Change type to hash_table. Update dependent calls and types.
22383 * lto-streamer-out.c: Update output_block::string_hash_table dependent
22386 * lto-streamer.c (htab_t tree_htab):
22387 Change type to hash_table. Update dependent calls and types.
22389 * omp-low.c: Include gimplify-ctx.h.
22391 * passes.c (htab_t name_to_pass_map):
22392 Change type to hash_table. Update dependent calls and types.
22393 (pass_traverse): Rename to passes_pass_traverse.
22395 * plugin.c (htab_t event_tab):
22396 Change type to hash_table. Update dependent calls and types.
22398 * postreload-gcse.c (htab_t expr_table):
22399 Change type to hash_table. Update dependent calls and types.
22400 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
22402 * sese.c (debug_rename_map_1): Make extern.
22403 (htab_t copy_bb_and_scalar_dependences::rename_map):
22404 Change type to hash_table. Update dependent calls and types.
22406 * sese.h (extern debug_rename_map): Move to .c file.
22408 * store-motion.c (htab_t store_motion_mems_table):
22409 Change type to hash_table. Update dependent calls and types.
22411 * trans-mem.c (htab_t tm_new_mem_hash):
22412 Change type to hash_table. Update dependent calls and types.
22414 * tree-browser.c (htab_t TB_up_ht):
22415 Change type to hash_table. Update dependent calls and types.
22417 * tree-cfg.c (htab_t discriminator_per_locus):
22418 Change type to hash_table. Update dependent calls and types.
22420 * tree-complex.c: Include tree-hasher.h
22421 (htab_t complex_variable_components):
22422 Change type to hash_table. Update dependent calls and types.
22424 * tree-eh.c (htab_t finally_tree):
22425 Change type to hash_table. Update dependent calls and types.
22427 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22428 struct int_tree_hasher.
22429 (extern int_tree_map_eq): Likewise.
22430 (uid_decl_map_hash): Removed.
22431 (extern decl_tree_map_eq): Likewise.
22433 * tree-hasher.h: New.
22434 (struct int_tree_hasher): New.
22435 (typedef int_tree_htab_type): New.
22437 * tree-inline.c: Include gimplify-ctx.h.
22439 * tree-mudflap.c: Include gimplify-ctx.h.
22441 * tree-parloops.c: Include tree-hasher.h.
22442 (htab_t eliminate_local_variables_stmt::decl_address):
22443 Change type to hash_table. Update dependent calls and types.
22444 (htab_t separate_decls_in_region::decl_copies): Likewise.
22446 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22447 Change type to hash_table. Update dependent calls and types.
22449 * tree-sra.c (candidates):
22450 Change type to hash_table. Update dependent calls and types.
22452 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22454 (int_tree_map_hash): Likewise.
22456 * tree-ssa-dom.c (htab_t avail_exprs):
22457 Change type to hash_table. Update dependent calls and types.
22459 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22460 Change type to hash_table. Update dependent calls and types.
22462 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22463 Change type to hash_table. Update dependent calls and types.
22465 * tree-ssa-phiopt.c (seen_ssa_names):
22466 Change type to hash_table. Update dependent calls and types.
22468 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22469 Change type to hash_table. Update dependent calls and types.
22471 * tree-ssa-uncprop.c (equiv):
22472 Change type to hash_table. Update dependent calls and types.
22474 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22476 PR rtl-optimization/57003
22477 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22478 call note_stores with kill_clobbered_value callback again after
22479 killing regs_invalidated_by_call.
22481 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22483 * config/aarch64/aarch64-simd.md
22484 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22485 (aarch64_simd_bsl<mode>): Likewise.
22486 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22488 2013-04-25 Marek Polacek <polacek@redhat.com>
22490 PR tree-optimization/57066
22491 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22493 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22495 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22497 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22499 * config/aarch64/aarch64-builtins.c
22500 (aarch64_fold_builtin): New.
22501 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22502 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22503 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22504 * config/aarch64/arm_neon.h
22505 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22507 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22508 Tejas Belagod <tejas.belagod@arm.com>
22510 * config/aarch64/aarch64-builtins.c
22511 (aarch64_gimple_fold_builtin): New.
22512 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22513 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22514 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22515 (addvv4sf): Update.
22516 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22518 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22520 * config/aarch64/aarch64.md
22521 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22523 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22525 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22526 (*ngcsi_uxtw): New pattern.
22528 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22529 Julian Brown <julian@codesourcery.com>
22531 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22532 (TB_DREG): Add T_V4HF.
22533 (v4hf_UP): New macro.
22534 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22535 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22536 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22538 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22539 (arm_vector_mode_supported_p): Handle V4HF.
22540 (arm_mangle_map): Handle V4HFmode.
22541 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22542 * config/arm/arm_neon_builtins.def: Add entries for
22543 vcvtv4hfv4sf, vcvtv4sfv4hf.
22544 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22545 (neon_vcvtv4hfv4sf): Likewise.
22546 * config/arm/neon-gen.ml: Handle half-precision floating point
22548 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22549 * config/arm/arm_neon.h: Regenerate.
22550 * config/arm/neon.ml (type elts): Add F16.
22551 (type vectype): Add T_float16x4, T_floatHF.
22552 (type vecmode): Add V4HF.
22553 (type features): Add Requires_FP_bit feature.
22554 (elt_width): Handle F16.
22555 (elt_class): Likewise.
22556 (elt_of_class_width): Likewise.
22557 (mode_of_elt): Refactor.
22558 (type_for_elt): Handle F16, fix error messages.
22559 (vectype_size): Handle T_float16x4.
22560 (vcvt_sh): New function.
22561 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
22562 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
22563 (string_of_mode): Handle V4HF.
22564 * doc/arm-neon-intrinsics.texi: Regenerate.
22566 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22568 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
22569 format specifier in 'X' case.
22571 2013-04-25 Alan Modra <amodra@gmail.com>
22574 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
22575 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
22576 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
22577 Repeat for many other rotate/shift and mask patterns using subregs.
22578 Name lshiftrt insns.
22579 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
22580 on WORDS_BIG_ENDIAN.
22582 2013-04-25 Alan Modra <amodra@gmail.com>
22584 * config.gcc: Support little-endian powerpc-linux targets.
22585 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
22586 (LINK_OS_LINUX_SPEC): Define.
22587 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
22588 Preserve MASK_LITTLE_ENDIAN.
22589 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
22590 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
22591 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
22592 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
22593 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
22594 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
22595 Correct fp word order for little-endian. Don't shift toc entries
22596 smaller than a word for little-endian.
22597 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
22598 (bswapdi2 splits): Correct low-part subreg for little-endian.
22599 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
22600 low/high where such is correct only for be.
22601 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
22602 little-endian for -mcall-aixdesc.
22604 2013-04-25 Alan Modra <amodra@gmail.com>
22606 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22607 replace_equiv_address_nv.
22609 2013-04-25 Alan Modra <amodra@gmail.com>
22611 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22613 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22616 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22617 * rtl.h (struct rtx_def): ...
22619 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22621 PR rtl-optimizations/57046
22622 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22623 for multi-reg splits.
22625 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22627 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22629 2013-04-24 Sterling Augustine <saugustine@google.com>
22631 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22632 (comp_dir_string, debug_str_dwo_section): New.
22633 (DEBUG_STR_DWO_SECTION): Rename to ...
22634 (DEBUG_DWO_STR_SECTION): ... this.
22635 (DEBUG_NORM_STR_SECTION): Delete.
22636 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22637 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22638 (find_AT_string): Move most logic to ...
22639 (find_AT_string_in_table): ... here. New.
22640 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22641 add_skeleton_AT_string. Delete logic.
22642 (output_skeleton_debug_sections): Remove call to
22643 add_top_level_skeleton_die_attrs.
22644 (add_comp_dir_attribute): Move logic to comp_dir_string.
22645 (dwarf2out_init): Initialize debug_str_dwo_section.
22646 (output_indirect_string): Call find_string_form.
22647 (output_indirect_strings): Rewrite.
22648 (prune_unused_types): Empty skeleton_debug_str_hash.
22649 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22650 (dwarf2out_finish): Call output_indirect_strings.
22652 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22654 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22656 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22658 * rtl.h (struct rtx_def): Add comment for field jump.
22659 (LRA_SUBREG_P): New macro.
22660 * recog.c (register_operand): Check LRA_SUBREG_P.
22661 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22663 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22664 (lra_final_code_change): Skip subreg change for operators.
22665 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22666 if there are no operand changes.
22667 * lra-constraints.c (curr_insn_set): New.
22668 (match_reload): Set LRA_SUBREG_P.
22669 (emit_spill_move): Ditto.
22670 (check_and_process_move): Use curr_insn_set. Process only single
22671 set insns. Don't initialize sec_mem_p and change_p.
22672 (simplify_operand_subreg): Use LRA_SUBREG_P.
22673 (reg_in_class_p): New function.
22674 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22675 of #ifdef. Add code to remove cycling.
22676 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22677 non-null disp. Reload inner instead of disp when base and index
22678 are null. Try to put lo_sum into register.
22679 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22680 (check_and_process_move): Move code for move cost check to
22681 simple_move_p. Remove equiv_substitution.
22682 (simple_move_p): New function.
22683 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22684 curr_insn_set. Call check_and_process_move only for single set
22685 insns. Use the new function. Move call of check_and_process_move
22686 after operand equiv substitution and address process.
22688 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22690 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22691 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22692 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22694 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22696 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22698 2013-04-24 Marek Polacek <polacek@redhat.com>
22700 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22701 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22702 (select_loops_exit_conditions): Likewise.
22703 (number_of_iterations_for_all_loops): Likewise.
22704 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22705 (scev_analysis): Likewise.
22707 2013-04-02 Catherine Moore <clm@codesourcery.com>
22708 Chao-ying Fu <fu@mips.com>
22710 * config/mips/micromips.md (jraddiusp): New pattern.
22711 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22712 instruction if possible.
22714 2013-04-24 Alan Modra <amodra@gmail.com>
22716 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22718 2013-04-24 Julian Brown <julian@codesourcery.com>
22719 Chung-Lin Tang <cltang@codesourcery.com>
22721 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22722 dependency behavior in enumeration type DIE generation. Add TODO note
22723 to comments about future DW_FORM_sdata/udata re-work of related code.
22725 2013-04-23 Lawrence Crowl <crowl@google.com>
22727 * Makefile.in: Update as needed below.
22729 * hash-table.h (class hash_table):
22730 Correct many methods with parameter types compare_type to the correct
22731 value_type. (Correct code was unlikely to notice the change.)
22732 (hash_table::elements_with_deleted) New.
22733 (class hashtable::iterator): New.
22734 (hashtable::begin()): New.
22735 (hashtable::end()): New.
22736 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22738 * statistics.c (statistics_hashes):
22739 Change type to hash_table. Update dependent calls and types.
22741 * tree-into-ssa.c (var_infos):
22742 Change type to hash_table. Update dependent calls and types.
22744 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22745 Change type to hash_table. Update dependent calls and types.
22747 * tree-ssa-loop-im.c (struct mem_ref.refs):
22748 Change type to hash_table. Update dependent calls and types.
22750 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22751 Change type to hash_table. Update dependent calls and types.
22753 * tree-ssa-sccvn.c (vn_tables_s::nary):
22754 Change type to hash_table. Update dependent calls and types.
22755 (vn_tables_s::phis): Likewise.
22756 (vn_tables_s::references): Likewise.
22758 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22759 (vn_reference_eq): Update parameter and return types.
22761 * tree-ssa-structalias.c (pointer_equiv_class_table):
22762 Change type to hash_table. Update dependent calls and types.
22763 (location_equiv_class_table): Likewise.
22765 * tree-vect-data-refs.c: Consequential changes for making
22766 peeling a hash_table.
22768 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22769 (destroy_loop_vec_info): Dependent hash_table update.
22771 * tree-vectorizer.h (peeling_htab):
22772 Change type to hash_table. Update dependent calls and types.
22774 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22776 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22777 to check the register content is equal or not.
22778 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22779 to assign register content record.
22780 * lra-eliminations.c (update_reg_eliminate): Use
22781 lra_update_reg_val_offset to update register content offset.
22782 * lra-int.h (struct lra_reg): Add offset member.
22783 (lra_reg_val_equal_p): New static inline function.
22784 (lra_update_reg_val_offset): New static inline function.
22785 (lra_assign_reg_val): New static inline function.
22786 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22787 to assign register content record.
22788 (initialize_lra_reg_info_element): Initial offset to zero.
22790 2013-04-23 Catherine Moore <clm@codesourcery.com>
22792 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22793 operands. Record compression.
22795 2013-04-23 Xinliang David Li <davidxl@google.com>
22797 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22799 2013-04-23 Richard Biener <rguenther@suse.de>
22801 PR middle-end/57036
22802 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22803 parameter, only add abnormal goto edges from the copied body
22804 if the call could perform abnormal gotos.
22805 (copy_cfg_body): Adjust.
22807 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22809 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22811 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22813 * coretypes.h (gimple_stmt_iterator): Add struct to make
22816 2013-04-23 Richard Biener <rguenther@suse.de>
22818 PR tree-optimization/57026
22819 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22820 from SSA names occuring in abnormal PHI nodes.
22822 2013-04-22 Andi Kleen <ak@linux.intel.com>
22824 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22826 2013-04-22 Andi Kleen <ak@linux.intel.com>
22828 * lto/lto.c (print_lto_report_1): Declare early.
22829 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22831 2013-04-22 Andi Kleen <ak@linux.intel.com>
22833 * common.opt (-flto-report-wpa): Add.
22834 * doc/invoke.texi (-flto-report-wpa): Add.
22835 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22838 2013-04-22 Xinliang David Li <davidxl@google.com>
22840 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22841 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22842 * Makefile.in: New dependency
22844 David Daney <ddaney.cavm@gmail.com>
22846 * configure.ac (gcc_cv_as_micromips_support): Use the
22847 --fatal-warnings option.
22848 * configure: Regenerate.
22850 2013-04-22 Marek Polacek <polacek@redhat.com>
22853 * tsan.c (instrument_expr): Don't instrument expression
22854 in case its size is zero.
22856 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22860 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22862 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22864 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22866 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22867 (gimple_stmt_iterator): New typedef.
22868 * gimple.h (gimple_stmt_iterator): Rename to...
22869 (gimple_stmt_iterator_d): ... This.
22870 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22871 trees be valid for GIMPLE and GENERIC.
22872 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22873 * gimple-fold.c (gimple_fold_call): Call target hook
22874 gimple_fold_builtin.
22875 * hooks.c (hook_bool_gsiptr_false): New.
22876 * hooks.h (hook_bool_gsiptr_false): New.
22877 * target.def (fold_stmt): New.
22878 * doc/tm.texi: Regenerate.
22880 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22883 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22884 a set sp if no stack realignment.
22886 2013-04-22 Nick Clifton <nickc@redhat.com>
22888 * config.gcc (tilegx-linux): Extend extra_objs rather than
22890 (tilepro-linux): Likewise.
22892 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22894 * config/aarch64/aarch64-builtins.c
22896 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22897 (VAR<1-12>): Add MAP parameter.
22898 (BUILTIN_*): Likewise.
22899 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22900 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22901 (aarch64_ushl_n<mode>): Likewise.
22902 (aarch64_sshr_n<mode>): Likewise.
22903 (aarch64_ushr_n<mode>): Likewise.
22904 (aarch64_<maxmin><mode>): Likewise.
22905 (aarch64_sqrt<mode>): Likewise.
22906 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22907 (vshr<q>_n_*): Likewise.
22909 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22911 * config/aarch64/aarch64-builtins.c
22912 (aarch64_simd_builtin_type_mode): Handle SF types.
22914 (BUILTIN_GPF): Define.
22915 (aarch64_init_simd_builtins): Handle SF types.
22916 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22917 (frecps): Likewise.
22918 (frecpx): Likewise.
22919 * config/aarch64/aarch64-simd.md
22920 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22921 (aarch64_frecpe<mode>): New.
22922 (aarch64_frecps<mode>): Likewise.
22923 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22924 (v8type): Add frecp<esx>.
22925 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22926 (aarch64_frecps<mode>): Likewise.
22927 * config/aarch64/iterators.md (FRECP): New.
22928 (frecp_suffix): Likewise.
22929 * config/aarch64/arm_neon.h
22930 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22932 2013-04-22 Christian Bruel <christian.bruel@st.com>
22935 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22936 (REG_CLASS_NAMES): Idem.
22937 (REG_CLASS_CONTENTS): Idem.
22938 (REGCLASS_HAS_FP_REG): Idem.
22939 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22940 (sh_conditional_register_usage): Idem.
22942 2013-04-21 Jeff Law <law@redhat.com>
22944 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22945 (ssa_forward_propagate_and_combine): Use it.
22947 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22949 * lra.c: Update the flow chart diagram.
22951 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22953 PR rtl-optimization/56847
22954 * lra-constraints.c (process_alt_operands): Discourage alternative
22955 with non-matche doffsettable memory constraint fro memory with
22958 2013-04-19 Richard Biener <rguenther@suse.de>
22960 PR tree-optimization/56982
22961 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22963 * gimplify.c (gimplify_call_expr): Notice special calls.
22964 (gimplify_modify_expr): Likewise.
22965 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22966 abnormal control flow receivers.
22967 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22968 in the same way as cfun->has_nonlocal_labels.
22969 (gimple_purge_dead_abnormal_call_edges): Likewise.
22970 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22971 receivers start a basic-block.
22973 2013-04-19 Richard Biener <rguenther@suse.de>
22975 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22977 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22978 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22979 (SLP_TREE_LOAD_PERMUTATION): Add.
22980 (vect_transform_slp_perm_load): Adjust prototype.
22981 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22982 (vect_free_slp_instance): Likewise.
22983 (vect_create_new_slp_node): Likewise.
22984 (vect_supported_slp_permutation_p): Remove.
22985 (vect_slp_rearrange_stmts): Adjust.
22986 (vect_supported_load_permutation_p): Likewise. Inline
22987 vect_supported_slp_permutation_p here.
22988 (vect_analyze_slp_instance): Compute load permutations per
22989 slp node instead of per instance.
22990 (vect_get_slp_defs): Adjust.
22991 (vect_transform_slp_perm_load): Likewise.
22992 (vect_schedule_slp_instance): Remove redundant code.
22993 (vect_schedule_slp): Remove hack for PR56270, add it ...
22994 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22995 CSE loads for SLP. Adjust.
22997 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22999 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
23000 spelling in two comments.
23002 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
23005 * config/arm/arm.c (load_multiple_sequence): Require SP
23006 as base register for loads if SP is in the register list.
23008 2013-04-19 Martin Jambor <mjambor@suse.cz>
23010 PR tree-optimization/56718
23011 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
23012 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
23013 and made public. Adjusted all callers.
23014 (ipa_intraprocedural_devirtualization): New function.
23015 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
23016 (ipa_intraprocedural_devirtualization): Likewise.
23017 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
23019 2013-04-19 Richard Biener <rguenther@suse.de>
23021 PR tree-optimization/57000
23022 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
23024 2013-04-19 Terry Guo <terry.guo@arm.com>
23026 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
23028 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
23029 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
23030 (cortex_m4_fmacs): Use new reservations.
23031 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
23033 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
23035 PR rtl-optimization/56999
23036 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
23038 (lra_coalesce): Remove split_origin_bitmap and related code.
23039 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
23040 ranges if necessary.
23042 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
23044 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
23046 (ix86_expand_call): Remove clobbered_registers array and use
23047 x86_64_ms_sysv_extra_clobbered_registers instead.
23048 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
23050 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
23052 * config/i386/i386.md (*call_rex64_ms_sysv): Use
23053 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
23054 (*call_value_rex64_ms_sysv): Ditto.
23056 2013-04-18 Cary Coutant <ccoutant@google.com>
23058 * dwarf2out.c (output_pubnames): Check die_perennial_p of
23059 parent instead of die_mark.
23061 2013-04-18 Diego Novillo <dnovillo@google.com>
23063 * gimple.c (create_gimple_tmp): New.
23064 (get_expr_type): New.
23065 (build_assign): New.
23066 (build_type_cast): New.
23067 * gimple.h (enum ssa_mode): Define.
23068 (gimple_seq_set_location): New.
23069 * asan.c (build_check_stmt): Change some gimple_build_* calls
23070 to use build_assign and build_type_cast.
23072 2013-04-18 Richard Biener <rguenther@suse.de>
23074 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
23075 handle negative step. Remove redundant checks.
23076 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
23077 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
23078 for negative step and grouped loads fail to vectorize.
23080 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23082 * emit-rtl.c (reset_insn_used_flags): New function.
23083 (reset_all_used_flags): Use it.
23084 (verify_insn_sharing): New function.
23085 (verify_rtl_sharing): Fix verification for SEQUENCEs.
23087 2013-04-18 Jakub Jelinek <jakub@redhat.com>
23089 PR tree-optimization/56984
23090 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
23091 and (x >> M) >= N don't register any assertion if N << M is the
23094 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23096 * lower-subreg.c (resolve_simple_move): If called self-recursive,
23097 do not delete_insn insns that have not yet been emitted, only
23098 unlink them with remove_insn.
23099 * df-scan.c (df_insn_delete): Revert r197492.
23101 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
23103 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
23104 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
23106 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23108 * config/arm/arm.md (movsicc_insn): Convert define_insn into
23109 define_insn_and_split.
23110 (and_scc,ior_scc,negscc): Likewise.
23111 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
23113 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23115 * config/arm/arm.c (use_return_insn): Return 0 for targets that
23116 can benefit from using a sequence of LDRD instructions in epilogue
23117 instead of a single LDM instruction.
23119 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
23122 * doc/extend.texi: Fix typo.
23124 2013-04-17 Richard Biener <rguenther@suse.de>
23126 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
23127 (vect_build_slp_tree): ... here.
23128 (vect_build_slp_tree_1): Compute which stmts of the SLP group
23129 match. Remove special-casing of mismatched complex loads.
23130 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
23131 re-try the match with swapped commutative operands.
23132 (vect_supported_load_permutation_p): Remove special-casing of
23133 mismatched complex loads.
23134 (vect_analyze_slp_instance): Adjust.
23136 2013-04-17 Richard Biener <rguenther@suse.de>
23138 PR rtl-optimization/56921
23139 * cfgloop.h (struct loop): Add simple_loop_desc member.
23140 (struct niter_desc): Mark with GTY(()).
23141 (simple_loop_desc): Do not use aux field but simple_loop_desc.
23142 * loop-iv.c (get_simple_loop_desc): Likewise.
23143 (free_simple_loop_desc): Likewise.
23146 2013-04-16 Richard Biener <rguenther@suse.de>
23148 PR rtl-optimization/56921
23149 * loop-init.c (pass_rtl_move_loop_invariants): Add
23150 TODO_do_not_ggc_collect to todo_flags_finish.
23151 (pass_rtl_unswitch): Same.
23152 (pass_rtl_unroll_and_peel_loops): Same.
23153 (pass_rtl_doloop): Same.
23155 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
23157 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
23158 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
23159 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
23161 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
23162 * tree-streamer.c (record_common_node): Adjust reference in comment.
23164 2013-04-17 Terry Guo <terry.guo@arm.com>
23166 * config/arm/cortex-m4.md: Add a new bypass.
23168 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23170 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
23172 (*subs_<optab><mode>_multp2): New pattern.
23173 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23174 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23176 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23178 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
23179 (*subs_mul_imm_<mode>): New pattern.
23181 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
23184 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
23185 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
23186 (vsx_movti_32bit): Change j->wa to O->wa.
23188 2013-04-16 Richard Biener <rguenther@suse.de>
23190 PR rtl-optimization/56921
23191 * loop-init.c (pass_rtl_move_loop_invariants): Add
23192 TODO_do_not_ggc_collect to todo_flags_finish.
23193 (pass_rtl_unswitch): Same.
23194 (pass_rtl_unroll_and_peel_loops): Same.
23195 (pass_rtl_doloop): Same.
23197 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
23199 * config/arm/arm.c (emit_multi_reg_push): New declaration
23200 for an existing function.
23201 (arm_emit_strd_push): New function.
23202 (arm_expand_prologue): Used here.
23203 (arm_emit_ldrd_pop): New function.
23204 (arm_expand_epilogue): Used here.
23205 (arm_get_frame_offsets): Update condition.
23206 (arm_emit_multi_reg_pop): Add a special case for load of a single
23207 register with writeback.
23209 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
23211 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
23214 2013-04-16 Richard Biener <rguenther@suse.de>
23216 PR tree-optimization/56756
23217 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
23218 (first_mem_ref_loc): New.
23219 (execute_sm): Place the load temporarily before a previous
23220 access instead of in the latch edge to ensure its SSA dependencies
23221 are defined at points dominating the load.
23223 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
23225 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
23226 correct fix by moving header and footer insn to the footer of
23227 the merged basic block. Clear BB_END of the merged-away block.
23229 PR middle-end/43631
23230 * emit-rtl.c (make_note_raw): New function.
23231 (link_insn_into_chain): New static inline function.
23232 (add_insn): Use it.
23233 (add_insn_before, add_insn_after): Factor insn chain linking code...
23234 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
23235 using link_insn_into_chain.
23236 (note_outside_basic_block_p): New helper function for emit_note_after
23237 and emit_note_before.
23238 (emit_note_after): Use nobb variant of add_insn_after if the note
23239 should not be contained in a basic block.
23240 (emit_note_before): Use nobb variant of add_insn_before if the note
23241 should not be contained in a basic block.
23242 (emit_note_copy): Use make_note_raw.
23243 (emit_note): Likewise.
23244 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
23245 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
23246 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
23247 the moved barrier the tail of the basic block it follows.
23248 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
23250 2013-04-15 Jakub Jelinek <jakub@redhat.com>
23252 PR tree-optimization/56962
23253 * gimple-ssa-strength-reduction.c (record_increment): Only set
23254 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
23255 either rhs1 or rhs2 is equal to c->base_expr.
23257 2013-04-15 Richard Biener <rguenther@suse.de>
23259 PR tree-optimization/56933
23260 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
23262 (GROUP_READ_WRITE_DEPENDENCE): Remove.
23263 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
23264 * tree-vect-data-refs.c (vect_analyze_group_access): Move
23265 dependence check ...
23266 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
23268 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
23269 GROUP_READ_WRITE_DEPENDENCE.
23271 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23273 * emit-rtl.c (reset_all_used_flags): New function.
23274 (verify_rtl_sharing): Call reset_all_used_flags before and after
23275 performing the checks.
23277 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23279 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
23280 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
23281 * config/arm/constraints.md (De): New constraint.
23282 * config/arm/neon.md (anddi3_neon): Delete.
23283 (neon_vand<mode>): Expand to standard anddi3 pattern.
23284 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
23285 Move earlier in the file.
23286 (neon_inv_logic_op2): Likewise.
23287 (arm_anddi_operand_neon): New predicate.
23289 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23291 * configure.ac (gcc_cv_ld_as_needed): Set
23292 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
23293 Use -z ignore, -z record on *-*-solaris2*.
23294 (HAVE_LD_AS_NEEDED): Update comment.
23295 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
23296 * configure: Regenerate.
23297 * config.in: Regenerate.
23298 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
23299 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
23300 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
23301 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
23302 equivalents. Fix markup.
23303 * doc/tm.texi: Regenerate.
23305 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
23307 * config/i386/i386.opt: New option mstack-protector-guard=.
23308 * config/i386/i386-opts.h: Add enum stack_protector_guard.
23309 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
23310 TARGET_SSP_TLS_GUARD.
23311 * config/i386/i386.c (ix86_option_override_internal): Set
23312 ix86_stack_protector_guard.
23313 * config/i386/i386.md (stack_protect_set): Enable for
23314 TARGET_SSP_TLS_GUARD only.
23315 (stack_protect_set_<mode>): Ditto.
23316 (stack_protect_test): Ditto.
23317 (stack_protect_test_<mode>): Ditto.
23318 * doc/invoke.texi (i386 Option): Document.
23320 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
23323 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
23324 (S_MODES): Set H_MODE bit.
23325 (SF_MODES): Set only S_MODE and SF_MODE bits.
23326 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
23327 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
23328 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
23329 <MODE_FLOAT>: Likewise.
23331 2013-04-15 Joey Ye <joey.ye@arm.com>
23333 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
23335 2013-04-15 Joey Ye <joey.ye@arm.com>
23337 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
23339 (thumb_far_jump_used_p): Count instruction size and set
23342 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
23344 * reorg.c (fill_simple_delay_slots): Reindent block of code.
23345 * resource.c (mark_target_live_regs): Reformat conditional block.
23347 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23349 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
23350 notes, they are emitted only just before final.
23351 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
23353 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23355 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
23356 * cfgrtl.c (delete_insn): Call it here instead.
23357 * lra-spills.c (lra_final_code_change): Use delete_insn.
23358 * haifa-sched.c (sched_remove_insn): Likewise.
23359 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
23360 returning to the nop pool.
23361 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
23362 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
23364 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23366 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
23367 * doc/tm.texi: Regenerated.
23369 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
23371 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
23374 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23376 * df-core.c (df_find_def): Compare register numbers.
23377 (df_find_use): Likewise.
23379 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
23382 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
23383 lra_in_progress for return.
23385 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23387 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
23388 define_insn into define_insn_and_split and emit movsicc patterns.
23390 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23392 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
23394 2013-04-12 Richard Biener <rguenther@suse.de>
23396 * tree-pass.h (TODO_do_not_ggc_collect): New.
23397 * passes.c (execute_one_ipa_transform_pass): Honor
23398 TODO_do_not_ggc_collect.
23399 (execute_one_pass): Likewise.
23402 2013-04-10 Richard Biener <rguenther@suse.de>
23404 * passes.c (init_optimization_passes): Remove reload pass.
23405 * ira.c (do_reload): Merge into ...
23407 (rest_of_handle_reload): Remove.
23408 (pass_reload): Likewise.
23409 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23410 of reload for vzeroupper pass placement.
23412 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23414 PR tree-optimization/56918
23415 PR tree-optimization/56920
23416 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23417 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23418 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23419 use wide_mul_with_sign method.
23421 2013-04-12 Richard Biener <rguenther@suse.de>
23423 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23424 not be considered a gimple constant.
23426 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23428 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23429 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23432 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23433 Jakub Jelinek <jakub@redhat.com>
23435 * opts.c: Include diagnostic-color.h.
23436 (common_handle_option): Handle OPT_fdiagnostics_color_.
23437 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23438 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23439 (diagnostic-color.o): New.
23440 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23441 (diagnostic_color_rule): New enum.
23442 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23443 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23444 the location string.
23445 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23446 either NULL, or color kind.
23447 * diagnostic-color.c: New file.
23448 * diagnostic-color.h: New file.
23449 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23451 * doc/invoke.texi (-fdiagnostics-color): Document.
23452 * pretty-print.h (pp_show_color): Define.
23453 (struct pretty_print_info): Add show_color field.
23454 * diagnostic.c: Include diagnostic-color.h.
23455 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23456 macros. Colorize error:, warning: etc. strings and also the location
23458 (diagnostic_show_locus): Colorize the caret line.
23459 * pretty-print.c: Include diagnostic-color.h.
23460 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23461 inside of %< %> quotes or quoted through q format modifier.
23463 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23465 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23467 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23469 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23470 code in CC_NZ mode.
23471 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23474 2013-04-11 Marek Polacek <polacek@redhat.com>
23476 PR tree-optimization/48184
23477 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23479 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23481 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23482 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23483 (skip_simple_arithmetic): Tidy up.
23484 * tree.h (skip_simple_constant_arithmetic): Declare.
23486 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23488 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23490 2013-04-11 Richard Biener <rguenther@suse.de>
23492 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23493 generate vector constants.
23495 2013-04-11 Richard Biener <rguenther@suse.de>
23497 PR tree-optimization/56878
23498 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23499 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23501 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23502 Prefer to align the DR with the most invariant base address.
23504 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23506 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23509 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23511 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23512 floating-point vector comparisons against 0.
23514 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23516 PR tree-optimization/56899
23517 * fold-const.c (extract_muldiv_1): Apply distributive law
23518 only if TYPE_OVERFLOW_WRAPS (ctype).
23520 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23523 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23524 loading has side effect.
23526 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23528 * config/sparc/sparc.c: Include tree-pass.h.
23529 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23530 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23531 head of file. Change return type. Split off gate function.
23532 (sparc_gate_work_around_errata): New function.
23533 (pass_work_around_errata): New pass definition.
23534 (insert_pass_work_around_errata) New pass insert definition to
23535 insert pass_work_around_errata just after delayed-branch scheduling.
23536 (sparc_option_override): Insert the pass.
23537 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23539 2013-04-10 David S. Miller <davem@davemloft.net>
23541 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23542 or -mcpu=hypersparc.
23544 * target.def (cstore_mode): New hook.
23545 * target.h: Include insn-codes.h
23546 * targhooks.c: Likewise.
23547 (default_cstore_mode): New function.
23548 * targhooks.h: Declare it.
23549 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23550 * doc/tm.texi: Rebuild.
23551 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23552 target hook, rather than inspecting the insn_data.
23553 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23554 (TARGET_CSTORE_MODE): Redefine.
23555 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
23557 * config/sparc/predicates.md (cstore_result_operand): New special
23559 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
23560 Use it for operand 0.
23561 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
23562 (*snesi_special): Likewise.
23563 (*snesi_zero): Likewise.
23564 (*seqsi_zero): Likewise.
23565 (*sltu_insn): Likewise.
23566 (*sgeu_insn): Likewise.
23567 (*seqdi_special): Make operand 0 and comparison operation be of
23569 (*snedi_special): Likewise.
23570 (*snedi_special_vis3): Likewise.
23571 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
23572 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
23573 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
23574 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
23575 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
23576 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
23577 (*sltu_extend_sp64): Likewise.
23578 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
23579 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
23580 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
23581 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
23582 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
23584 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
23586 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
23587 (aarch64_start_file): Use the new function.
23589 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23590 Jason Merrill <jason@redhat.com>
23592 * common.opt: Add -gdwarf.
23593 * opts.c (common_handle_option): Handle it.
23594 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
23596 2013-04-10 Richard Biener <rguenther@suse.de>
23598 * passes.c (execute_todo): Do not call ggc_collect conditional here.
23599 (execute_one_ipa_transform_pass): But unconditionally here.
23600 (execute_one_pass): And here.
23601 (init_optimization_passes): Remove reload pass.
23602 * tree-pass.h (TODO_ggc_collect): Remove.
23603 (pass_reload): Likewise.
23604 * ira.c (do_reload): Merge into ...
23606 (rest_of_handle_reload): Remove.
23607 (pass_reload): Likewise.
23608 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23609 of reload for vzeroupper pass placement.
23610 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23611 and todo_flags_finish of all passes.
23613 2013-04-10 Richard Biener <rguenther@suse.de>
23615 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23616 first_const_oprnd field, rename first_def_type to first_op_type.
23617 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23618 (vect_get_and_check_slp_defs): Always use the type of the
23619 operand. Allow mixed vect_external_def, vect_constant_def types.
23620 (vect_get_constant_vectors): Handle mixed vect_external_def,
23621 vect_constant_def types.
23623 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23625 PR tree-optimization/55524
23626 * tree-ssa-math-opts.c
23627 (convert_mult_to_fma): Don't use an fms construct
23628 when we don't have an fms operation, but fnma, and it looks
23629 likely that we'll be able to use the latter.
23631 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23633 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23635 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23636 inline fail caused by overwritable functions.
23638 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23640 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23641 unnecessary bits in the constant power of two case.
23643 2013-04-10 Richard Biener <rguenther@suse.de>
23645 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23646 broken code swapping operands.
23647 (vect_build_slp_tree): Do not compute load permutations here.
23648 (vect_analyze_slp_instance): Compute load permutations here,
23649 after building the SLP tree.
23651 2013-04-09 Christian Bruel <christian.bruel@st.com>
23653 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23654 of next/prev_real_insn.
23656 2013-04-09 Jan Hubicka <jh@suse.cz>
23658 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23659 Drop aliased parameter.
23660 (function_and_variable_visibility): Do not handle alias pairs.
23661 * cgraph.c (varpool_externally_visible_p): Update prototype.
23662 * varpool.c (varpool_add_new_variable): Update.
23664 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23666 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23668 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23670 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23672 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23674 2013-04-09 Marek Polacek <polacek@redhat.com>
23676 PR tree-optimization/48762
23677 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23679 2013-04-09 Richard Biener <rguenther@suse.de>
23681 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23683 (vect_build_slp_tree): Likewise.
23684 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23685 calculating the cost of a SLP instance.
23686 (vect_analyze_slp_instance): Use it from here, after building
23689 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23691 PR middle-end/56883
23692 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23693 expand_omp_for_static_chunk): Use simple_p = true in
23694 force_gimple_operand_gsi calls when assigning to addressable decls.
23696 2013-04-09 Jeff Law <law@redhat.com>
23698 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23699 when the boolean was created by converting a wider object which
23700 had a boolean range.
23702 2013-04-09 Richard Biener <rguenther@suse.de>
23704 * tree-vectorizer.h (slp_void_p): Remove.
23705 (slp_tree): Typedef before _slp_tree declaration.
23706 (struct _slp_tree): Use a vector of slp_tree as children.
23707 (vect_get_place_in_interleaving_chain): Remove.
23708 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23710 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23712 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23713 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23714 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23715 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23716 Use slp_node instead of slp_void_p and adjust.
23718 2013-04-09 Richard Biener <rguenther@suse.de>
23720 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23721 work that is not necessary.
23723 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23725 PR tree-optimization/56854
23726 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23727 forward into clobber stmts if it would change MEM_REF lhs into
23730 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23732 * tree.c (type_hash_lookup, type_hash_add): Make static.
23733 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23735 2013-04-09 Richard Biener <rguenther@suse.de>
23737 * tree.h (unsave_expr_now): Remove.
23738 * tree-inline.c (mark_local_for_remap_r): Remove.
23739 (unsave_expr_1): Likewise.
23740 (unsave_r): Likewise.
23741 (unsave_expr_now): Likewise.
23742 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23743 (propagate_tree_value): Likewise.
23745 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23747 * doc/rtl.texi (sequence): Rewrite documentation to match the
23748 current use of SEQUENCE rtl objects.
23749 * rtl.def (SEQUENCE): Likewise.
23751 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23752 Update documentation.
23753 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23754 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23756 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23758 2013-04-08 Teresa Johnson <tejohnson@google.com>
23760 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23761 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23763 (estimate_edge_size_and_time): Add comment to suggest using rounding
23765 (estimate_node_size_and_time): Ditto.
23766 (remap_edge_change_prob): Use helper rounding divide methods.
23767 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23768 (gimple_mod_pow2_value_transform): Ditto.
23769 (gimple_mod_subtract_transform): Ditto.
23770 (gimple_ic_transform): Ditto.
23771 (gimple_stringops_transform): Ditto.
23772 * stmt.c (conditional_probability): Ditto.
23773 (emit_case_dispatch_table): Ditto.
23774 * lto-cgraph.c (merge_profile_summaries): Ditto.
23775 * tree-optimize.c (execute_fixup_cfg): Ditto.
23776 * cfgcleanup.c (try_forward_edges): Ditto.
23777 * cfgloopmanip.c (scale_loop_profile): Ditto.
23779 (duplicate_loop_to_header_edge): Ditto.
23780 (lv_adjust_loop_entry_edge): Ditto.
23781 * tree-vect-loop.c (vect_transform_loop): Ditto.
23782 * profile.c (compute_branch_probabilities): Ditto.
23783 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23784 * lto-streamer-in.c (input_cfg): Ditto.
23785 * gimple-streamer-in.c (input_bb): Ditto.
23786 * ipa-cp.c (update_profiling_info): Ditto.
23787 (update_specialized_profile): Ditto.
23788 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23789 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23791 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23792 (compute_trg_info): Ditto.
23793 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23794 (purge_dead_edges): Ditto.
23795 * loop-unswitch.c (unswitch_loop): Ditto.
23796 * cgraphclones.c (cgraph_clone_edge): Ditto.
23797 (cgraph_clone_node): Ditto.
23798 * tree-inline.c (copy_bb): Ditto.
23799 (copy_edges_for_bb): Ditto.
23800 (initialize_cfun): Ditto.
23801 (copy_cfg_body): Ditto.
23802 (expand_call_inline): Ditto.
23804 2013-04-08 Kai Tietz <ktietz@redhat.com>
23806 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23807 TARGET_CYGWIN64 by TARGET_64BIT.
23809 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23811 * config/epiphany/epiphany.md (GPR_1): New constant.
23812 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23813 * config/epiphany/epiphany.c (gen_compare_reg):
23814 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23815 is already in place.
23816 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23817 Don't require being called during rtl expansion; If y operlaps r0,
23819 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23820 (epiphany_expand_epilogue): Likewise.
23822 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23823 Don't use CC_FPmode for ORDERED / UNORDERED.
23824 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23826 * config/epiphany/constraints.md (CnL): New constraint.
23827 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23828 * config/epiphany/predicates.md (add_operand): Allow 1024.
23830 * config/epiphany/epiphany.md (logical_op): New code iterator.
23831 (op_mnc): New code attribute.
23832 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23833 (mov_f+1, mov_f+2): New peephole2 patterns.
23835 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23836 (cstoresi4): Also allow re-use of zero result when doing a NE
23837 comparison to a non-zero operand.
23838 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23840 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23841 Use gen_addsi3_i / gen_subsi3_i.
23843 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23847 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23848 contain anything but clobbers, at most one __builtin_stack_restore,
23849 optionally debug stmts and final resx, and if it has at least one
23850 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23851 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23852 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23853 which isn't defaut definition, remove them.
23854 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23855 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23856 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23857 with MEM_REF LHS with SSA_NAME address.
23859 2013-04-08 Jeff Law <law@redhat.com>
23861 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23863 2013-04-08 Richard Biener <rguenther@suse.de>
23865 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23867 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23868 determined vector type.
23869 (vect_analyze_data_refs): Likewise.
23870 (vect_get_new_vect_var): Adjust.
23871 (vect_create_destination_var): Preserve SSA name versions.
23872 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23873 not dump anything here.
23875 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23877 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23878 Add member lr_slot_known.
23879 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23881 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23882 Remove code that sets lr_slot_offset according to what a previous
23883 version of epiphany_emit_save_restore used to do.
23884 (epiphany_emit_save_restore): When doing an lr save or restore,
23885 set/verify lr_slot_known and lr_slot_offset.
23887 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23890 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23893 2013-04-08 Richard Biener <rguenther@suse.de>
23895 * alias.c (find_base_term): Fix thinko in previous change.
23897 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23899 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23900 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23901 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23902 if possible to compute val.
23903 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23904 For QImode integers don't require anything about precision. Use
23905 const_with_all_bytes_same to find out if the constant doesn't have
23906 repeated bytes in it.
23908 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23910 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23913 2013-04-08 Marek Polacek <polacek@redhat.com>
23915 PR rtl-optimization/48182
23916 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23919 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23922 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23923 nuses, make sure we have a label.
23925 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23928 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23929 (rs6000_emit_swdiv_low_precision): Remove.
23930 (rs6000_emit_swdiv): Rewrite to handle between one and four
23931 iterations of Newton-Raphson generally; modify required number of
23932 iterations for some cases.
23933 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23935 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23937 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23938 set-but-unused variable.
23940 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23941 basic blocks of released function bodies garbage-collectable.
23943 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23944 (struct rtl_opt_pass): Add TODO_df_finish.
23946 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23948 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23950 * config/arm/constraints.md (q): New constraint.
23951 * config/arm/ldrdstrd.md: New file.
23952 * config/arm/arm.md (ldrdstrd.md) New include.
23953 (arm_movdi): Use "q" instead of "r" constraint
23954 for double-word memory access.
23955 (movdf_soft_insn): Likewise.
23956 * config/arm/vfp.md (movdi_vfp): Likewise.
23957 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23958 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23959 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23960 (mem_ok_for_ldrd_strd): Likewise.
23961 (output_move_double): Update assertion.
23963 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23965 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23967 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23969 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23970 define_insn_and_split.
23971 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23973 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23975 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23976 define_insn_and_split.
23977 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23978 (shiftsi3_compare): New pattern.
23979 (rrx): New pattern.
23980 * config/arm/unspecs.md (UNSPEC_RRX): New.
23982 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23984 * config/arm/arm.md (negdi_extendsidi): New pattern.
23985 (negdi_zero_extendsidi): Likewise.
23987 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23989 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23990 define_insn_and_split.
23991 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23992 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23994 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23996 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23997 define_insn_and_split.
23998 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23999 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
24001 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24003 * config/arm/arm.md (subsi3_carryin): New pattern.
24004 (subsi3_carryin_const): Likewise.
24005 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
24006 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
24008 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24010 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
24012 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24014 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
24015 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
24017 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24019 * config/arm/arm.c (arm_expand_builtin): Change fcode
24020 type to unsigned int.
24022 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24024 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
24026 2013-04-04 Ian Lance Taylor <iant@google.com>
24028 * doc/standards.texi (Standards): The Go frontend supports the Go 1
24031 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
24033 PR middle-end/56729
24034 * df-scan.c (df_insn_delete): Disable failing assert.
24036 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24038 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
24039 New function prototype.
24040 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
24041 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
24042 (arm_builtin_vectorized_function): New function.
24044 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24046 * config/arm/arm_neon_builtins.def: New file.
24047 * config/arm/arm.c (neon_builtin_data): Move contents to
24048 arm_neon_builtins.def.
24049 (enum arm_builtins): Include neon builtin definitions.
24050 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
24051 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
24053 2013-04-04 Marek Polacek <polacek@redhat.com>
24055 PR tree-optimization/48186
24056 * predict.c (maybe_hot_frequency_p): Return false if
24057 HOT_BB_FREQUENCY_FRACTION is 0.
24058 (cgraph_maybe_hot_edge_p): Likewise.
24060 2013-04-04 Richard Biener <rguenther@suse.de>
24062 PR tree-optimization/56826
24063 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
24066 2013-04-04 Richard Biener <rguenther@suse.de>
24068 PR tree-optimization/56213
24069 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
24070 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
24072 2013-04-04 Richard Biener <rguenther@suse.de>
24074 PR tree-optimization/56837
24075 * tree-loop-distribution.c (classify_partition): For non-zero
24076 values require that the value has the same precision as its
24077 mode to be useful as memset value.
24079 2013-04-03 Nick Clifton <nickc@redhat.com>
24081 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
24082 (fmssf4): Use fmsf.s on E3V5 architectures.
24083 (fnmasf4): Use fnmaf.s on E3V5 architectures.
24084 (fnmssf4): Use fnmsf.s on E3V5 architectures.
24086 2013-04-03 Jeff Law <law@redhat.com>
24088 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
24089 (lra-eliminations.o): Likewise.
24091 2013-04-03 Teresa Johnson <tejohnson@google.com>
24093 * gcov-io.c (compute_working_sets): Moved most of body of old
24094 compute_working_sets here from profile.c.
24095 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
24096 (gcov_working_set_t): Moved typedef here from basic-block.h
24097 (compute_working_set): Declare.
24098 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
24099 (get_working_sets): Renamed from compute_working_set,
24100 replace most of body with call to new compute_working_sets.
24101 (get_exec_counts): Replace call to compute_working_sets
24102 to get_working_sets.
24103 * profile.h (get_working_sets): Renamed from compute_working_set.
24104 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
24105 to get_working_sets.
24106 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
24107 * gcov-dump.c (dump_working_sets): New function.
24109 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
24111 * hwint.c (sext_hwi, zext_hwi): New functions.
24112 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
24113 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
24114 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
24115 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
24116 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
24117 (sext_hwi, zext_hwi): New functions.
24119 2013-04-03 Jeff Law <law@redhat.com>
24121 PR tree-optimization/56799
24122 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
24123 back test for widening conversion erroneously dropped in prior change.
24125 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24128 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
24129 instead of next_real_insn.
24131 2013-04-03 Marek Polacek <polacek@redhat.com>
24134 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
24136 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24139 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
24141 (thumb1_output_casesi): Likewise.
24142 (thumb2_output_casesi): Likewise.
24144 2013-04-03 Richard Biener <rguenther@suse.de>
24146 PR tree-optimization/56817
24147 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
24149 (tree_unroll_loops_completely_1): ... new function to manually
24150 walk the loop tree, properly defering outer loops of unrolled
24151 loops to later iterations.
24153 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24155 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
24156 (vectorizable_load): Likewise.
24157 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
24158 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
24160 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24162 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
24165 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
24167 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
24169 2013-04-03 Bin Cheng <bin.cheng@arm.com>
24171 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
24173 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24175 PR tree-optimization/56790
24176 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
24179 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24181 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
24183 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
24184 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
24187 2013-04-03 Jakub Jelinek <jakub@redhat.com>
24190 * tree.h (force_folding_builtin_constant_p): New decl.
24191 * builtins.c (force_folding_builtin_constant_p): New variable.
24192 (fold_builtin_constant_p): Fold immediately also if
24193 force_folding_builtin_constant_p.
24195 2013-04-03 Richard Biener <rguenther@suse.de>
24197 PR tree-optimization/56812
24198 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
24199 DRs of the same interleaving chain are independent.
24201 2013-04-02 Jason Merrill <jason@redhat.com>
24203 * gdbinit.in (pbb): Use debug fn.
24205 2013-04-02 Lawrence Crowl <crowl@google.com>
24207 * sese.h (struct ivtype_map_elt_s): Remove unused.
24208 (extern debug_ivtype_map): Remove unused.
24209 (extern eq_ivtype_map_elts): Remove unused.
24210 * sese.c (debug_ivtype_map): Removed unused.
24211 (debug_ivtype_map_1): Removed unused.
24212 (debug_ivtype_elt): Remove unused.
24213 (eq_ivtype_map_elts): Remove unused.
24215 2013-04-02 Kai Tietz <ktietz@redhat.com>
24218 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
24219 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
24220 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
24222 (legitimize_pe_coff_symbol): Likewise.
24223 (is_imported_p): New helper-function.
24224 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
24225 for Windows x64 targets.
24226 (ix86_expand_prologue): Optimize for pe-coff targets.
24227 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
24228 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
24229 medium/large code-model.
24230 (legitimize_pic_address): Likewise.
24231 (legitimize_tls_address): Likewise.
24232 (ix86_expand_call): Likewise.
24233 (x86_output_mi_thunk): Likewise.
24234 (get_dllimport_decl): Add new beimport argument.
24235 (construct_plt_address): Don't assert for x64 pe-coff targets.
24236 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
24238 (SYMBOL_FLAG_STUBVAR): New macro.
24239 (SYMBOL_REF_STUBVAR_P): Likewise.
24240 * config/i386/winnt.c (stub_list): New structure.
24241 (stub_head): New local variable.
24242 (i386_pe_record_stub): New function.
24243 (i386_pe_file_end): Emit refptr-stubs.
24245 2013-04-02 Jakub Jelinek <jakub@redhat.com>
24247 PR rtl-optimization/56745
24248 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
24249 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
24252 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
24253 and both of them are MEM_REFs, just compare first argument for
24254 equality and attempt to deal even with differing offsets.
24257 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
24258 of gimple_clobber_p to be MEM_REF.
24259 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
24260 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
24261 after gimplification.
24262 * asan.c (get_mem_ref_of_assignment): Don't instrument
24263 gimple_clobber_p stmts.
24264 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
24265 gimple_clobber_p stmt if they have MEM_REF lhs and
24266 are dead because of another gimple_clobber_p stmt.
24267 * tree-ssa-live.c (clear_unused_block_pointer): Treat
24268 gimple_clobber_p stmts like debug stmts.
24269 (remove_unused_locals): Remove clobbers with MEM_REF lhs
24270 that refer to unused VAR_DECLs or uninitialized values.
24271 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
24272 gimple_clobber_p stmts if they refer to removed parameters.
24273 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
24276 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
24278 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
24279 using SWI48 mode attribute.
24281 2013-04-02 Wei Mi <wmi@google.com>
24283 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
24284 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
24285 *<rotate_insn><mode>3_mask in i386.md.
24287 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
24289 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
24291 2013-04-02 Richard Biener <rguenther@suse.de>
24293 PR tree-optimization/56778
24294 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
24295 Runtime alias tests are not supported for gather loads.
24296 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
24297 stmts referenced from SSA operands before updating SSA form.
24299 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
24300 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24302 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
24303 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
24304 * config/arm/cortex-a53.md: New file.
24305 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
24306 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
24307 * config/arm/arm.c (arm_issue_rate): Likewise.
24308 * config/arm/arm-tune.md: Regenerate
24309 * config/arm/arm-tables.opt: Regenerate.
24310 * config/arm/arm-cores.def: Add cortex-a53.
24312 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
24314 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
24317 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24319 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
24320 scalar load/store operations using B/H registers.
24321 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
24323 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24325 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
24327 * config/aarch64/aarch64.c
24328 (aarch64_simd_scalar_immediate_valid_for_move): New.
24329 * config/aarch64/aarch64-protos.h
24330 (aarch64_simd_scalar_immediate_valid_for_move): New.
24331 * config/aarch64/constraints.md (Dh, Dq): New.
24332 * config/aarch64/iterators.md (hq): New.
24334 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
24336 * reorg.c (get_branch_condition): Deal with conditional returns.
24337 (fill_simple_delay_slots): Remove dead code dealing with jumps.
24339 2013-04-01 Wei Mi <wmi@google.com>
24341 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
24342 Truncate operand 2 using %b asm operand modifier.
24343 (*<shift_insn><mode>3_mask): Ditto.
24344 (*<rotate_insn><mode>3_mask): Ditto.
24346 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
24348 PR middle-end/56798
24349 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
24351 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
24353 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
24355 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
24357 2013-03-30 Lawrence Crowl <crowl@google.com>
24359 * dse.c (clear_alias_sets): Remove never set.
24360 (disqualified_clear_alias_sets): Remove never set.
24361 (clear_alias_mode_pool): Remove never set.
24362 (dse_step0): Remove condition that is never true.
24363 (canon_address): Remove condition that is never true.
24364 (dse_step7): Remove condition that is never true.
24365 (rest_of_handle_dse): Remove condition that is never true.
24366 (rest_of_handle_dse::did_global): Remove never read from above.
24367 (dse_step2_spill): Remove never called from above.
24368 (dse_step5_spill): Remove never called from above.
24370 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
24372 * doc/md.texi (Standard Names) <casesi>: Update documentation for
24373 JUMP_TABLE_DATA changes.
24374 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
24375 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
24376 (Insns) <jump_table_data>: New entry.
24377 * doc/tm.texi: Regenerate.
24379 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
24381 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
24382 for table jump at the end of a basic block using tablejump_p.
24383 * targhooks.c (default_invalid_within_doloop): Likewise.
24384 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
24385 target hook implementation that is identical to the default hook.
24386 (rs6000_invalid_within_doloop): Remove.
24388 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
24389 unused variable from tablejump_p call.
24391 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
24392 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
24393 (INSN_DELETED_P): Likewise.
24394 (emit_jump_table_data): New prototype.
24395 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
24396 after 4th as unused.
24397 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
24398 * sched-vis.c (print_insn): Likewise.
24399 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
24400 insn for compatibility with back ends that use next_active_insn to
24401 identify jump table data.
24402 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
24403 (remove_insn): Likewise.
24404 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
24406 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
24407 (emit_jump_table_data): New function.
24409 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24410 basic block, a JUMP_TABLE_DATA never is.
24411 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24412 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24413 off from code handling real insns.
24414 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24415 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24416 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24417 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24418 is not a NONDEBUG_INSN_P.
24419 * ira-costs.c (scan_one_insn): Likewise.
24420 * jump.c (mark_all_labels): Likewise.
24421 (mark_jump_label_1): Likewise.
24422 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24423 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24425 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24426 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24427 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24428 JUMP_TABLE_DATA_P insns.
24429 (calculate_elim_costs_all_insns): Likewise.
24430 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24431 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24432 (delete_output_reload): Code style fixups.
24433 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24434 insn flags on this non-insn.
24435 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24436 as scheduling barriers, for pre-change compatibility.
24437 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24438 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24440 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24441 redundant JUMP_TABLE_DATA_P test.
24442 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24443 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24444 (frv_for_each_packet): Likewise.
24445 * config/i386/i386.c (min_insn_size): Likewise.
24446 (ix86_avoid_jump_mispredicts): Likewise.
24447 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24448 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24449 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24450 (mips16_insn_length): Robustify.
24451 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24452 (mips16_split_long_branches): Likewise.
24453 * config/pa/pa.c (pa_combine_instructions): Likewise.
24454 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24455 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24456 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24457 as contributing to pool range lengths.
24458 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24459 Remove redundant JUMP_TABLE_DATA_P test.
24460 (sh_loop_align): Likewise.
24461 (split_branches): Likewise.
24462 (sh_insn_length_adjustment): Likewise.
24463 * config/spu/spu.c (get_branch_target): Likewise.
24465 2013-03-29 Jan Hubicka <jh@suse.cz>
24467 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24468 gcov streaming; stream hot bb threshold to ltrans.
24469 * predict.c (get_hot_bb_threshold): Break out from ....
24470 (maybe_hot_count_p): ... here.
24471 (set_hot_bb_threshold): New function.
24472 * lto-section-in.c (lto_section_name): Add profile.
24473 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24474 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24475 and data-streamer.h
24476 (histogram_entry): New structure.
24477 (histogram, histogram_pool): New global vars.
24478 (histogram_hash): New structure.
24479 (histogram_hash::hash): New method.
24480 (histogram_hash::equal): Likewise.
24481 (account_time_size): New function.
24482 (cmp_counts): New function.
24483 (dump_histogram): New function.
24484 (ipa_profile_generate_summary): New function.
24485 (ipa_profile_write_summary): New function.
24486 (ipa_profile_read_summary): New function.
24487 (ipa_profile): Decide on threshold.
24488 (pass_ipa_profile): Add ipa_profile_write_summary and
24489 ipa_profile_read_summary.
24490 * Makefile.in (ipa.o): Update dependencies.
24491 * lto-streamer.h (LTO_section_ipa_profile): New section.
24493 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24495 * tree.h (VAR_P): New.
24497 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24500 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24502 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24504 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24505 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24506 (control_flow_insn_p): Likewise.
24507 * cfgrtl.c (duplicate_insn_chain): Likewise.
24508 * final.c (get_attr_length_1): Likewise.
24509 (shorten_branches): Likewise.
24510 (final_scan_insn): Likewise.
24511 * function.c (instantiate_virtual_regs): Likewise.
24512 * gcse.c (insert_insn_end_basic_block): Likewise.
24513 * ira-costs.c (scan_one_insn): Likewise.
24514 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24515 * lra.c (check_rtl): Likewise.
24516 * reload1.c (elimination_costs_in_insn): Likewise.
24517 * reorg.c (follow_jumps): Likewise.
24519 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24520 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24521 (thumb_far_jump_used_p): Likewise.
24522 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24523 (workaround_speculation): Likewise.
24524 (add_sched_insns_for_speculation): Likewise.
24525 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24526 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24527 (frv_for_each_packet): Likewise.
24528 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24529 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24530 (final_emit_insn_group_barriers): Likewise.
24531 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24532 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24533 (mips16_insn_length): Likewise.
24534 * config/pa/pa.c (pa_reorg): Likewise.
24535 (pa_combine_instructions): Likewise.
24536 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24537 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24538 (sh_reorg): Likewise.
24539 (split_branches): Likewise.
24540 * config/spu/spu.c (get_branch_target): Likewise.
24542 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24545 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24547 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24548 Fix declaration name.
24550 2013-03-28 Lawrence Crowl <crowl@google.com>
24552 * graphds.h (struct graph.indicies): Remove unused.
24553 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24554 (SCOP_ORIGINAL_PDDRS): Remove unused.
24555 * sese.h (extern insert_loop_close_phis): Removed unused.
24556 (extern insert_guard_phis): Removed unused.
24557 (extern ivtype_map_elt_info): Removed unused.
24558 (new_ivtype_map_elt): Removed unused.
24559 * sese.c (ivtype_map_elt_info): Removed unused.
24561 2013-03-28 Lawrence Crowl <crowl@google.com>
24563 * Makefile.in: Add several missing include dependences.
24565 (test-dump.o): New. This object is not added to any executable,
24566 but is present for ad-hoc testing.
24568 (debug (const bitmap_head_def &)): New.
24569 (debug (const bitmap_head_def *)): New.
24571 (extern debug (const bitmap_head_def &)): New.
24572 (extern debug (const bitmap_head_def *)): New.
24574 (debug (edge_def &)): New.
24575 (debug (edge_def *)): New.
24577 (debug (basic_block_def &)): New.
24578 (debug (basic_block_def *)): New.
24580 (dump_node (const_tree, int, FILE *)): Correct source file.
24582 (debug (die_struct &)): New.
24583 (debug (die_struct *)): New.
24585 (extern debug (die_struct &)): New.
24586 (extern debug (die_struct *)): New.
24587 * gimple-pretty-print.c
24588 (debug (gimple_statement_d &)): New.
24589 (debug (gimple_statement_d *)): New.
24590 * gimple-pretty-print.h
24591 (extern debug (gimple_statement_d &)): New.
24592 (extern debug (gimple_statement_d *)): New.
24594 (debug (ira_allocno_copy &)): New.
24595 (debug (ira_allocno_copy *)): New.
24596 (debug (ira_allocno &)): New.
24597 (debug (ira_allocno *)): New.
24599 (extern debug (ira_allocno_copy &)): New.
24600 (extern debug (ira_allocno_copy *)): New.
24601 (extern debug (ira_allocno &)): New.
24602 (extern debug (ira_allocno *)): New.
24604 (debug (live_range &)): New.
24605 (debug (live_range *)): New.
24607 (debug (lra_live_range &)): New.
24608 (debug (lra_live_range *)): New.
24610 (debug (lra_live_range &)): New.
24611 (debug (lra_live_range *)): New.
24613 (debug (omega_pb_d &)): New.
24614 (debug (omega_pb_d *)): New.
24616 (extern debug (omega_pb_d &)): New.
24617 (extern debug (omega_pb_d *)): New.
24619 (debug (const rtx_def &)): New.
24620 (debug (const rtx_def *)): New.
24622 (debug_tree (tree): Move within file.
24623 (debug_raw (const tree_node &)): New.
24624 (debug_raw (const tree_node *)): New.
24625 (dump_tree_via_hooks (const tree_node *, int)): New.
24626 (debug (const tree_node &)): New.
24627 (debug (const tree_node *)): New.
24628 (debug_verbose (const tree_node &)): New.
24629 (debug_verbose (const tree_node *)): New.
24630 (debug_head (const tree_node &)): New.
24631 (debug_head (const tree_node *)): New.
24632 (debug_body (const tree_node &)): New.
24633 (debug_body (const tree_node *)): New.
24634 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24635 (debug (vec<tree, va_gc> &)): New.
24636 (debug (vec<tree, va_gc> *)): New.
24638 (extern debug (const rtx_def &)): New.
24639 (extern debug (const rtx_def *)): New.
24641 (debug_raw (simple_bitmap_def &)): New.
24642 (debug_raw (simple_bitmap_def *)): New.
24643 (debug (simple_bitmap_def &)): New.
24644 (debug (simple_bitmap_def *)): New.
24646 (extern debug (simple_bitmap_def &)): New.
24647 (extern debug (simple_bitmap_def *)): New.
24648 (extern debug_raw (simple_bitmap_def &)): New.
24649 (extern debug_raw (simple_bitmap_def *)): New.
24651 (debug (vinsn_def &)): New.
24652 (debug (vinsn_def *)): New.
24653 (debug_verbose (vinsn_def &)): New.
24654 (debug_verbose (vinsn_def *)): New.
24655 (debug (expr_def &)): New.
24656 (debug (expr_def *)): New.
24657 (debug_verbose (expr_def &)): New.
24658 (debug_verbose (expr_def *)): New.
24659 (debug (vec<rtx> &)): New.
24660 (debug (vec<rtx> *)): New.
24662 (extern debug (vinsn_def &)): New.
24663 (extern debug (vinsn_def *)): New.
24664 (extern debug_verbose (vinsn_def &)): New.
24665 (extern debug_verbose (vinsn_def *)): New.
24666 (extern debug (expr_def &)): New.
24667 (extern debug (expr_def *)): New.
24668 (extern debug_verbose (expr_def &)): New.
24669 (extern debug_verbose (expr_def *)): New.
24670 (extern debug (vec<rtx> &)): New.
24671 (extern debug (vec<rtx> *)): New.
24673 (_list_iter_cond_expr): Make inline instead of static.
24675 (debug (sreal &)): New.
24676 (debug (sreal *)): New.
24678 (extern debug (sreal &)): New.
24679 (extern debug (sreal *)): New.
24681 (extern debug_raw (const tree_node &)): New.
24682 (extern debug_raw (const tree_node *)): New.
24683 (extern debug (const tree_node &)): New.
24684 (extern debug (const tree_node *)): New.
24685 (extern debug_verbose (const tree_node &)): New.
24686 (extern debug_verbose (const tree_node *)): New.
24687 (extern debug_head (const tree_node &)): New.
24688 (extern debug_head (const tree_node *)): New.
24689 (extern debug_body (const tree_node &)): New.
24690 (extern debug_body (const tree_node *)): New.
24691 (extern debug (vec<tree, va_gc> &)): New.
24692 (extern debug (vec<tree, va_gc> *)): New.
24694 (debug (struct loop &)): New.
24695 (debug (struct loop *)): New.
24696 (debug_verbose (struct loop &)): New.
24697 (debug_verbose (struct loop *)): New.
24698 * tree-dump.c: Add header dependence.
24700 (extern debug (struct loop &)): New.
24701 (extern debug (struct loop *)): New.
24702 (extern debug_verbose (struct loop &)): New.
24703 (extern debug_verbose (struct loop *)): New.
24705 (debug (data_reference &)): New.
24706 (debug (data_reference *)): New.
24707 (debug (vec<data_reference_p> &)): New.
24708 (debug (vec<data_reference_p> *)): New.
24709 (debug (vec<ddr_p> &)): New.
24710 (debug (vec<ddr_p> *)): New.
24712 (extern debug (data_reference &)): New.
24713 (extern debug (data_reference *)): New.
24714 (extern debug (vec<data_reference_p> &)): New.
24715 (extern debug (vec<data_reference_p> *)): New.
24716 (extern debug (vec<ddr_p> &)): New.
24717 (extern debug (vec<ddr_p> *)): New.
24719 (debug (pt_solution &)): New.
24720 (debug (pt_solution *)): New.
24722 (extern debug (pt_solution &)): New.
24723 (extern debug (pt_solution *)): New.
24725 (debug (_var_map &)): New.
24726 (debug (_var_map *)): New.
24727 (debug (tree_live_info_d &)): New.
24728 (debug (tree_live_info_d *)): New.
24730 (extern debug (_var_map &)): New.
24731 (extern debug (_var_map *)): New.
24732 (extern debug (tree_live_info_d &)): New.
24733 (extern debug (tree_live_info_d *)): New.
24735 2013-03-28 Jan Hubicka <jh@suse.cz>
24737 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24739 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24741 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24742 record only when desired or required.
24744 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24746 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24747 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24749 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24751 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24752 (*andsi3_compare0_uxtw): New pattern.
24753 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24754 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24756 2013-03-28 Jan Hubicka <jh@suse.cz>
24758 * data-streamer-in.c (streamer_read_gcov_count): New function.
24759 * gimple-streamer-out.c: Include value-prof.h.
24760 (output_gimple_stmt): Output histogram.
24761 (output_bb): Use streamer_write_gcov_count.
24762 * value-prof.c: Include data-streamer.h
24763 (dump_histogram_value): Add HIST_TYPE_MAX.
24764 (stream_out_histogram_value): New function.
24765 (stream_in_histogram_value): New function.
24766 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24767 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24768 * data-streamer-out.c (streamer_write_gcov_count): New function.
24769 (streamer_write_gcov_count_stream): New function.
24770 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24771 (lto_output_node): Likewise.
24772 (input_node, input_edge): Likewise.
24773 * lto-streamer-out.c (output_cfg): Update streaming.
24774 * lto-streamer-in.c (input_cfg): Likewise.
24775 * data-streamer.h (streamer_write_gcov_count,
24776 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24777 * gimple-streamer-in.c: Include value-prof.h
24778 (input_gimple_stmt): Input histograms.
24779 (input_bb): Update profile streaming.
24781 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24783 * genmodes.c (emit_max_int): New function.
24784 (emit_insn_modes_h): Added call to emit_max_function.
24785 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24787 * machmode.def: Fixed comment.
24789 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24791 * combine.c (try_combine): Removed useless assert.
24792 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24794 2013-03-28 Marek Polacek <polacek@redhat.com>
24795 Richard Biener <rguenther@suse.de>
24797 PR tree-optimization/56695
24798 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24799 build signed result of a vector comparison.
24800 * tree-cfg.c (verify_gimple_comparison): Check that a result
24801 of a vector comparison has signed type.
24803 2013-03-28 Richard Biener <rguenther@suse.de>
24805 PR tree-optimization/37021
24806 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24807 do not restrict gaps between groups.
24808 * tree-vect-stmts.c (vectorizable_load): Properly account for
24809 a gap between groups.
24811 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24813 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24814 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24817 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24819 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24820 * doc/extend.texi (Named Address Spaces): Ditto.
24821 (Variable Attributes): Ditto.
24823 2013-03-27 Kai Tietz <ktietz@redhat.com>
24825 * config.build: Add support for cygwin x64 target.
24826 * config.gcc: Likewise.
24827 * config.host: Likewise.
24828 * configure.ac: Likewise
24829 * configure: Regenerated.
24831 2013-03-27 Kai Tietz <ktietz@redhat.com>
24833 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24834 * config/i386/t-cygwin-w64: New file.
24835 * config/i386/cygwin-w64.h: New file.
24836 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24837 and add support for x64-cygwin target.
24838 (CPP_SPEC): Likewise.
24839 (CXX_WRAP_SPEC_LIST): Undefine before define.
24840 (LIBGCJ_SONAME): Use 15 as version.
24842 2013-03-27 Richard Biener <rguenther@suse.de>
24844 PR tree-optimization/56716
24845 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24846 dumping for ref nodes.
24848 2013-03-27 Martin Jambor <mjambor@suse.cz>
24850 PR tree-optimization/55334
24851 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24852 restricted pointers to arrays.
24854 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24856 * Makefile.in (.SUFFIXES): Add .cc.
24857 (.c.o): Apply same recipe for implicit rule .cc.o.
24859 2013-03-27 Richard Biener <rguenther@suse.de>
24861 PR tree-optimization/37021
24862 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24863 REALPART/IMAGPART_EXPRs around the supported refs.
24864 * tree-ssa-structalias.c (find_func_aliases): Assume that
24865 floating-point values are not used to transfer pointers.
24867 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24869 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24870 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24871 * doc/tm.texi: Regenerate.
24872 * targhooks.h (default_has_ifunc_p): New.
24873 * targhooks.c (default_has_ifunc_p): Ditto.
24874 * config/linux-protos.h: New file.
24875 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24876 hook for linux which disables support of indirect functions in android.
24877 * config/linux-android.c: New file.
24878 * config/t-linux-android.c: Ditto.
24879 * config.gcc: Added new object file linux-android.o.
24880 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24881 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24882 * varasm.c (do_assemble_alias): Likewise.
24883 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24884 doesn't support indirect functions.
24885 * configure: Regenerate.
24887 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24890 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24891 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24894 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24896 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24898 2013-03-27 Terry Guo <terry.guo@arm.com>
24900 * config/arm/arm-cores.def: Added core cortex-r7.
24901 * config/arm/arm-tune.md: Regenerated.
24902 * config/arm/arm-tables.opt: Regenerated.
24903 * doc/invoke.texi: Added entry for core cortex-r7.
24905 2013-03-27 Walter Lee <walt@tilera.com>
24907 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24908 double-decrement of next_scratch_regno.
24910 2013-03-27 Walter Lee <walt@tilera.com>
24912 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24914 (insn_v1mulus): Ditto.
24915 (insn_v2muls): Ditto.
24917 2013-03-27 Walter Lee <walt@tilera.com>
24919 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24920 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24922 2013-03-27 Walter Lee <walt@tilera.com>
24924 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24925 (*sibcall_value): Ditto.
24927 2013-03-27 Walter Lee <walt@tilera.com>
24929 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24930 (insn_mnz_v8qi): ... this ...
24931 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24933 (insn_v<n>mnz): Replaced by ...
24934 (insn_v1mnz): ... this ...
24935 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24937 (insn_mz_<mode>): Replaced by ...
24938 (insn_mz_v8qi): ... this ...
24939 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24941 (insn_v<n>mz): Replaced by ...
24942 (insn_v1mz): ... this ...
24943 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24946 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24948 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24950 2013-03-26 Roland McGrath <mcgrathr@google.com>
24952 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24953 than fprintf with a non-constant, non-format string.
24955 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24957 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24958 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24959 operand 0 predicate.
24960 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24961 attribute. Use general_x64nomem_operand as operand 1 predicate.
24962 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24963 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24964 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24965 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24966 (mov<mode>_insv_1): Remove expander. Merge insn with
24967 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24968 Use general_x64nomem_operand as operand 1 predicate.
24969 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24970 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24971 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24972 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24973 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24974 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24975 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24976 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24977 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24978 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24979 (general_x64nomem_operand): Ditto.
24981 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24983 * config/rtems.opt: Add -pthread option.
24985 2013-03-26 Richard Biener <rguenther@suse.de>
24987 * alias.c (find_base_term): Avoid redundant and not used recursion.
24988 (base_alias_check): Get the initial base term from the caller.
24989 (true_dependence_1): Compute and pass base terms to base_alias_check.
24990 (write_dependence_p): Likewise.
24991 (may_alias_p): Likewise.
24993 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24995 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24996 PC-relative load in SI modes and above only.
24998 2013-03-26 Xinyu Qi <xyqi@marvell.com>
25000 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
25001 * config/arm/iwmmxt.md (WCGR0): Update.
25002 (WCGR1, WCGR2, WCGR3): Likewise.
25004 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
25006 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
25007 Use x64 and nox64 isa attributes.
25009 2013-03-26 Richard Biener <rguenther@suse.de>
25011 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
25012 alignment computations and rely on get_object_alignment_1
25013 for the !TYPE_P case.
25014 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
25016 2013-03-26 Walter Lee <walt@tilera.com>
25018 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
25019 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
25021 2013-03-25 Jeff Law <law@redhat.com>
25023 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
25024 check for INTEGRAL_TYPE_P that was missing due to checking in
25025 wrong version of prior patch.
25027 2013-03-25 Walter Lee <walt@tilera.com>
25029 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
25030 TILEGX_INSN_SHUFFLEBYTES1.
25031 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
25033 (tilegx_builtins): Ditto.
25034 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
25036 2013-03-25 Walter Lee <walt@tilera.com>
25038 * config/tilegx/tilegx.md (floatsisf2): New pattern.
25039 (floatunssisf2): New pattern.
25040 (floatsidf2): New pattern.
25041 (floatunssidf2): New pattern.
25043 2013-03-25 Walter Lee <walt@tilera.com>
25045 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
25046 tests for constraint J, K, N, P.
25048 2013-03-25 Walter Lee <walt@tilera.com>
25050 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
25051 Use indirect/pcrel encoding.
25052 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
25055 2013-03-25 Steve Ellcey <sellcey@mips.com>
25057 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
25058 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
25059 * config/mips/mips.c (mips_option_override): Set IMADD default.
25060 * config/mips/mips.h (PTF_AVOID_IMADD): New.
25061 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
25062 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
25063 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
25064 * doc/invoke.texi (-mimadd/-mno-imadd): New.
25066 2013-03-25 Jeff Law <law@redhat.com>
25068 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
25069 slightly to avoid creating and folding useless trees. Simplify
25070 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
25072 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
25074 * config/i386/i386.md (*zero_extendsidi2): Merge with
25075 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
25076 * config/i386/predicates.md (x86_64_zext_operand): Rename from
25077 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
25078 targets. Clarify comment.
25080 2013-03-25 Martin Jambor <mjambor@suse.cz>
25082 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
25083 pass-through jump functions differently.
25084 (ipa_read_jump_function): Likewise. Also use setter functions to set
25087 2013-03-25 Martin Jambor <mjambor@suse.cz>
25089 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
25090 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
25092 (ipa_get_indirect_edge_target): New function.
25093 (devirtualization_time_bonus): New parameter known_aggs, pass it to
25094 ipa_get_indirect_edge_target. Update all callers.
25095 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
25096 ipa_get_indirect_edge_target_1 instead of calling
25097 ipa_get_indirect_edge_target.
25098 (create_specialized_node): Pass aggvlas to
25099 ipcp_discover_new_direct_edges.
25101 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25103 * config/arm/arm.md (f_sels, f_seld): New types.
25104 (*cmov<mode>): New pattern.
25105 * config/arm/predicates.md (arm_vsel_comparison_operator): New
25108 2013-03-25 Kai Tietz <ktietz@redhat.com>
25110 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
25111 POSIX-printf for mingw-hosted builds.
25113 2013-03-25 Richard Biener <rguenther@suse.de>
25115 PR middle-end/56694
25116 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
25117 must-not-throw stmt location.
25119 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25121 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
25122 Emit load-acquire versions when acq is true.
25123 (arm_emit_store_exclusive): Add rel parameter.
25124 Emit store-release versions when rel is true.
25125 (arm_split_compare_and_swap): Use acquire-release instructions
25127 of barriers when appropriate.
25128 (arm_split_atomic_op): Likewise.
25129 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
25130 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
25131 (VUNSPEC_SLX): Likewise.
25132 (VUNSPEC_LDA): Likewise.
25133 (VUNSPEC_STL): Likewise.
25134 * config/arm/sync.md (atomic_load<mode>): New pattern.
25135 (atomic_store<mode>): Likewise.
25136 (arm_load_acquire_exclusive<mode>): Likewise.
25137 (arm_load_acquire_exclusivesi): Likewise.
25138 (arm_load_acquire_exclusivedi): Likewise.
25139 (arm_store_release_exclusive<mode>): Likewise.
25141 2013-03-25 Catherine Moore <clm@codesourcery.com>
25143 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
25144 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
25145 * config/mip/predicates.md (lwsp_swsp_operand,
25146 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
25147 sb16_operand, db4_operand, db7_operand, ib3_operand,
25148 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
25149 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
25150 andi16_operand): New predicates.
25151 * config/mips/mips.md (compression): New attribute.
25152 (enabled): New attribute.
25153 (length): Consider compression in computing length.
25154 (shift_compression): New code attribute.
25155 (*add<mode>3): New operands. Record compression.
25156 (sub<mode>3): Likewise.
25157 (one_cmpl<mode>2): Likewise.
25158 (*and<mode>3): Likewise.
25159 (*ior<mode>3): Likewise.
25160 (unnamed pattern for xor): Likewise.
25161 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
25162 (*<optab><mode>3): Likewise.
25163 (*mov<mode>_internal: Likewise.
25164 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
25165 (mips_unsigned_immediate_p): New.
25166 (umips_lwsp_swsp_address_p): New.
25167 (m16_based_address_p): New.
25168 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
25169 (mips_unsigned_immediate_p): New prototype.
25170 (lwsp_swsp_address_p): New prototype.
25171 (m16_based_address_p): New prototype.
25172 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
25173 (mips_signed_immediate_p): New function.
25174 (m16_based_address_p): New function.
25175 (lwsp_swsp_address_p): New function.
25176 (mips_print_operand_punctuation): Recognize short delay slot insns
25177 for microMIPS.add<mode>3"
25179 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25182 * config/arm/iterators.md (v_cmp_result): New mode attribute.
25183 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
25185 2013-03-25 Richard Biener <rguenther@suse.de>
25187 PR tree-optimization/56689
25188 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
25191 2013-03-25 Richard Biener <rguenther@suse.de>
25193 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
25195 (memory_references): Likewise.
25196 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
25197 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
25198 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
25199 (gather_mem_refs_in_loops): Fold into ...
25200 (analyze_memory_references): ... this. Move initialization
25201 to tree_ssa_lim_initialize.
25202 (fill_always_executed_in): Rename to ...
25203 (fill_always_executed_in_1): ... this.
25204 (fill_always_executed_in): Move contains_call computation to
25205 this new function from ...
25206 (tree_ssa_lim_initialize): ... here.
25207 (tree_ssa_lim): Call fill_always_executed_in.
25209 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25211 * postreload.c (reload_combine): Fix code detecting returns.
25213 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25215 * function.c (emit_use_return_register_into_block): On cc0 targets,
25216 do not emit the sequence between cc0 setter and user.
25218 2013-03-25 Kai Tietz <ktietz@redhat.com>
25220 * config/i386/predicates.md (local_symbolic_operand): Interpret
25221 dll-imported symbols as none-local.
25223 2013-03-25 Richard Biener <rguenther@suse.de>
25225 * tree-ssa-loop-im.c (struct depend): Remove.
25226 (struct lim_aux_data): Make depends a vec of gimples.
25227 (free_lim_aux_data): Adjust.
25228 (add_dependency): Likewise.
25229 (set_level): Likewise.
25231 2013-03-25 Richard Biener <rguenther@suse.de>
25233 PR middle-end/56434
25234 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
25235 the pointer returned by calls with ECF_MALLOC set.
25237 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25239 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
25241 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25243 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
25244 using MMXMODE mode iterator.
25245 (*move<mode>_internal): Merge with *movv2sf_internal and
25246 *movv2sf_internal_rex64 using MMXMODE mode iterator.
25248 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25250 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
25251 (record_last_mem_set_info): Likewise.
25253 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
25254 of XNEWVEC followed by memset.
25255 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
25257 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25259 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
25260 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
25261 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
25262 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
25263 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
25264 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
25265 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
25266 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
25267 BARRIER_P instead of GET_CODE.
25269 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25271 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
25272 inaccuracy in the probing code.
25274 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
25275 (ctrapdi4): Likewise.
25277 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25279 * calls.c (expand_call): Add missing guard to code handling return
25280 of non-BLKmode structures in MSB.
25281 * function.c (expand_function_end): Likewise.
25283 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25285 * combine.c (try_combine): Adjust comment. Do not add the set of
25286 insn #0 if the destination indirectly is set or dies in insn #2.
25287 Tidy up code to distribute a new note.
25289 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25291 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
25292 also for alternatives 16 and 17.
25294 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25296 * config/i386/sse.md (*mov<mode>_internal): Merge with
25297 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
25298 Emit insn template depending on type attribute. Use
25299 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25300 movd instead of movq mnemonic for interunit moves. Rewrite mode
25301 attribute calculation. Remove unit attribute calculation.
25302 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
25303 Set prefix_data16 attribute for DImode ssemov types.
25304 Use Ym instead of y for SSE-MMX conversion alternatives.
25305 Reorder operand constraints.
25307 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
25309 * df.h (df_insn_delete): Adjust prototype.
25310 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
25311 and let it decide whether mark the basic block dirty.
25312 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
25313 * df-scan.c (df_insn_info_delete): New helper function, split
25314 off from df_insn_delete.
25315 (df_scan_free_bb_info): Use it.
25316 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
25318 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
25319 that the insn is actually an insn and it has a non-NULL basic block.
25320 Do not mark basic block dirty if only deleting a DEBUG_INSN.
25322 2013-03-22 Richard Biener <rguenther@suse.de>
25324 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
25326 (mem_ref_alloc): Do not allocate them.
25327 (refs_independent_p): Do not query or maintain a cache.
25329 2013-03-22 Richard Biener <rguenther@suse.de>
25331 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
25332 (gather_mem_refs_in_loops): Do not compute it.
25333 (analyze_memory_references): Do not allocate it.
25334 (tree_ssa_lim_finalize): Do not free it.
25335 (for_all_locs_in_loop): Do not query all_refs_in_loop.
25337 2013-03-22 Richard Biener <rguenther@suse.de>
25339 * is-a.h (as_a): Use gcc_checking_assert.
25341 2013-03-22 Ian Bolton <ian.bolton@arm.com>
25343 * config/aarch64/aarch64.c (aarch64_print_operand): New
25344 format specifier for printing a constant in hex.
25345 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
25346 format specifier for printing second operand.
25348 2013-03-22 Richard Biener <rguenther@suse.de>
25350 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
25352 (gather_mem_refs_in_loops): Perform store accumulation here.
25353 (create_vop_ref_mapping_loop): Remove.
25354 (create_vop_ref_mapping): Likewise.
25355 (analyze_memory_references): Initialize refs_stored_in_loop.
25356 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
25357 (record_indep_loop): Remove.
25358 (record_dep_loop): New function.
25359 (ref_indep_loop_p_1): Adjust to only walk over references
25360 in the loop, not its subloops.
25361 (ref_indep_loop_p): Rename to ...
25362 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
25363 maintaining a more fine-grained cache.
25364 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
25365 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
25367 2013-03-22 Richard Biener <rguenther@suse.de>
25369 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
25370 (struct mem_ref): Make accesses_in_loop a vec of a vec of
25371 aggregate mem_ref_loc.
25372 (free_mem_ref_locs): Inline into ...
25373 (memref_free): ... this and adjust.
25374 (mem_ref_alloc): Adjust.
25375 (mem_ref_locs_alloc): Remove.
25376 (record_mem_ref_loc): Adjust.
25377 (get_all_locs_in_loop): Rewrite into ...
25378 (for_all_locs_in_loop): ... this iterator.
25379 (rewrite_mem_ref_loc): New functor.
25380 (rewrite_mem_refs): Use for_all_locs_in_loop.
25381 (sm_set_flag_if_changed): New functor.
25382 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
25383 (ref_always_accessed): New functor.
25384 (ref_always_accessed_p): Use for_all_locs_in_loop.
25386 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25388 * tree-pass.h (PROP_gimple_lvec): New.
25389 * passes.c (dump_properties): Handle PROP_gimple_lvec.
25390 (init_optimization_passes): Move pass_lower_vector.
25391 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
25393 (pass_lower_vector): Provide PROP_gimple_lvec.
25394 (pass_lower_vector_ssa): Likewise.
25395 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
25397 2013-03-21 Mark Wielaard <mjw@redhat.com>
25399 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
25401 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25403 * config/i386/i386.md (*movdi_internal): Disparage slightly
25404 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
25405 conversion alternatives.
25407 2013-03-21 Jakub Jelinek <jakub@redhat.com>
25409 PR middle-end/48087
25410 * diagnostic.def (DK_WERROR): New kind.
25411 * diagnostic.h (werrorcount): Define.
25412 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25413 promoted to DK_ERROR, increment DK_WERROR counter instead of
25415 * toplev.c (toplev_main): Call print_ignored_options even if
25416 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25417 even if just werrorcount is non-zero.
25420 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25422 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25423 (string_cst_pool_decl): New function.
25424 (optimize_one_addr_into_implicit_ptr): New function.
25425 (resolve_addr_in_expr): Optimize DWARF location expression
25426 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25427 which doesn't live in memory, but has DW_AT_location or
25428 DW_AT_const_value, or refers to a string literal, into
25429 DW_OP_GNU_implicit_pointer.
25430 (optimize_location_into_implicit_ptr): New function.
25431 (resolve_addr): If removing DW_AT_location of a variable because
25432 it was DW_OP_addr of address of the variable, but the variable doesn't
25433 live in memory, try to emit const value attribute for the initializer.
25435 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25437 * tree.h (VECTOR_TYPE_P): New macro.
25438 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25439 TYPE_MODE): Use it.
25440 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25441 VEC_COND_EXPR cannot be lvalues.
25442 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25444 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25446 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25447 Restrict the transformation to equal modes.
25449 2013-03-21 Richard Biener <rguenther@suse.de>
25451 PR tree-optimization/39326
25452 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25453 (MEM_ANALYZABLE): Adjust.
25454 (record_mem_ref_loc): Move bitmap ops ...
25455 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25456 unanalyzable refs, do not record locations for it.
25457 (analyze_memory_references): Allocate ref zero as shared
25459 (refs_independent_p): Do not test for unanalyzed mems here.
25460 (ref_indep_loop_p_1): Special-case disambiguation against
25461 the unanalyzed ref.
25462 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25464 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25466 * config/arm/arm-protos.h (tune_params): Add
25467 prefer_neon_for_64bits field.
25468 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25469 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25470 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25471 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25472 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25473 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25474 (arm_option_override): Handle -mneon-for-64bits new option.
25475 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25476 (prefer_neon_for_64bits): Declare new variable.
25477 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25478 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25479 (arch_enabled): Handle new arch types. Remove support for onlya8
25481 (one_cmpldi2): Use new arch names.
25482 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25483 * config/arm/arm.opt (mneon-for-64bits): Add option.
25484 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25485 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25486 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25488 * doc/invoke.texi (-mneon-for-64bits): Document.
25490 2013-03-21 Richard Biener <rguenther@suse.de>
25492 PR tree-optimization/39326
25493 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25494 (sort_bbs_in_loop_postorder_cmp): New function.
25495 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25497 2013-03-21 Richard Biener <rguenther@suse.de>
25499 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25500 (vect_insert_into_interleaving_chain): Likewise.
25501 (vect_drs_dependent_in_basic_block): Inline ...
25502 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25504 (vect_analyze_data_ref_dependence): ... here. Simplify.
25505 (vect_check_interleaving): Simplify.
25506 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25507 (vect_slp_analyze_data_ref_dependences): ... this new function.
25508 (dr_group_sort_cmp): New function.
25509 (vect_analyze_data_ref_accesses): Compute data-reference groups
25510 here instead of in vect_analyze_data_ref_dependence. Use
25511 a more efficient algorithm.
25512 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25513 vect_slp_analyze_data_ref_dependences. Call
25514 vect_analyze_data_ref_accesses earlier.
25515 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25516 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25517 (vect_slp_analyze_data_ref_dependences): New prototype.
25519 2013-03-21 Richard Biener <rguenther@suse.de>
25521 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25522 ref is stored in the loop.
25523 (find_refs_for_sm): Walk only over all stores.
25524 (store_motion_loop): Allocate from lim_bitmap_obstack.
25525 (store_motion): Likewise.
25527 2013-03-21 Richard Biener <rguenther@suse.de>
25529 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25530 Update virtual SSA form.
25532 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25534 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25535 * configure: Regenerate.
25536 * config.in: Regenerate.
25537 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25538 if !HAVE_LD_EH_FRAME_CIEV3.
25540 2013-03-21 Richard Biener <rguenther@suse.de>
25542 * tree-cfg.c (verify_expr_no_block): New function.
25543 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25544 nor DECL_VALUE_EXPR have locations with associated blocks.
25545 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25546 (clear_unused_block_pointer): Remove code dealing with
25547 blocks in DECL_DEBUG_EXPR locations.
25549 2013-03-21 Richard Biener <rguenther@suse.de>
25551 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25552 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25553 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25554 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
25555 instead of DECL_DEBUG_EXPR_IS_FROM.
25556 * gimplify.c (gimplify_modify_expr): Likewise.
25557 * tree-cfg.c (verify_expr_location_1): Likewise.
25558 * tree-complex.c (create_one_component_var): Likewise.
25559 * tree-sra.c (create_access_replacement): Likewise.
25560 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
25561 (clear_unused_block_pointer): Likewise.
25562 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
25563 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
25564 * var-tracking.c (var_debug_decl): Likewise.
25565 (track_expr_p): Likewise.
25566 * tree-inline.c (add_local_variables): Likewise. Set
25567 DECL_HAS_DEBUG_EXPR_P after copying it.
25568 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
25569 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
25571 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25574 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
25575 * configure: Regenerate.
25576 * config.in: Regenerate.
25577 * config/i386/i386.md (*movdf_internal): Use
25578 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25579 movd instead of movq mnemonic for interunit moves.
25580 (*movdi_internal): Ditto.
25582 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25584 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
25585 (abd<mode>_3): New pattern.
25586 (aba<mode>_3): New pattern.
25587 (fabd<mode>_3): New pattern.
25589 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25591 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
25592 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
25593 occurrence of REGISTER_PREFIX as its empty string.
25595 2013-03-20 Jeff Law <law@redhat.com>
25597 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
25598 addititional equivalences for equality comparisons between an SSA_NAME
25599 and a constant where the SSA_NAME was set from a widening conversion.
25601 2013-03-20 Walter Lee <walt@tilera.com>
25603 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25605 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25607 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25608 depending on type attribute.
25609 (*movti_internal): Ditto.
25610 (*movtf_internal): Ditto.
25611 (*movxf_internal): Ditto.
25612 (*movdf_internal): Ditto.
25613 (*movsf_internal): Ditto.
25615 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25617 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25618 maybe_vex for sselog1 and ssemov types.
25619 (*movdi_internal): Reorder operand constraints.
25620 (*movsi_internal): Ditto. Set prefix attribute to
25621 maybe_vex for sselog1 and ssemov types.
25622 (*movtf_internal): Set prefix attribute to maybe_vex
25623 for sselog1 and ssemov types.
25624 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25625 DImode ssemov types. Reorder operand constraints.
25626 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25627 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25628 attribute for SImode ssemov types. Reorder operand constraints.
25630 2013-03-20 Martin Jambor <mjambor@suse.cz>
25632 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25633 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25635 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25637 * config/rs6000/predicates.md (indexed_address, update_address_mem
25638 update_indexed_address_mem): New predicates.
25639 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25640 attribute for load/store instructions.
25641 * config/rs6000/dfp.md (movsd_store): Likewise.
25642 (movsd_load): Likewise.
25643 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25644 (unnamed HI->DI extend define_insn): Likewise.
25645 (unnamed SI->DI extend define_insn): Likewise.
25646 (unnamed QI->SI extend define_insn): Likewise.
25647 (unnamed QI->HI extend define_insn): Likewise.
25648 (unnamed HI->SI extend define_insn): Likewise.
25649 (unnamed HI->SI extend define_insn): Likewise.
25650 (extendsfdf2_fpr): Likewise.
25651 (movsi_internal1): Likewise.
25652 (movsi_internal1_single): Likewise.
25653 (movhi_internal): Likewise.
25654 (movqi_internal): Likewise.
25655 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25656 attribute for load/store instructions.
25657 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25659 (mov<mode>_softfloat): Likewise.
25660 (mov<mode>_hardfloat32): Likewise.
25661 (mov<mode>_hardfloat64): Likewise.
25662 (mov<mode>_softfloat64): Likewise.
25663 (movdi_internal32): Likewise.
25664 (movdi_internal64): Likewise.
25665 (probe_stack_<mode>): Likewise.
25667 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25669 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25670 floating point, and decimal floating point to reload iterator.
25672 * config/rs6000/constraints.md (wl constraint): New constraints to
25673 return FLOAT_REGS if certain options are used to reduce the number
25674 of separate patterns that exist in the file.
25675 (wx constraint): Likewise.
25676 (wz constraint): Likewise.
25678 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25679 -mdebug=reg, print wg, wl, wx, and wz constraints.
25680 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25681 Initialize the reload functions for 64-bit binary/decimal floating
25683 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25684 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25685 create the buffer on the stack to overcome not having a 32-bit
25687 (rs6000_emit_move): Likewise.
25688 (rs6000_secondary_memory_needed_rtx): Likewise.
25689 (rs6000_alloc_sdmode_stack_slot): Likewise.
25690 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25691 via xxlxor, just like DFmode 0.0.
25693 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25694 define as 1 if we are running on a power7 or newer.
25695 (enum r6000_reg_class_enum): Add new constraints.
25697 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25698 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25699 with other moves by using conditional constraits (wg). Use LFIWZX
25700 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25701 (movsd splitter): Likewise.
25702 (movsd_hardfloat): Likewise.
25703 (movsd_softfloat): Likewise.
25705 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25706 binary and decimal floating point moves.
25707 (fmove_ok): New attributes to combine binary and decimal floating
25708 point moves, and to combine power6x (mfpgpr) moves along normal
25710 (real_value_to_target): Likewise.
25711 (f32_lr): Likewise.
25712 (f32_lm): Likewise.
25713 (f32_li): Likewise.
25714 (f32_sr): Likewise.
25715 (f32_sm): Likewise.
25716 (f32_si): Likewise.
25717 (movsf): Combine binary and decimal floating point moves. Combine
25718 power6x (mfpgpr) moves with other moves by using conditional
25719 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25720 (mov<mode> for SFmode/SDmode); Likewise.
25721 (SFmode/SDmode splitters): Likewise.
25722 (movsf_hardfloat): Likewise.
25723 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25724 (movsf_softfloat): Likewise.
25725 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25727 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25728 wx and wz constraints.
25730 * config/rs6000/constraints.md (wg constraint): New constraint to
25731 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25733 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25736 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25737 -mdebug=reg, print wg, wl, wx, and wz constraints.
25738 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25739 Initialize the reload functions for 64-bit binary/decimal floating
25741 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25742 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25743 create the buffer on the stack to overcome not having a 32-bit
25745 (rs6000_emit_move): Likewise.
25746 (rs6000_secondary_memory_needed_rtx): Likewise.
25747 (rs6000_alloc_sdmode_stack_slot): Likewise.
25748 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25749 via xxlxor, just like DFmode 0.0.
25751 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25752 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25753 with other moves by using conditional constraits (wg). Use LFIWZX
25754 and STFIWX for loading SDmode on power7.
25755 (movdd splitters): Likewise.
25756 (movdd_hardfloat32): Likewise.
25757 (movdd_softfloat32): Likewise.
25758 (movdd_hardfloat64_mfpgpr): Likewise.
25759 (movdd_hardfloat64): Likewise.
25760 (movdd_softfloat64): Likewise.
25762 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25763 64-bit binary and decimal floating point moves.
25764 (FMOVE64X): Likewise.
25765 (movdf): Combine 64-bit binary and decimal floating point moves.
25766 Combine power6x (mfpgpr) moves with other moves by using
25767 conditional constraits (wg).
25768 (mov<mode> for DFmode/DDmode): Likewise.
25769 (DFmode/DDmode splitters): Likewise.
25770 (movdf_hardfloat32): Likewise.
25771 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25772 (movdf_softfloat32): Likewise.
25773 (movdf_hardfloat64_mfpgpr): Likewise.
25774 (movdf_hardfloat64): Likewise.
25775 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25776 (movdf_softfloat64): Likewise.
25777 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25778 (reload_<mode>_load): Move to later in the file so they aren't in
25779 the middle of the floating point move insns.
25780 (reload_<mode>_store): Likewise.
25782 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25785 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25786 constraint if -mdebug=reg.
25787 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25788 Enable using dd reload support if needed.
25790 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25791 binary and decimal floating point moves in rs6000.md.
25792 (movtd_internal): Likewise.
25794 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25795 decimal floating point moves.
25797 (movtf_internal): Likewise.
25798 (mov<mode>_internal, TDmode/TFmode): Likewise.
25799 (movtf_softfloat): Likewise.
25800 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25802 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25803 movdi_internal64, using wg constraint for move direct operations.
25804 (movdi_internal64): Likewise.
25806 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25807 MODES_TIEABLE_P for selected modes. Print the numerical value of
25808 the various virtual registers. Use GPR/FPR first/last values,
25809 instead of hard coding the register numbers. Print which modes
25810 have reload functions registered.
25811 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25812 settings before/after setting cpu, target and subtarget settings.
25813 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25814 and for secondary reload failures in rs6000_secondary_reload_inner.
25815 (rs6000_secondary_reload_fail): Likewise.
25816 (rs6000_secondary_reload_inner): Likewise.
25818 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25819 macros for first/last GPR and FPR registers.
25820 (LAST_GPR_REGNO): Likewise.
25821 (FIRST_FPR_REGNO): Likewise.
25822 (LAST_FPR_REGNO): Likewise.
25824 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25825 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25826 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25827 (vcond<mode><mode>): Likewise.
25828 (vcondu<mode><mode>): Likewise.
25829 (vector_gtu<mode>): Likewise.
25830 (vector_gte<mode>): Likewise.
25831 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25832 to prevent the compiler from converting DImode operations to TImode.
25833 (ior<mode>3): Likewise.
25834 (and<mode>3): Likewise.
25835 (one_cmpl<mode>2): Likewise.
25836 (nor<mode>3): Likewise.
25837 (andc<mode>3): Likewise.
25839 * config/rs6000/constraints.md (wt constraint): New constraint
25840 that returns VSX_REGS if TImode is allowed in VSX registers.
25842 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25843 constant under VSX.
25845 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25846 similar to TImode, but it is restricted to being in the GPRs.
25848 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25849 TImode to occupy a single VSX register.
25851 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25852 -mvsx-timode for power7/power8.
25853 (power7 cpu): Likewise.
25854 (power8 cpu): Likewise.
25856 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25857 sure that TFmode/TDmode take up two registers if they are ever
25858 allowed in the upper VSX registers.
25859 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25861 (rs6000_init_hard_regno_mode_ok): Likewise.
25862 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25863 constraint. Print if LRA is turned on.
25864 (rs6000_option_override_internal): Give an error if -mvsx-timode
25865 and VSX is not enabled.
25866 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25867 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25868 to reg+offset addressing. Use PTImode when checking offset
25869 addresses for validity.
25870 (reg_offset_addressing_ok_p): Likewise.
25871 (rs6000_legitimate_offset_address_p): Likewise.
25872 (rs6000_legitimize_address): Likewise.
25873 (rs6000_legitimize_reload_address): Likewise.
25874 (rs6000_legitimate_address_p): Likewise.
25875 (rs6000_eliminate_indexed_memrefs): Likewise.
25876 (rs6000_emit_move): Likewise.
25877 (rs6000_secondary_reload): Likewise.
25878 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25879 reloads to fpr registers to continue to use reg+offset addressing,
25880 but 64-bit reloads to altivec registers need reg+reg addressing.
25881 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25882 it. Treat LO_SUM like a PLUS operation.
25883 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25884 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25885 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25886 registers to share a register with a smaller sized type, since VSX
25887 puts scalars in the upper 64-bits.
25888 (print_operand): Add support for PTImode.
25889 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25890 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25891 registers, but don't have arithmetic support.
25892 (rs6000_memory_move_cost): Add test for VSX.
25893 (rs6000_opt_masks): Add -mvsx-timode.
25895 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25898 (VSr): Use wt constraint for TImode.
25899 (VSv): Drop TImode support.
25900 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25901 (vsx_movti_64bit): Likewise.
25902 (vsx_movti_32bit): Likewise.
25903 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25904 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25905 one '?' on the appropriate output constraint. Do not allow TImode
25906 logical operations on 32-bit systems.
25907 (vsx_ior<mode>3): Likewise.
25908 (vsx_xor<mode>3): Likewise.
25909 (vsx_one_cmpl<mode>2): Likewise.
25910 (vsx_nor<mode>3): Likewise.
25911 (vsx_andc<mode>3): Likewise.
25912 (vsx_concat_<mode>): Likewise.
25913 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25915 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25916 OPTION_MASK_VSX_TIMODE.
25917 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25918 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25920 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25921 (TI2 iterator): New iterator for TImode, PTImode.
25922 (wd mode attribute): Add values for vector types.
25923 (movti_string): Replace TI move operations with operations for TImode
25924 and PTImode. Add support for TImode being allowed in VSX registers.
25925 (mov<mode>_string, TImode/PTImode): Likewise.
25926 (movti_ppc64): Likewise.
25927 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25928 (TI mode splitters): Likewise.
25930 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25933 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25935 PR tree-optimization/56355
25936 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25937 Also handle integers with undefined overflow.
25939 2013-03-20 Catherine Moore <clm@codesourcery.com>
25940 Maciej W. Rozycki <macro@codesourcery.com>
25941 Tom de Vries <tom@codesourcery.com>
25942 Nathan Sidwell <nathan@codesourcery.com>
25943 Iain Sandoe <iain@codesourcery.com>
25944 Nathan Froyd <froydnj@codesourcery.com>
25945 Chao-ying Fu <fu@mips.com>
25947 * doc/extend.texi: (micromips, nomicromips, nocompression):
25948 Document new function attributes.
25949 * doc/invoke.texi (minterlink-compressed, mmicromips,
25950 m14k, m14ke, m14kec): Document new options.
25951 (minterlink-mips16): Update documentation.
25952 * doc/md.texi (ZC, ZD): Document new constraints.
25953 * configure.ac (gcc_cv_as_micromips): Check if linker
25954 supports the .set micromips directive.
25955 * configure: Regenerate.
25956 * config.in: Regenerate.
25957 * config/mips/mips-tables.opt: Regenerate.
25958 * config/mips/micromips.md: New file.
25959 * constraints.md (ZC, ZD): New constraints.
25960 * config/mips/predicates.md (movep_src_register): New predicate.
25961 (movep_src_operand): New predicate.
25962 (non_volatile_mem_operand): New predicate.
25963 * config/mips/mips.md (multimem): New type.
25964 (length): Differentiate between 17-bit and 18-bit branch offsets.
25965 (MOVEP1, MOVEP2): New mode iterator.
25966 (mov_<load>l): Use ZC constraint.
25967 (mov_<load>r): Likewise.
25968 (mov_<store>l): Likewise.
25969 (mov_<store>r): Likewise.
25970 (*branch_equality<mode>_inverted): Add microMIPS support.
25971 (*branch_equality<mode>): Likewise.
25972 (*jump_absolute): Likewise.
25973 (indirect_jump_<mode>): Likewise.
25974 (tablejump_<mode>): Likewise.
25975 (<optab>_internal): Likewise.
25976 (sibcall_internal): Likewise.
25977 (sibcall_value_internal): Likewise.
25978 (prefetch): Use constraint ZD.
25979 * config/mips/mips.opt (minterlink-compressed): New option.
25980 (minterlink-mips16): Now an alias for minterlink-compressed.
25981 (mmicromips): New option.
25982 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25983 (compare_and_swap_12): Likewise.
25984 (sync_add<mode>): Likewise.
25985 (sync_<optab>_12): Likewise.
25986 (sync_old_<optab>_12): Likewise.
25987 (sync_new_<optab>_12): Likewise.
25988 (sync_nand_12): Likewise.
25989 (sync_old_nand_12): Likewise.
25990 (sync_new_nand_12): Likewise.
25991 (sync_sub<mode>): Likewise.
25992 (sync_old_add<mode>): Likewise.
25993 (sync_old_sub<mode>): Likewise.
25994 (sync_new_add<mode>): Likewise.
25995 (sync_new_sub<mode>): Likewise.
25996 (sync_<optab><mode>): Likewise.
25997 (sync_old_<optab><mode>): Likewise.
25998 (sync_new_<optab><mode>): Likewise.
25999 (sync_nand<mode>): Likewise.
26000 (sync_old_nand<mode>): Likewise.
26001 (sync_new_nand<mode>): Likewise.
26002 (sync_lock_test_and_set<mode>): Likewise.
26003 (test_and_set_12): Likewise.
26004 (atomic_compare_and_swap<mode>): Likewise.
26005 (atomic_exchange<mode>_llsc): Likewise.
26006 (atomic_fetch_add<mode>_llsc): Likewise.
26007 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26008 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26009 (umips_save_restore_pattern_p): Likewise.
26010 (umips_load_store_pair_p): Likewise.
26011 (umips_output_load_store_pair): Likewise.
26012 (umips_movep_target_p): Likewise.
26013 (umips_12bit_offset_address_p): Likewise.
26014 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26015 (mips_base_mips16): Rename this...
26016 (mips_base_compression_flags): ...to this. Update all uses.
26017 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26018 (mips_mips16_decl_p): Delete.
26019 (mips_nomips16_decl_p): Delete.
26020 (mips_get_compress_on_flags): New function.
26021 (mips_get_compress_off_flags): New function.
26022 (mips_get_compress_mode): New function.
26023 (mips_get_compress_on_name): New function.
26024 (mips_get_compress_off_name): New function.
26025 (mips_insert_attributes): Support multiple compression types.
26026 (mips_merge_decl_attributes): Likewise.
26027 (umips_12bit_offset_address_p): New function.
26028 (mips_start_function_definition): Emit .set micromips directive.
26029 (mips_call_may_need_jalx_p): New function.
26030 (mips_function_ok_for_sibcall): Add microMIPS support.
26031 (mips_print_operand_punctuation): Support short delay slots and
26033 (umips_swm_mask, umips_swm_encoding): New.
26034 (umips_build_save_restore): New function.
26035 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26036 (was_mips16_p): Remove.
26037 (old_compression_mode): New.
26038 (mips_set_compression_mode): New function.
26039 (mips_set_current_function): Add microMIPS support.
26040 (mips_option_override): Likewise.
26041 (umips_save_restore_pattern_p): New function.
26042 (umips_output_save_restore): New function.
26043 (umips_load_store_pair_p_1): New function.
26044 (umips_load_store_pair_p): New function.
26045 (umips_output_load_store_pair_1): New function.
26046 (umips_output_load_store_pair): New function.
26047 (umips_movep_target_p) New function.
26048 (mips_prepare_pch_save): Add microMIPS support.
26049 * config/mips/mips.h (TARGET_COMPRESSION): New.
26050 (TARGET_CPU_CPP_BUILTINS): Update macro
26051 to use new compression flags and to support microMIPS.
26052 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26053 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26054 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26055 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26056 (ASM_SPEC): Support mmicromips and mno-micromips.
26057 (M16STORE_REG_P): New macro.
26058 (MIPS_CALL): Support TARGET_MICROMIPS.
26059 (MICROMIPS_J): New macro.
26060 (mips_base_mips16): Rename this...
26061 (mips_base_compression_flags): ...to this.
26062 (UMIPS_12BIT_OFFSET_P): New macro.
26063 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26064 (MULTILIB_DIRNAMES): Likewise.
26065 2013-03-20 Richard Biener <rguenther@suse.de>
26067 PR tree-optimization/56661
26068 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
26069 the result does not have to be distinct.
26071 2013-03-20 Richard Biener <rguenther@suse.de>
26073 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
26076 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
26077 Steven Bosscher <steven@gcc.gnu.org>
26079 PR rtl-optimization/56605
26080 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
26082 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26085 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
26086 that require movd instead of movq.
26088 2013-03-20 Richard Biener <rguenther@suse.de>
26090 * tree-ssa-structalias.c (struct variable_info): Add pointer
26091 to the first field of an aggregate with sub-vars. Make
26092 this and the pointer to the next subfield its ID.
26093 (vi_next): New function.
26094 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
26095 storedanything_id, integer_id): Increment by one.
26096 (new_var_info, get_call_vi, lookup_call_clobber_vi,
26097 get_call_clobber_vi): Adjust.
26098 (solution_set_expand): Simplify and speedup.
26099 (solution_set_add): Inline into ...
26100 (set_union_with_increment): ... this. Adjust accordingly.
26101 (do_sd_constraint): Likewise.
26102 (do_ds_constraint): Likewise.
26103 (do_complex_constraint): Simplify.
26104 (build_pred_graph): Adjust.
26105 (solve_graph): Likewise. Simplify and speedup.
26106 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
26107 get_constraint_for_component_ref, get_constraint_for_1,
26108 first_vi_for_offset, first_or_preceding_vi_for_offset,
26109 create_function_info_for, create_variable_info_for_1,
26110 create_variable_info_for, intra_create_variable_infos): Adjust.
26111 (init_base_vars): Push NULL for ID zero.
26112 (compute_points_to_sets): Adjust.
26114 2013-03-20 Richard Biener <rguenther@suse.de>
26116 * cfgloop.c (verify_loop_structure): Streamline and avoid
26117 ICEing on corrupt loop tree.
26118 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
26121 2013-03-20 Richard Biener <rguenther@suse.de>
26123 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
26124 check whether an SSA update is needed.
26126 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
26128 * config/mips/constraints.md (T): Rename to...
26132 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
26133 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
26135 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26137 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
26138 (*subsi3_carryin_uxtw): Likewise.
26140 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26142 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
26143 (*rorsi3_insn_uxtw): Likewise.
26145 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26147 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
26148 (*extrsi5_insn_uxtw): Likewise.
26150 2013-03-19 Richard Biener <rguenther@suse.de>
26152 PR tree-optimization/56273
26153 * passes.c (init_optimization_passes): Move second VRP after DOM.
26155 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
26157 * config/i386/i386.md (*movti_internal): Merge from
26158 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
26159 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
26160 nox64 isa attributes.
26162 2013-03-18 Richard Biener <rguenther@suse.de>
26164 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
26166 (merge_node_constraints): Likewise.
26167 (build_succ_graph): Likewise.
26168 (valid_graph_edge): Inline into single caller.
26169 (unify_nodes): Likewise. Use bitmap_set_bit return value
26171 (scc_visit): Fix formatting and variable use.
26172 (do_sd_constraint): Use gcc_checking_assert.
26173 (do_ds_constraint): Likewise.
26174 (do_complex_constraint): Likewise.
26175 (condense_visit): Likewise. Cleanup.
26176 (dump_pred_graph): New function.
26177 (perform_var_substitution): Dump the pred-graph before
26178 variable substitution.
26179 (find_equivalent_node): Use gcc_checking_assert.
26180 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
26182 2013-03-18 Richard Biener <rguenther@suse.de>
26184 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
26185 Remove cond_expr_stmt_list argument and do not gimplify the
26187 (vect_loop_versioning): Adjust.
26188 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
26189 Cleanup to use less temporaries.
26190 (vect_create_data_ref_ptr): Cleanup.
26192 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26194 PR tree-optimization/56635
26195 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
26196 require types_compatible_p types.
26198 2013-03-18 Nick Clifton <nickc@redhat.com>
26200 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
26201 spurious backslash.
26203 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
26204 Add missing line to comment describing function.
26206 2013-03-18 Richard Biener <rguenther@suse.de>
26208 PR tree-optimization/56210
26209 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26210 Handle string / character search functions.
26211 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
26213 2013-03-18 Richard Biener <rguenther@suse.de>
26215 PR middle-end/56483
26216 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
26217 and implement properly.
26218 * gimple.h (gimple_cond_single_var_p): Remove.
26220 2013-03-18 Richard Biener <rguenther@suse.de>
26222 * tree-data-ref.h (find_data_references_in_loop): Declare.
26223 * tree-data-ref.c (get_references_in_stmt): Use a stack
26224 vector pre-allocated in the callers.
26225 (find_data_references_in_stmt): Adjust.
26226 (graphite_find_data_references_in_stmt): Likewise.
26227 (create_rdg_vertices): Likewise.
26228 (find_data_references_in_loop): Export.
26229 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
26230 Compute dependences here...
26231 (vect_analyze_data_refs): ...not here. When we encounter
26232 a non-vectorizable data reference in basic-block vectorization
26233 truncate the data reference vector. Do not bother to
26234 fixup data-dependence information for gather loads.
26235 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
26236 of data references, as reported.
26238 2013-03-18 Richard Biener <rguenther@suse.de>
26240 PR tree-optimization/3713
26241 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
26242 has_constants and expr.
26243 (stmt_has_constants): Properly valueize SSA names when deciding
26244 whether the stmt has constants.
26246 2013-03-18 Richard Biener <rguenther@suse.de>
26248 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
26249 whole function when there is nothing to do.
26250 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
26251 * tree-vectorizer.c (vectorize_loops): Update virtual and
26252 loop-closed SSA once.
26253 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
26255 2013-03-18 Richard Biener <rguenther@suse.de>
26257 PR middle-end/56113
26258 * domwalk.c (bb_postorder): New global static.
26259 (cmp_bb_postorder): New function.
26260 (walk_dominator_tree): Replace scheme imposing an order for
26261 visiting dominator sons by one sorting them at the time they
26262 are pushed on the stack.
26264 2013-03-18 Richard Biener <rguenther@suse.de>
26266 PR tree-optimization/39326
26267 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
26268 (struct mem_ref): Replace mem member with ao_ref typed member.
26269 (MEM_ANALYZABLE): Adjust.
26270 (memref_eq): Likewise.
26271 (mem_ref_alloc): Likewise.
26272 (gather_mem_refs_stmt): Likewise.
26273 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
26274 (execute_sm_if_changed_flag_set): Adjust.
26275 (execute_sm): Likewise.
26276 (ref_always_accessed_p): Likewise.
26277 (refs_independent_p): Likewise.
26278 (can_sm_ref_p): Likewise.
26280 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26283 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
26284 return 1 even for !unsignedp.
26286 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26288 * config/i386/i386.md (isa): Add x64 and nox64.
26289 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
26290 (*pushtf): Enable *roF alternative for x64 isa only.
26291 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
26292 mode attribute of integer alternatives to DImode for TARGET_64BIT.
26293 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
26294 (*movtf_internal): Merge from *movtf_internal_rex64 and
26295 *movtf_internal_sse. Use x64 and nox64 isa attributes.
26296 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
26297 nox64 isa attributes.
26298 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
26299 nox64 isa attributes.
26300 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
26302 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26304 * config/alpha/alpha.c (TARGET_LRA_P): New define.
26306 2013-03-17 Jakub Jelinek <jakub@redhat.com>
26309 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
26310 class names. Remove trailing comma after "ALL_REGS".
26312 2013-03-16 Jan Hubicka <jh@suse.cz>
26314 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
26315 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
26316 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
26317 of cgraph_get_create_node.
26318 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
26320 2013-03-16 Jason Merrill <jason@redhat.com>
26323 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
26324 with DW_AT_default_value.
26326 2013-03-16 Jakub Jelinek <jakub@redhat.com>
26328 * BASE-VER: Set to 4.9.0.
26330 2013-03-14 Andi Kleen <ak@linux.intel.com>
26333 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
26334 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
26335 Document _x* TSX intrinsics.
26337 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
26338 David Holsgrove <david.holsgrove@xilinx.com>
26340 * configure.ac: Add MicroBlaze TLS support detection.
26341 * configure: Regenerate.
26342 * config/microblaze/microblaze-protos.h
26343 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
26344 symbol_mentioned_p, label_mentioned_p): Add prototypes.
26345 * config/microblaze/microblaze.c (microblaze_address_type): Add
26346 ADDRESS_TLS and tls_reloc address types.
26347 (microblaze_address_info): Add tls_reloc.
26348 (TARGET_HAVE_TLS): Define.
26349 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
26350 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
26351 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
26352 load_tls_operand, microblaze_call_tls_get_addr,
26353 microblaze_legitimize_tls_address): New functions.
26354 (microblaze_classify_unspec): Handle UNSPEC_TLS.
26355 (get_base_reg): Use microblaze_tls_symbol_p.
26356 (microblaze_classify_address): Handle TLS.
26357 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
26358 label_mentioned_p and microblaze_tls_referenced_p.
26359 (microblaze_legitimize_address): Handle TLS.
26360 (microblaze_address_insns): Handle ADDRESS_TLS.
26361 (pic_address_needs_scratch): Handle TLS.
26362 (print_operand_address): Handle TLS.
26363 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
26364 (microblaze_expand_move): Handle TLS.
26365 (microblaze_legitimate_constant_p): Check
26366 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
26367 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
26368 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
26369 (PIC_OFFSET_TABLE_REGNUM): Set.
26370 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
26371 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
26372 (addsi3, movsi_internal2, movdf_internal): Update constraints
26373 * config/microblaze/predicates.md (arith_plus_operand): Define
26374 (move_operand): Redefine as move_src_operand,
26375 check microblaze_tls_referenced_p.
26377 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26379 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
26380 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
26382 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26384 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
26387 2013-03-14 Jakub Jelinek <jakub@redhat.com>
26389 PR tree-optimization/53265
26390 * common.opt (Waggressive-loop-optimizations): New option.
26391 * tree-ssa-loop-niter.c: Include tree-pass.h.
26392 (do_warn_aggressive_loop_optimizations): New function.
26393 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
26394 if number_of_latch_executions returned constant.
26395 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
26396 early. If number_of_latch_executions returned constant, set
26397 nb_iterations_upper_bound back to it.
26398 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
26400 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
26401 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
26403 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
26404 (MULTILIB_OSDIRNAMES): Set.
26405 * genmultilib: If defaultosdirname doesn't start with :: , set
26406 defaultosdirname2 instead, clear it and emit two . multilib_raw
26407 entries instead of just one.
26409 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26411 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26412 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26413 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26414 (SUBTARGET_OVERRIDE_OPTIONS): New.
26416 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26419 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26420 (musermode): Convert to Var(TARGET_USERMODE).
26421 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26422 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26423 * config/sh/sh.c (sh_option_override): Use
26424 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26425 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26427 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26429 (udivsi3_i4_single, divsi3_i4_single): Use
26430 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26432 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26434 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26437 2013-03-13 Richard Biener <rguenther@suse.de>
26439 PR tree-optimization/56608
26440 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26441 calls when vectorizing basic-blocks.
26443 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26446 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26449 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26451 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26453 2013-03-11 Jan Hubicka <jh@suse.cz>
26456 * lto-streamer-out.c (output_symbol_p): Skip references from
26457 constructors of external variables.
26459 2013-03-11 Jan Hubicka <jh@suse.cz>
26461 PR middle-end/56571
26462 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26464 * emit-rtl.c (verify_rtx_sharing): Likewise.
26465 (copy_insn_1): Likewise.
26466 * rtl.c (copy_rtx): Likewise.
26468 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26471 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26472 output_operand_lossage message.
26474 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26477 * arm.c (shift_op): Validate RTL pattern on the fly.
26478 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26481 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26484 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26485 new call patterns that clobber %fr12.
26486 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26487 split and postreload patterns.
26488 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26489 registers %fr12 and %fr12R as call used.
26491 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26493 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26494 (canon_address, record_store, replace_read, check_mem_read_rtx,
26495 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26496 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26497 rest_of_handle_dse): Likewise.
26499 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26501 PR middle-end/56524
26502 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26504 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26505 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26506 (save_optabs_if_changed): Replace with...
26507 (init_tree_optimization_optabs): ...this.
26508 * optabs.c (save_optabs_if_changed): Rename to...
26509 (init_tree_optimization_optabs): ...this. Take the optimization node
26510 as argument. Do nothing if the base optabs are already correct.
26511 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26512 to recompute optabs.
26513 * function.h (function): Remove optabs field.
26514 * function.c (invoke_set_current_function_hook): Call
26515 init_tree_optimization_optabs. Use the result to initialize
26518 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26520 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26521 if GTMA_HAS_NO_INSTRUMENTATION.
26522 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26523 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26524 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26525 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26526 GTMA_HAS_NO_INSTRUMENTATION.
26528 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26530 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26533 2013-03-08 Marek Polacek <polacek@redhat.com>
26534 Jakub Jelinek <jakub@redhat.com>
26536 PR tree-optimization/56478
26537 * predict.c (is_comparison_with_loop_invariant_p): Change the
26538 type of loop_step to tree.
26539 (predict_loops): Adjust.
26540 (predict_iv_comparison): Perform the computations on double_ints.
26542 2013-03-08 Richard Biener <rguenther@suse.de>
26544 PR tree-optimization/56570
26545 * tree-cfg.c (verify_expr_location_1): Verify locations for
26547 * tree-sra.c (create_access_replacement): Strip locations
26548 from DECL_DEBUG_EXPRs.
26550 2013-03-08 Richard Biener <rguenther@suse.de>
26552 * tree-inline.c (expand_call_inline): Do not associate
26553 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26554 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
26556 2013-03-08 Richard Biener <rguenther@suse.de>
26558 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
26559 or block changes with -Og. Fix for location / block encoding
26560 changes and PHI arguments with locations.
26562 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
26564 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
26566 (struct output_info): Likewise.
26567 (register_overhead): Remove bad gcc_assert.
26568 (bitmap_find_bit): If there is only a single bitmap element, do not
26569 count a miss as a search.
26570 (print_statistics): Update for counter type changes.
26571 (dump_bitmap_statistics): Likewise. Print headers such that they
26572 are properly lined up with the printed counters.
26574 2013-03-07 Jakub Jelinek <jakub@redhat.com>
26576 PR tree-optimization/56559
26577 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
26578 check that it has only a single use.
26580 2013-03-07 Richard Biener <rguenther@suse.de>
26582 * doc/invoke.texi (fwhole-program): Discourage use in combination
26585 2013-03-06 Jakub Jelinek <jakub@redhat.com>
26587 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
26589 PR tree-optimization/56539
26590 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
26591 instead of GSI_CONTINUE_LINKING as last argument to
26592 force_gimple_operand_gsi. Adjust function comment.
26594 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
26597 PR middle-end/56548
26598 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
26599 promoted mode, convert the result back to the original mode.
26601 2013-03-06 Richard Biener <rguenther@suse.de>
26603 PR middle-end/56294
26604 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26605 (insert_updated_phi_nodes_compare_uids): New function.
26606 (update_ssa): Sort symbols_to_rename after UID before
26607 traversing it to insert PHI nodes.
26609 2013-03-06 Richard Biener <rguenther@suse.de>
26611 PR middle-end/50494
26612 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26613 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26616 2013-02-13 Richard Biener <rguenther@suse.de>
26619 * varasm.c (output_constant_def_1): Get the decl representing
26620 the constant as argument.
26621 (output_constant_def): Wrap output_constant_def_1.
26622 (make_decl_rtl): Use output_constant_def_1 with the decl
26623 representing the constant.
26624 (build_constant_desc): Optionally re-use a decl already
26625 representing the constant.
26626 (tree_output_constant_def): Adjust.
26628 2013-03-06 Joey Ye <joey.ye@arm.com>
26631 * gcc.c (convert_white_space): New function.
26632 (main): Handles white space in function name.
26634 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26637 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26638 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26639 to SH_DIV_CALL_TABLE for TARGET_SH2.
26640 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26642 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26643 call-table options.
26645 2013-03-05 Sterling Augustine <saugustine@google.com>
26646 Cary Coutant <ccoutant@google.com>
26649 * dwarf2out.c (resolve_addr): Don't call
26650 remove_loc_list_addr_table_entries a second time for the same
26653 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26656 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26657 (avoid_complex_debug_insns): New function.
26658 (expand_debug_locations): Call it.
26660 PR rtl-optimization/56484
26661 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26662 lifetimes of hard registers on small register class machines.
26664 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26666 * config/microblaze/microblaze-protos.h: Rename
26667 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26668 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26670 (microblaze_fast_interrupt_function_p): New function.
26671 (microblaze_is_interrupt_handler): Rename to
26672 microblaze_is_interrupt_variant and add fast_interrupt check.
26673 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26674 (save_restore_insns): Likewise.
26675 (compute_frame_size): Likewise.
26676 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26677 (microblaze_globalize_label): Likewise.
26678 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26679 * config/microblaze/microblaze.md: Use wrapper
26680 microblaze_is_interrupt_variant.
26682 2013-03-05 Kai Tietz <ktietz@redhat.com>
26684 * sdbout.c (sdbout_one_type): Switch to current function's section
26685 supporting cold/hot.
26687 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26689 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26692 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26694 PR middle-end/56461
26695 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26696 if VALGRIND_GET_VBITS is defined, temporarily make object
26697 memory all defined, and restore previous valgrind addressability
26698 and definability afterwards. Free this_object at the end.
26700 PR middle-end/56461
26701 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26702 right before calling lra_create_live_ranges, also call it
26703 when clearing live_p. Only call lra_clear_live_ranges
26704 at the end if live_p.
26706 PR middle-end/56461
26707 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26709 2013-03-05 Richard Biener <rguenther@suse.de>
26711 PR tree-optimization/56521
26712 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26715 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26718 * except.h (remove_unreachable_eh_regions): New prototype.
26719 * except.c (remove_eh_handler_splicer): New function, split out
26720 of remove_eh_handler.
26721 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26722 warning about running it on many EH regions one at a time.
26723 (remove_unreachable_eh_regions_worker): New function, walk the
26724 EH tree in depth-first order and remove non-marked regions.
26725 (remove_unreachable_eh_regions): New function.
26726 * tree-eh.c (mark_reachable_handlers): New function, split out
26727 from remove_unreachable_handlers.
26728 (remove_unreachable_handlers): Use mark_reachable_handlers and
26729 remove_unreachable_eh_regions.
26730 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26731 and remove_unreachable_eh_regions.
26733 2013-03-05 Richard Biener <rguenther@suse.de>
26735 PR middle-end/56525
26736 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26737 not freeing them until the end.
26739 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26741 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26743 2013-03-05 Richard Biener <rguenther@suse.de>
26745 PR tree-optimization/56270
26746 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26747 of loads after scheduling an SLP instance.
26749 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26751 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26753 (check_gcc_parallelize): Run guality.exp as a separate job from
26754 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26755 struct-layout-1.exp with stackalign.exp.
26757 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26759 PR middle-end/56461
26760 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26761 load_index sbitmap even if some bit in it isn't set.
26763 PR middle-end/56461
26764 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26765 (discover_iteration_bound_by_body_walk): Change queues to
26766 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26767 spelling in comment. Call safe_push on queues[bound_index] directly.
26768 Release queues[queue_index] in every iteration unconditionally.
26769 Release bounds vector.
26771 PR middle-end/56461
26772 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26773 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26774 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26777 2013-03-05 Richard Biener <rguenther@suse.de>
26780 * tree-inline.c (remap_blocks_to_null): New function.
26781 (expand_call_inline): When expanding a call stmt without
26782 an associated block inline remap all callee blocks to NULL.
26784 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26786 PR rtl-optimization/56494
26787 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26788 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26789 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26791 PR middle-end/56461
26792 * sel-sched-ir.c (free_sched_pools): Release
26793 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26794 if succs_info_pool.max_top isn't -1.
26797 * opts.c (opts_obstack, opts_concat): Moved to...
26798 * opts-common.c (opts_obstack, opts_concat): ... here.
26800 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26802 PR middle-end/56461
26803 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26805 2013-03-04 Martin Jambor <mjambor@suse.cz>
26807 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26808 all appropriate places.
26810 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26812 PR tree-optimization/56424
26813 * ipa-split.c (split_function): Do not set the RSO flag if result is
26814 not by reference and its type is a register type.
26816 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26818 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26819 (microblaze_legitimate_pic_operand): Likewise
26820 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26821 new function microblaze_legitimate_pic_operand
26822 * config/microblaze/microblaze-protos.h
26823 (microblaze_legitimate_pic_operand): Declare.
26825 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26827 * config/microblaze/predicates.md (call_insn_simple_operand):
26828 New predicate for supported rtx code types.
26829 * config/microblaze/microblaze.md (call_internal1): Use
26830 call_insn_simple_operand predicate.
26832 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26834 PR middle-end/56461
26835 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26836 partitions.ordered_remove.
26838 PR middle-end/56461
26839 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26840 vec_oprnds0.create (1) for modifier == NONE.
26842 PR middle-end/56461
26843 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26844 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26845 vec_oprnds1 right before pushing anything to it for
26848 PR middle-end/56461
26849 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26850 set nbbs to 0 instead of having separate code path.
26851 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26852 instead of false as last argument if returning NULL.
26854 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26856 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26857 the attribute is now called "target" instead of "option".
26858 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26859 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26860 attribute/pragma name for TARGET_OPTION_VALID_P and
26861 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26862 * doc/tm.texi: Regenerated.
26864 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26866 * config/microblaze/microblaze.c:
26867 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26868 * config/microblaze/microblaze.h: Add -mxl-reorder to
26870 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26871 instructions emitted if TARGET_REORDER.
26872 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26873 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26876 2013-03-01 Xinliang David Li <davidxl@google.com>
26878 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26881 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26883 PR middle-end/56461
26884 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26885 vector even when returning true. Fix up function comment formatting.
26887 PR middle-end/56461
26888 * ira-build.c (ira_loop_nodes_count): New variable.
26889 (create_loop_tree_nodes): Initialize it.
26890 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26892 PR middle-end/56461
26893 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26894 method on dr_chain and result_chain.
26895 * tree-vect-stmts.c (vectorizable_store): Only call
26896 result_chain.create if j == 0.
26898 PR middle-end/56461
26899 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26900 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26901 before overwriting it.
26903 2013-03-01 Tobias Burnus <burnus@net-b.de>
26905 * doc/extended.texi (C Extensions): Change order in @menu
26907 (Other MIPS Built-in Functions): Move last MIPS entry before
26908 "picoChip Built-in Functions".
26909 (SH Built-in Functions): Move after RX Built-in Functions.
26910 * doc/gcc.texi (Introduction): Change order in @menu
26912 * doc/md.texi (Constraints): Ditto.
26913 * gty.texi (Type Information): Ditto.
26914 (User-provided marking routines for template types): Make
26916 * doc/invoke.texi (AArch64 Options): Move before
26917 "Adapteva Epiphany Options".
26919 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26920 Jakub Jelinek <jakub@redhat.com>
26923 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26924 no_address_safety_analysis attribute.
26925 * doc/extend.texi (no_address_safety_attribute): Rename to
26926 no_sanitize_address attribute, mention no_address_safety_analysis
26927 attribute as deprecated alias.
26929 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26931 PR middle-end/56461
26932 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26933 type to vec<vec<tree> > *.
26934 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26935 to be vec<tree> instead of vec<tree> *, set vec_defs
26936 to vNULL and call vec_defs.create (number_of_vects), adjust other
26938 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26939 vectorizable_condition): Adjust vect_get_slp_defs callers.
26941 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26943 * config/aarch64/aarch64.c
26944 (aarch64_float_const_representable): Remove unused variable.
26946 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26948 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26950 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26952 * config/aarch64/aarch64-builtins.c
26953 (aarch64_init_simd_builtins): Make static.
26955 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26957 * config/aarch64/aarch64.c
26958 (aarch64_simd_make_constant): Make static.
26960 2013-02-28 Martin Jambor <mjambor@suse.cz>
26962 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26963 with no initialization to the RHS of debug statements.
26965 2013-02-28 Martin Jambor <mjambor@suse.cz>
26967 PR tree-optimization/56294
26968 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26970 (get_access_replacement): Do not call create_access_replacement.
26971 Assert a replacement exists.
26972 (get_repl_default_def_ssa_name): Create the replacement declaration
26975 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26977 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26978 final_end_function.
26980 2013-02-28 Marek Polacek <polacek@redhat.com>
26982 PR rtl-optimization/56466
26983 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26984 if we're changing a loop.
26985 (peel_loops_completely): Likewise.
26987 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26990 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26992 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26995 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26996 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26997 INTX_FTYPE_FX, FX_FTYPE_INTX.
26998 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
27000 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27002 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
27003 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
27004 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
27005 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
27006 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
27007 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
27008 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
27009 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
27010 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
27011 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
27012 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
27013 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
27014 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
27015 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
27016 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
27017 (avrxmega6): Increase max flash segments from 5 to 6.
27018 * config/avr/t-multilib: Regenerate.
27019 * config/avr/avr-tables.opt: Regenerate.
27020 * doc/avr-mmcu.texi: Regenerate.
27022 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27024 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
27025 (avr_device_to_arch): Rename to avr_device_to_ld.
27026 (avr_device_to_as): New prototype.
27027 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
27028 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
27029 * config/avr/driver-avr.c (avr_device_to_as): New.
27030 (avr_device_to_arch): Rename to avr_device_to_ld.
27032 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27034 PR middle-end/56461
27035 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
27036 method on dr_chain and result_chain.
27038 PR middle-end/56461
27039 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
27040 pointer_set_destroy on not_executed_last_iteration.
27042 PR middle-end/56461
27043 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
27045 PR middle-end/56461
27046 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
27047 FOR_EACH_DEFINED_FUNCTION when freeing state.
27049 PR middle-end/56461
27050 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
27052 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
27055 PR middle-end/56461
27056 * ipa-cp.c (decide_whether_version_node): Call vec_free on
27057 known_aggs[i].items and release known_aggs vector.
27059 PR middle-end/56461
27060 * ipa-reference.c (propagate): Free node_info even for alias nodes.
27062 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27064 * config/microblaze/microblaze.c (microblaze_emit_compare):
27065 Use xor for EQ/NE comparisions.
27066 * config/microblaze/microblaze.md (cstoresf4): Add constraints
27067 (cbranchsf4): Adjust operator to comparison_operator.
27069 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27071 PR middle-end/56461
27072 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
27074 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
27075 vec_safe_push, always update *slot.
27076 (redirect_edge_var_map_clear): Use vec_free.
27077 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
27078 (free_var_map_entry): Use vec_free.
27079 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
27080 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
27082 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
27084 PR middle-end/45472
27085 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
27086 when the may_trap_p bit of the exprs being merged differs.
27087 Reorder tests for speculativeness in the logical and operator.
27089 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27091 * incpath.c (add_standard_paths): Use reconcat instead of concat
27092 where appropriate and avoid leaking memory.
27094 * opts.h: Include obstack.h.
27095 (opts_concat): New prototype.
27096 (opts_obstack): New declaration.
27097 * opts.c (opts_concat): New function.
27098 (opts_obstack): New variable.
27099 (init_options_struct): Call gcc_init_obstack on opts_obstack.
27100 (finish_options): Use opts_concat instead of concat
27101 and XOBNEWVEC instead of XNEWVEC.
27102 * opts-common.c (generate_canonical_option, decode_cmdline_option,
27103 generate_option): Likewise.
27104 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
27105 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
27108 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
27109 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
27111 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27113 PR middle-end/56461
27114 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
27116 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27118 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
27119 (arm_block_move_unaligned_straight): Likewise.
27120 (arm_adjust_block_mem): Likewise.
27122 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27125 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
27126 temp, cond and label.
27127 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
27130 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
27131 * config/c6x/c6x.h (dbx_register_map): Update declaration.
27134 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
27135 of prologue/epilogue functions.
27138 * config/tilegx/tilegx.c (tilegx_expand_prologue):
27139 Remove unused variable cfa_offset.
27140 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
27143 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
27144 type promotion to unsigned.
27147 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
27148 for HOST_WIDE_INT of 32 bit / same size as int.
27149 (arm_block_move_unaligned_straight): Likewise.
27150 (arm_adjust_block_mem): Likewise.
27153 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
27156 2013-02-26 Marek Polacek <polacek@redhat.com>
27158 PR tree-optimization/56426
27159 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
27161 2013-02-26 Richard Biener <rguenther@suse.de>
27164 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
27165 unused variable loops.
27167 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27169 PR tree-optimization/56448
27170 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
27171 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
27172 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
27173 later operands of the references, or even first operand for
27174 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
27176 PR tree-optimization/56443
27177 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
27178 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
27179 to type_for_mode langhook.
27181 2013-02-25 Matt Turner <mattst88@gmail.com>
27183 * doc/invoke.texi: Document r4700.
27185 2013-02-25 Richard Biener <rguenther@suse.de>
27187 PR tree-optimization/56175
27188 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
27190 (simplify_bitwise_binary): ... here. Also guard the conversion
27191 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
27193 2013-02-25 Catherine Moore <clm@codesourcery.com>
27196 2013-02-24 Catherine Moore <clm@codesourcery.com>
27197 Maciej W. Rozycki <macro@codesourcery.com>
27198 Tom de Vries <tom@codesourcery.com>
27199 Nathan Sidwell <nathan@codesourcery.com>
27200 Iain Sandoe <iain@codesourcery.com>
27201 Nathan Froyd <froydnj@codesourcery.com>
27202 Chao-ying Fu <fu@mips.com>
27204 * doc/extend.texi: (micromips, nomicromips, nocompression):
27205 Document new function attributes.
27206 * doc/invoke.texi (minterlink-compressed, mmicromips,
27207 m14k, m14ke, m14kec): Document new options.
27208 (minterlink-mips16): Update documentation.
27209 * doc/md.texi (ZC, ZD): Document new constraints.
27210 * configure.ac (gcc_cv_as_micromips): Check if linker
27211 supports the .set micromips directive.
27212 * configure: Regenerate.
27213 * config.in: Regenerate.
27214 * config/mips/mips-tables.opt: Regenerate.
27215 * config/mips/micromips.md: New file.
27216 * constraints.md (ZC, AD): New constraints.
27217 * config/mips/predicates.md (movep_src_register): New predicate.
27218 (movep_src_operand): New predicate.
27219 (non_volatile_mem_operand): New predicate.
27220 * config/mips/mips.md (multimem): New type.
27221 (length): Differentiate between 17-bit and 18-bit branch offsets.
27222 (MOVEP1, MOVEP2): New mode iterator.
27223 (mov_<load>l): Use ZC constraint.
27224 (mov_<load>r): Likewise.
27225 (mov_<store>l): Likewise.
27226 (mov_<store>r): Likewise.
27227 (*branch_equality<mode>_inverted): Add microMIPS support.
27228 (*branch_equality<mode>): Likewise.
27229 (*jump_absolute): Likewise.
27230 (indirect_jump_<mode>): Likewise.
27231 (tablejump_<mode>): Likewise.
27232 (<optab>_internal): Likewise.
27233 (sibcall_internal): Likewise.
27234 (sibcall_value_internal): Likewise.
27235 (prefetch): Use constraint ZD.
27236 * config/mips/mips.opt (minterlink-compressed): New option.
27237 (minterlink-mips16): Now an alias for minterlink-compressed.
27238 (mmicromips): New option.
27239 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27240 (compare_and_swap_12): Likewise.
27241 (sync_add<mode>): Likewise.
27242 (sync_<optab>_12): Likewise.
27243 (sync_old_<optab>_12): Likewise.
27244 (sync_new_<optab>_12): Likewise.
27245 (sync_nand_12): Likewise.
27246 (sync_old_nand_12): Likewise.
27247 (sync_new_nand_12): Likewise.
27248 (sync_sub<mode>): Likewise.
27249 (sync_old_add<mode>): Likewise.
27250 (sync_old_sub<mode>): Likewise.
27251 (sync_new_add<mode>): Likewise.
27252 (sync_new_sub<mode>): Likewise.
27253 (sync_<optab><mode>): Likewise.
27254 (sync_old_<optab><mode>): Likewise.
27255 (sync_new_<optab><mode>): Likewise.
27256 (sync_nand<mode>): Likewise.
27257 (sync_old_nand<mode>): Likewise.
27258 (sync_new_nand<mode>): Likewise.
27259 (sync_lock_test_and_set<mode>): Likewise.
27260 (test_and_set_12): Likewise.
27261 (atomic_compare_and_swap<mode>): Likewise.
27262 (atomic_exchange<mode>_llsc): Likewise.
27263 (atomic_fetch_add<mode>_llsc): Likewise.
27264 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27265 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27266 (umips_save_restore_pattern_p): Likewise.
27267 (umips_load_store_pair_p): Likewise.
27268 (umips_output_load_store_pair): Likewise.
27269 (umips_movep_target_p): Likewise.
27270 (umips_12bit_offset_address_p): Likewise.
27271 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27272 (mips_base_mips16): Rename this...
27273 (mips_base_compression_flags): ...to this. Update all uses.
27274 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27275 (mips_mips16_decl_p): Delete.
27276 (mips_nomips16_decl_p): Delete.
27277 (mips_get_compress_on_flags): New function.
27278 (mips_get_compress_off_flags): New function.
27279 (mips_get_compress_mode): New function.
27280 (mips_get_compress_on_name): New function.
27281 (mips_get_compress_off_name): New function.
27282 (mips_insert_attributes): Support multiple compression types.
27283 (mips_merge_decl_attributes): Likewise.
27284 (umips_12bit_offset_address_p): New function.
27285 (mips_start_function_definition): Emit .set micromips directive.
27286 (mips_call_may_need_jalx_p): New function.
27287 (mips_function_ok_for_sibcall): Add microMIPS support.
27288 (mips_print_operand_punctuation): Support short delay slots and
27290 (umips_swm_mask, umips_swm_encoding): New.
27291 (umips_build_save_restore): New function.
27292 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27293 (was_mips16_p): Remove.
27294 (old_compression_mode): New.
27295 (mips_set_compression_mode): New function.
27296 (mips_set_current_function): Add microMIPS support.
27297 (mips_option_override): Likewise.
27298 (umips_save_restore_pattern_p): New function.
27299 (umips_output_save_restore): New function.
27300 (umips_load_store_pair_p_1): New function.
27301 (umips_load_store_pair_p): New function.
27302 (umips_output_load_store_pair_1): New function.
27303 (umips_output_load_store_pair): New function.
27304 (umips_movep_target_p) New function.
27305 (mips_prepare_pch_save): Add microMIPS support.
27306 * config/mips/mips.h (TARGET_COMPRESSION): New.
27307 (TARGET_CPU_CPP_BUILTINS): Update macro
27308 to use new compression flags and to support microMIPS.
27309 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27310 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27311 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27312 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27313 (ASM_SPEC): Support mmicromips and mno-micromips.
27314 (M16STORE_REG_P): New macro.
27315 (MIPS_CALL): Support TARGET_MICROMIPS.
27316 (MICROMIPS_J): New macro.
27317 (mips_base_mips16): Rename this...
27318 (mips_base_compression_flags): ...to this.
27319 (UMIPS_12BIT_OFFSET_P): New macro.
27320 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27321 (MULTILIB_DIRNAMES): Likewise.
27323 2013-02-25 Tom de Vries <tom@codesourcery.com>
27325 PR rtl-optimization/56131
27326 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
27327 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
27328 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
27330 2013-02-25 Tobias Burnus <burnus@net-b.de>
27332 * doc/invoke.texi (-fsanitize=): Move from optimization
27333 to debugging options.
27335 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27337 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
27339 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27340 Alexander Monakov <amonakov@ispras.ru>
27342 PR middle-end/56077
27343 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
27344 flush pending lists also on non-jumps. Adjust comment.
27346 2013-02-24 Catherine Moore <clm@codesourcery.com>
27347 Maciej W. Rozycki <macro@codesourcery.com>
27348 Tom de Vries <tom@codesourcery.com>
27349 Nathan Sidwell <nathan@codesourcery.com>
27350 Iain Sandoe <iain@codesourcery.com>
27351 Nathan Froyd <froydnj@codesourcery.com>
27352 Chao-ying Fu <fu@mips.com>
27354 * doc/extend.texi: (micromips, nomicromips, nocompression):
27355 Document new function attributes.
27356 * doc/invoke.texi (minterlink-compressed, mmicromips,
27357 m14k, m14ke, m14kec): Document new options.
27358 (minterlink-mips16): Update documentation.
27359 * doc/md.texi (ZC, ZD): Document new constraints.
27360 * configure.ac (gcc_cv_as_micromips): Check if linker
27361 supports the .set micromips directive.
27362 * configure: Regenerate.
27363 * config.in: Regenerate.
27364 * config/mips/mips-tables.opt: Regenerate.
27365 * config/mips/micromips.md: New file.
27366 * constraints.md (ZC, AD): New constraints.
27367 * config/mips/predicates.md (movep_src_register): New predicate.
27368 (movep_src_operand): New predicate.
27369 (non_volatile_mem_operand): New predicate.
27370 * config/mips/mips.md (multimem): New type.
27371 (length): Differentiate between 17-bit and 18-bit branch offsets.
27372 (MOVEP1, MOVEP2): New mode iterator.
27373 (mov_<load>l): Use ZC constraint.
27374 (mov_<load>r): Likewise.
27375 (mov_<store>l): Likewise.
27376 (mov_<store>r): Likewise.
27377 (*branch_equality<mode>_inverted): Add microMIPS support.
27378 (*branch_equality<mode>): Likewise.
27379 (*jump_absolute): Likewise.
27380 (indirect_jump_<mode>): Likewise.
27381 (tablejump_<mode>): Likewise.
27382 (<optab>_internal): Likewise.
27383 (sibcall_internal): Likewise.
27384 (sibcall_value_internal): Likewise.
27385 (prefetch): Use constraint ZD.
27386 * config/mips/mips.opt (minterlink-compressed): New option.
27387 (minterlink-mips16): Now an alias for minterlink-compressed.
27388 (mmicromips): New option.
27389 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27390 (compare_and_swap_12): Likewise.
27391 (sync_add<mode>): Likewise.
27392 (sync_<optab>_12): Likewise.
27393 (sync_old_<optab>_12): Likewise.
27394 (sync_new_<optab>_12): Likewise.
27395 (sync_nand_12): Likewise.
27396 (sync_old_nand_12): Likewise.
27397 (sync_new_nand_12): Likewise.
27398 (sync_sub<mode>): Likewise.
27399 (sync_old_add<mode>): Likewise.
27400 (sync_old_sub<mode>): Likewise.
27401 (sync_new_add<mode>): Likewise.
27402 (sync_new_sub<mode>): Likewise.
27403 (sync_<optab><mode>): Likewise.
27404 (sync_old_<optab><mode>): Likewise.
27405 (sync_new_<optab><mode>): Likewise.
27406 (sync_nand<mode>): Likewise.
27407 (sync_old_nand<mode>): Likewise.
27408 (sync_new_nand<mode>): Likewise.
27409 (sync_lock_test_and_set<mode>): Likewise.
27410 (test_and_set_12): Likewise.
27411 (atomic_compare_and_swap<mode>): Likewise.
27412 (atomic_exchange<mode>_llsc): Likewise.
27413 (atomic_fetch_add<mode>_llsc): Likewise.
27414 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27415 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27416 (umips_save_restore_pattern_p): Likewise.
27417 (umips_load_store_pair_p): Likewise.
27418 (umips_output_load_store_pair): Likewise.
27419 (umips_movep_target_p): Likewise.
27420 (umips_12bit_offset_address_p): Likewise.
27421 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27422 (mips_base_mips16): Rename this...
27423 (mips_base_compression_flags): ...to this. Update all uses.
27424 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27425 (mips_mips16_decl_p): Delete.
27426 (mips_nomips16_decl_p): Delete.
27427 (mips_get_compress_on_flags): New function.
27428 (mips_get_compress_off_flags): New function.
27429 (mips_get_compress_mode): New function.
27430 (mips_get_compress_on_name): New function.
27431 (mips_get_compress_off_name): New function.
27432 (mips_insert_attributes): Support multiple compression types.
27433 (mips_merge_decl_attributes): Likewise.
27434 (umips_12bit_offset_address_p): New function.
27435 (mips_start_function_definition): Emit .set micromips directive.
27436 (mips_call_may_need_jalx_p): New function.
27437 (mips_function_ok_for_sibcall): Add microMIPS support.
27438 (mips_print_operand_punctuation): Support short delay slots and
27440 (umips_swm_mask, umips_swm_encoding): New.
27441 (umips_build_save_restore): New function.
27442 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27443 (was_mips16_p): Remove.
27444 (old_compression_mode): New.
27445 (mips_set_compression_mode): New function.
27446 (mips_set_current_function): Add microMIPS support.
27447 (mips_option_override): Likewise.
27448 (umips_save_restore_pattern_p): New function.
27449 (umips_output_save_restore): New function.
27450 (umips_load_store_pair_p_1): New function.
27451 (umips_load_store_pair_p): New function.
27452 (umips_output_load_store_pair_1): New function.
27453 (umips_output_load_store_pair): New function.
27454 (umips_movep_target_p) New function.
27455 (mips_prepare_pch_save): Add microMIPS support.
27456 * config/mips/mips.h (TARGET_COMPRESSION): New.
27457 (TARGET_CPU_CPP_BUILTINS): Update macro
27458 to use new compression flags and to support microMIPS.
27459 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27460 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27461 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27462 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27463 (ASM_SPEC): Support mmicromips and mno-micromips.
27464 (M16STORE_REG_P): New macro.
27465 (MIPS_CALL): Support TARGET_MICROMIPS.
27466 (MICROMIPS_J): New macro.
27467 (mips_base_mips16): Rename this...
27468 (mips_base_compression_flags): ...to this.
27469 (UMIPS_12BIT_OFFSET_P): New macro.
27470 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27471 (MULTILIB_DIRNAMES): Likewise.
27473 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27476 * target-globals.c (save_target_globals): For init_reg_sets and
27477 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27479 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27481 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27482 * config/aarch64/t-aarch64
27483 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27485 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27487 PR inline-asm/56148
27488 * lra-constraints.c (process_alt_operands): Reload operand
27489 conflicting with earlier clobber only if no more other conflicting
27492 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27495 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27496 if not linking a shared library.
27498 2013-02-22 Seth LaForge <sethml@google.com>
27500 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27502 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27504 * config/arm/arm.md (split for extendsidi): Update condition.
27505 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27506 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27507 (qhs_zextenddi_cstr): Likewise.
27509 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27511 PR middle-end/56420
27512 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27513 avoid signed wrapping.
27514 (expand_mult): Handle properly multiplication by
27515 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27516 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27517 in the compiler if coeff is HOST_WIDE_INT_MIN.
27518 (expand_divmod): Don't make ext_op1 static, change it's type to
27519 uhwi. Avoid undefined behavior in -INTVAL (op1).
27521 PR rtl-optimization/50339
27522 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27524 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27525 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27526 into splitting_ashiftrt field.
27527 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27529 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27532 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27534 PR middle-end/56108
27535 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27536 are sure to go irrevocable.
27538 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27540 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27541 scalars are valid operands.
27543 2013-02-21 Martin Jambor <mjambor@suse.cz>
27545 PR tree-optimization/56310
27546 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27547 only matching indices and non-negative final offsets.
27548 (intersect_aggregates_with_edge): Pass src_idx to
27549 agg_replacements_to_vector. Pass src_idx insstead of index to
27550 intersect_with_agg_replacements.
27552 2013-02-21 Martin Jambor <mjambor@suse.cz>
27554 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
27555 instead of hard-wired defaults.
27557 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
27559 * doc/invoke.texi (MIPS Options): Update documentation of the
27560 floating-point multiply-accumulate instruction restrictions.
27562 2013-02-21 Kostya Serebryany <kcc@google.com>
27564 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
27565 asan_shadow_offset on x86_64 linux.
27567 2013-02-21 Richard Biener <rguenther@suse.de>
27569 PR tree-optimization/56415
27571 2013-02-11 Richard Biener <rguenther@suse.de>
27573 PR tree-optimization/56273
27574 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27577 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27580 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
27583 PR inline-asm/56405
27584 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
27585 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
27587 2013-02-20 Jan Hubicka <jh@suse.cz>
27589 PR tree-optimization/56265
27590 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
27591 when target is referenced for first time.
27593 2013-02-20 Richard Biener <rguenther@suse.de>
27595 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
27596 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
27597 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
27598 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
27599 not return anything.
27600 (rename_ssa_copies): Do not remove unused locals.
27601 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27602 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27603 * passes.c (execute_function_todo): Do not schedule unused locals
27604 removal if cleanup_tree_cfg did something.
27605 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27606 about the number of removed locals.
27608 2013-02-20 Richard Biener <rguenther@suse.de>
27610 PR tree-optimization/56398
27611 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27613 2013-02-20 Martin Jambor <mjambor@suse.cz>
27615 PR tree-optimization/55334
27616 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27617 restricted pointers to arrays.
27619 2013-02-20 Richard Biener <rguenther@suse.de>
27620 Jakub Jelinek <jakub@redhat.com>
27622 PR tree-optimization/56396
27623 * tree-ssa-ccp.c (n_const_val): New static variable.
27624 (get_value): Return NULL for SSA names we don't have a lattice
27626 (ccp_initialize): Initialize n_const_val.
27627 * tree-ssa-copy.c (n_copy_of): New static variable.
27628 (init_copy_prop): Initialize n_copy_of.
27629 (get_value): Return NULL_TREE for SSA names we don't have a
27632 2013-02-20 Martin Jambor <mjambor@suse.cz>
27634 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27636 2013-02-20 Richard Biener <rguenther@suse.de>
27638 * genpreds.c (write_lookup_constraint): Do not compare first
27639 letter of the constraint again.
27641 2013-02-20 Richard Biener <rguenther@suse.de>
27643 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27645 (get_use_iv_cost): Terminate hashtable walk when coming across
27648 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27650 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27651 reassociation for avx2 targets.
27653 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27655 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27656 Add version check for v8.10.a to enable microblaze_has_clz
27657 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27658 version and TARGET_PATTERN_COMPARE check
27659 * config/microblaze/microblaze.md: New clzsi2 instruction
27661 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27663 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27664 function before branching.
27666 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27668 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27670 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27672 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27674 PR middle-end/55889
27675 * sel-sched.c: Include ira.h.
27676 (implicit_clobber_conflict_p): New function.
27677 (moveup_expr): Use it.
27678 * Makefile.in (sel-sched.o): Depend on ira.h.
27680 2013-02-19 Richard Biener <rguenther@suse.de>
27682 PR tree-optimization/56384
27683 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27684 (vn_hash_type): Split out from ...
27685 (vn_hash_constant_with_type): ... here.
27686 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27687 (vn_phi_eq): Compare types from vn_phi_s structure.
27688 (vn_phi_lookup): Populate vn_phi_s type.
27689 (vn_phi_insert): Likewise.
27691 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27693 PR tree-optimization/56350
27694 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27695 if haven't found reduction or nested cycle operand, rather than
27696 asserting we must find it.
27698 PR tree-optimization/56381
27699 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27702 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27703 Jakub Jelinek <jakub@redhat.com>
27706 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27707 (swap_optab_enable): Same.
27708 (init_all_optabs): Use argument instead of global.
27709 * tree.h (struct tree_optimization_option): New field target_optabs.
27710 * expr.h (init_all_optabs): Add argument to prototype.
27711 (TREE_OPTIMIZATION_OPTABS): New.
27712 (save_optabs_if_changed): Protoize.
27713 * optabs.h: Declare this_fn_optabs.
27714 * optabs.c (save_optabs_if_changed): New.
27715 Declare this_fn_optabs.
27716 (init_optabs): Add argument to init_all_optabs() call.
27717 * function.c (invoke_set_current_function_hook): Handle per
27719 * function.h (struct function): New field optabs.
27720 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27721 optimization_current_node has changed.
27722 * target-globals.h (save_target_globals_default_opts): Protoize.
27723 * target-globals.c (save_target_globals_default_opts): New.
27725 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27728 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27729 registers %fr12 and %fr12R as call used.
27732 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27733 and HImode, require all displacements to be an integer multiple of
27735 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27736 only allow QImode and HImode when reload is in progress and strict is
27737 true. Likewise for symbolic addresses. Use base14_operand to check
27738 displacements in REG+BASE addresses.
27740 2013-02-18 Richard Biener <rguenther@suse.de>
27742 PR tree-optimization/56366
27743 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27744 handle sign-conversion of outer-loop initial induction value.
27746 2013-02-18 Richard Biener <rguenther@suse.de>
27748 PR middle-end/56349
27749 * cfghooks.c (merge_blocks): If we merge a latch into another
27750 block adjust references to it.
27751 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27752 (verify_loop_structure): Verify that a recorded latch is in fact
27755 2013-02-18 Richard Biener <rguenther@suse.de>
27757 PR tree-optimization/56321
27758 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27759 order SSA name release and virtual operand unlinking.
27761 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27763 * config/microblaze/microblaze.md (save_stack_block): Define.
27764 (restore_stack_block): Likewise.
27766 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27768 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27769 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27770 * config/microblaze/microblaze.c (microblaze_option_override):
27771 Bail out early for PIC modes when target does not support PIC.
27773 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27775 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27776 Replace with a microblaze version.
27777 (microblaze_trampoline_init): Adapt for microblaze.
27778 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27781 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27782 Dodji Seketeli <dodji@redhat.com>
27785 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27786 (instrument_mem_region_access): Do not forget to always put
27787 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27788 0) statement, even for cases where either 'base' or 'base + len'
27789 are not instrumented -- because they have been previously
27790 instrumented. Simplify the logic by putting all the statements
27791 instrument 'base + len' inside a sequence, and then insert that
27792 sequence right before the current insertion point. Then, to
27793 instrument 'base + len', just get an iterator on that statement.
27794 And do not forget to update the pointer to iterator the function
27795 received as argument.
27797 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27799 PR rtl-optimization/56348
27800 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27802 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27804 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27805 (clean_graph_dump_file): Pass base to start_graph_dump.
27807 2013-02-14 Richard Henderson <rth@redhat.com>
27810 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27812 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27814 * collect2-aix.h: Define F_LOADONLY.
27816 2013-02-14 Richard Biener <rguenther@suse.de>
27819 * varasm.c (output_constant_def_1): Get the decl representing
27820 the constant as argument.
27821 (output_constant_def): Wrap output_constant_def_1.
27822 (make_decl_rtl): Use output_constant_def_1 with the decl
27823 representing the constant.
27824 (build_constant_desc): Optionally re-use a decl already
27825 representing the constant.
27826 (tree_output_constant_def): Adjust.
27828 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27831 * asan.c (instrument_builtin_call): Really put the length of the
27832 second source argument into src1_len.
27834 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27836 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27837 argument. If it is false, don't create edge from then_bb to
27839 (insert_if_then_before_iter): Pass true to it.
27840 (build_check_stmt): Pass false to it.
27841 (transform_statements): Flush hash table only on extended basic
27842 block boundaries, rather than at the beginning of every bb.
27843 Don't flush hash table on nonfreeing_call_p calls.
27844 * tree-flow.h (nonfreeing_call_p): New prototype.
27845 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27847 2013-02-13 David S. Miller <davem@davemloft.net>
27849 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27851 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27854 * ira.c (max_regno_before_ira): Move from ...
27856 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27859 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27861 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27863 2013-02-13 Richard Biener <rguenther@suse.de>
27866 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27867 globals in MEM_REFs.
27869 2013-02-13 Richard Biener <rguenther@suse.de>
27871 * loop-init.c (loop_optimizer_init): Clear loop state when
27872 re-initializing preserved loops.
27873 * loop-unswitch.c (unswitch_single_loop): Return whether
27874 we unswitched the loop. Do not verify loop state here.
27875 (unswitch_loops): When we unswitched a loop discover new loops.
27877 2013-02-13 Kostya Serebryany <kcc@google.com>
27879 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27881 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27883 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27885 Avoid instrumenting duplicated memory access in the same basic block
27886 * Makefile.in (asan.o): Add new dependency on hash-table.h
27887 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27888 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27889 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27890 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27891 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27892 (get_mem_ref_of_assignment): New functions.
27893 (get_mem_refs_of_builtin_call): Extract from
27894 instrument_builtin_call and tweak a little bit to make it fit with
27896 (instrument_builtin_call): Use the new
27897 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27898 of is_gimple_builtin_call.
27899 (instrument_derefs, instrument_mem_region_access): Insert the
27900 instrumented memory reference into the hash table.
27901 (maybe_instrument_assignment): Renamed instrument_assignment into
27902 this, and change it to advance the iterator when instrumentation
27903 actually happened and return true in that case. This makes it
27904 homogeneous with maybe_instrument_assignment, and thus give a
27905 chance to callers to be more 'regular'.
27906 (transform_statements): Clear the memory reference hash table
27907 whenever we enter a new BB, when we cross a function call, or when
27908 we are done transforming statements. Use
27909 maybe_instrument_assignment instead of instrumentation. No more
27910 need to special case maybe_instrument_assignment and advance the
27911 iterator after calling it; it's now handled just like
27912 maybe_instrument_call. Update comment.
27914 2013-02-13 Richard Biener <rguenther@suse.de>
27916 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27917 Fix loop discovery code.
27919 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27921 PR inline-asm/56148
27922 * lra-constraints.c (process_alt_operands): Match early clobber
27923 operand with itself. Check conflicts with earlyclobber only if
27924 the operand is not reloaded. Prefer to reload conflicting operand
27925 if earlyclobber and matching operands are the same.
27927 2013-02-12 Richard Biener <rguenther@suse.de>
27930 * lto-streamer-out.c (write_symbol): Do not output symbols
27931 for hard register variables.
27933 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27936 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27937 (umulsidi3_insn, mulsidi3_insn): New insns.
27939 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27941 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27942 (struct tune_params): Add vec_costs field.
27943 * config/arm/arm.c (arm_builtin_vectorization_cost)
27944 (arm_add_stmt_cost): New functions.
27945 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27946 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27947 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27948 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27949 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27950 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27951 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27953 2013-02-12 Richard Biener <rguenther@suse.de>
27956 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27957 decls again if possible.
27959 2013-02-12 Richard Biener <rguenther@suse.de>
27961 PR middle-end/56288
27962 * tree-ssa.c (verify_ssa_name): Fix check, move
27963 SSA_NAME_IN_FREE_LIST check up.
27965 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27966 Steven Bosscher <steven@gcc.gnu.org>
27968 PR rtl-optimization/56151
27969 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27970 equal to op0 or op1, and last_insn pattern is CODE operation
27971 with MEM dest and one of the operands matches that MEM.
27973 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27975 * doc/extend.texi: Document Function Multiversioning and "default"
27976 parameter string to target attribute.
27977 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27978 target attribute parameter is "default".
27979 (ix86_compare_version_priority): Remove checks for target attribute.
27980 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27981 Remove check for target attribute equal to NULL. Add assert.
27982 (ix86_generate_version_dispatcher_body): Change error to sorry.
27984 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27985 Jack Howarth <howarth@bromo.med.uc.edu>
27986 Patrick Marlier <patrick.marlier@gmail.com>
27989 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27990 define ENDFILE_SPEC as TM_DESTRUCTOR.
27991 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27993 2013-02-11 Alexander Potapenko <glider@google.com>
27994 Jack Howarth <howarth@bromo.med.uc.edu>
27995 Jakub Jelinek <jakub@redhat.com>
27998 * config/darwin.c (cdtor_record): Rename ctor_record.
27999 (sort_cdtor_records): Rename sort_ctor_records.
28000 (finalize_dtors): New routine to sort destructors by
28001 priority before use in assemble_integer.
28002 (machopic_asm_out_destructor): Use finalize_dtors if needed.
28004 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
28006 PR rtl-optimization/56275
28007 * simplify-rtx.c (avoid_constant_pool_reference): Check that
28008 offset is non-negative and less than cmode size before
28009 calling simplify_subreg.
28011 2013-02-11 Richard Biener <rguenther@suse.de>
28013 PR tree-optimization/56264
28014 * cfgloop.h (fix_loop_structure): Adjust prototype.
28015 * loop-init.c (fix_loop_structure): Return the number of
28016 newly discovered loops.
28017 * tree-cfgcleanup.c (repair_loop_structures): When new loops
28018 are discovered, do a full loop-closed SSA rewrite.
28020 2013-02-11 Richard Biener <rguenther@suse.de>
28022 PR tree-optimization/56273
28023 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
28025 (check_array_ref): Fix missing newline in dumps.
28026 (search_for_addr_array): Likewise.
28028 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
28030 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
28032 2013-02-09 Jakub Jelinek <jakub@redhat.com>
28035 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
28037 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
28039 PR rtl-optimization/56246
28040 * lra-constraints.c (simplify_operand_subreg): Try to reuse
28042 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
28043 constraints are satisfied.
28045 2013-02-08 Jeff Law <law@redhat.com>
28048 * emit-rtl.c (reg_is_parm_p): New function.
28049 * regs.h (reg_is_parm_p): New prototype.
28050 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
28051 callee-clobbered registers.
28053 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
28056 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
28057 If there is no implicit builtin declaration, just return NULL.
28059 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28061 * config/i386/sse.md (FMAMODEM): New mode iterator.
28062 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
28063 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
28065 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28067 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
28068 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
28069 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
28071 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28073 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
28074 (microblaze*-*-elf): Likewise.
28075 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
28077 * config/microblaze/microblaze-c.c: Add builtin defines for
28078 _LITTLE_ENDIAN and _BIG_ENDIAN.
28079 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
28080 add to TARGET_DEFAULT flags.
28081 Expand ASM_SPEC and LINK_SPEC.
28082 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
28083 * config/microblaze/microblaze.md: Update extendsidi2 and
28084 movdi_internal instructions to use low-order / high-order reg
28086 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
28087 options and inversemask / mask of LITTLE_ENDIAN.
28088 * config/microblaze/t-microblaze: Expand multilib options to
28089 include mlittle-endian (le) and update exceptions patterns.
28091 2013-02-08 Jakub Jelinek <jakub@redhat.com>
28093 PR rtl-optimization/56195
28094 * lra-constraints.c (get_reload_reg): Don't reuse regs
28095 if they have smaller mode than requested, if they have
28096 wider mode than requested, try to return a SUBREG.
28098 PR tree-optimization/56250
28099 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
28100 if type is unsigned and code isn't MULT_EXPR.
28102 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28104 PR tree-optimization/56064
28105 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
28106 bits according to mode.
28107 * fixed-value.h (fixed_from_double_int)
28108 (const_fixed_from_double_int): Adjust comments.
28110 2013-02-08 Richard Biener <rguenther@suse.de>
28113 * lto-streamer.h (struct data_in): Remove current_file, current_line
28114 and current_col members.
28115 * lto-streamer-out.c (lto_output_location): Stream changed bits
28116 en-block for efficiency.
28117 * lto-streamer-in.c (clear_line_info): Remove.
28118 (lto_input_location): Cache current file, line and column
28119 globally via local statics. Read changed bits en-block.
28120 (input_function): Do not call clear_line_info.
28121 (lto_read_body): Likewise.
28122 (lto_input_toplevel_asms): Likewise.
28124 2013-02-08 Michael Matz <matz@suse.de>
28126 PR tree-optimization/52448
28127 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
28128 (nt_call_phase): New static.
28129 (add_or_mark_expr): Only mark accesses with newer phase than any
28131 (nonfreeing_call_p): New.
28132 (nt_init_block): Update nt_call_phase, mark blocks as visited.
28133 (nt_fini_block): Keep blocks marked as visited.
28134 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
28136 2013-02-08 Richard Biener <rguenther@suse.de>
28138 * ira.c (ira): Free broken dominator information.
28140 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28142 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
28144 2013-02-08 Marek Polacek <polacek@redhat.com>
28146 * cfgloop.c (verify_loop_structure): Add more checking of headers.
28148 2013-02-08 Richard Biener <rguenther@suse.de>
28150 PR middle-end/56181
28151 * cfgloop.h (flow_loops_find): Adjust.
28152 (bb_loop_header_p): Declare.
28153 * cfgloop.c (bb_loop_header_p): New function split out from ...
28154 (flow_loops_find): ... here. Adjust function signature,
28155 support incremental loop structure update.
28156 (verify_loop_structure): Cleanup. Verify a loop is a loop.
28157 * cfgloopmanip.c (fix_loop_structure): Move ...
28158 * loop-init.c (fix_loop_structure): ... here.
28159 (apply_loop_flags): Split out from ...
28160 (loop_optimizer_init): ... here.
28161 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
28162 in incremental mode, only remove dead loops here.
28164 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28167 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
28168 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
28169 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
28170 (*round<mode>3.libgcc): New insns for fixed-modes.
28171 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
28172 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
28173 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
28174 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
28175 implementations. Define to __builtin_avr_absFX,
28176 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
28177 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
28178 __builtin_avr_countlsFX, respectively.
28179 * config/avr/avr-c.c (target.h): Include it.
28180 (enum avr_builtin_id): New enum.
28181 (avr_resolve_overloaded_builtin): New static function.
28182 (avr_register_target_pragmas): Use it to set
28183 targetm.resolve_overloaded_builtin.
28184 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
28185 tree nodes used by DEF_BUILTIN.
28186 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
28187 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
28188 <AVR_BUILTIN_xxBITS>: Same.
28190 2013-02-08 Richard Biener <rguenther@suse.de>
28192 * cfgloop.c (verify_loop_structure): Properly handle
28193 a loop exiting to another loop header.
28194 * ira-int.h (ira_loops): Remove.
28195 * ira.c (ira_loops): Remove.
28196 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
28197 (do_reload): Use loop_optimizer_finalize.
28198 * ira-build.c (create_loop_tree_nodes): Use get_loops and
28199 number_of_loops to access the loop tree.
28200 (more_one_region_p): Likewise.
28201 (finish_loop_tree_nodes): Likewise.
28202 (rebuild_regno_allocno_maps): Likewise.
28203 (mark_loops_for_removal): Likewise.
28204 (mark_all_loops_for_removal): Likewise.
28205 (remove_unnecessary_regions): Likewise.
28206 (ira_build): Likewise.
28207 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
28209 2013-02-08 Richard Biener <rguenther@suse.de>
28211 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
28212 * ipa-pure-const.c (analyze_function): Avoid calling
28213 mark_irreducible_loops twice.
28214 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
28216 2013-02-07 David S. Miller <davem@davemloft.net>
28218 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
28220 * var-tracking.c (vt_add_function_parameter): Test the presence of
28221 HAVE_window_save properly and do not remap argument registers when
28222 we have a leaf function.
28224 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
28227 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
28229 * config/i386/i386.c (ix86_print_operand): Ditto.
28231 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28233 * lra-constraints.c (process_alt_operands): Fix recently added comment.
28235 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28237 PR rtl-optimization/56225
28238 * lra-constraints.c (process_alt_operands): Check that reload hard
28239 reg can hold value for strict_low_part.
28241 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28244 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
28245 dwarf2out_end_function.
28246 (in_first_function_p, maybe_at_text_label_p,
28247 first_loclabel_num_not_at_text_label): New variables.
28248 (dwarf2out_var_location): In the first function find out
28249 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
28250 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
28253 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
28255 PR rtl-optimization/56178
28256 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
28257 SUBREG of a register. Tidy up related block of code.
28258 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
28259 note if the source is a register or a SUBREG of a register.
28261 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28264 * config/rs6000/rs6000.md (ptrm): New mode attr.
28265 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
28266 call_value_indirect_aix<pttrsize>,
28267 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
28270 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28272 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
28273 if -bnortl. Convert to strcmp and strncmp.
28275 2013-02-07 Alan Modra <amodra@gmail.com>
28278 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
28279 addresses won't wrap when offsetting.
28280 (rs6000_secondary_reload): Provide secondary reloads needed for
28281 wrapping LO_SUM addresses.
28283 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
28285 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
28286 MACH, just __MACH__.
28288 2013-02-06 Richard Biener <rguenther@suse.de>
28290 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
28291 instead of calling fix_loop_structure.
28293 2013-02-06 Jakub Jelinek <jakub@redhat.com>
28295 PR middle-end/56217
28296 * omp-low.c (use_pointer_for_field): Return false if
28297 lower_send_shared_vars doesn't generate any copy-out code.
28299 2013-02-06 Tom de Vries <tom@codesourcery.com>
28301 PR rtl-optimization/56131
28302 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
28303 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
28304 of the label is NULL. Add comment.
28306 2013-02-05 Jakub Jelinek <jakub@redhat.com>
28308 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
28311 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
28312 (STATIC_LIBTSAN_LIBS): Likewise.
28313 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
28314 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
28315 is defined, don't add anything else beyond that.
28316 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
28317 (LINK_COMMAND_SPEC): Use them.
28319 PR tree-optimization/56205
28320 * tree-stdarg.c (check_all_va_list_escapes): Return true if
28321 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
28322 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
28324 2013-02-05 Richard Biener <rguenther@suse.de>
28326 PR tree-optimization/53342
28327 PR tree-optimization/53185
28328 * tree-vectorizer.h (vect_check_strided_load): Remove.
28329 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
28330 not disallow peeling for vectorized strided loads.
28331 (vect_check_strided_load): Make static and simplify.
28332 (vect_analyze_data_refs): Adjust.
28333 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
28334 correctly when vectorizing strided loads.
28336 2013-02-05 Richard Biener <rguenther@suse.de>
28338 * doc/install.texi: Refer to ISL, not PPL.
28340 2013-02-05 Jan Hubicka <jh@suse.cz>
28342 PR tree-optimization/55789
28343 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
28345 2013-02-05 Jan Hubicka <jh@suse.cz>
28347 PR tree-optimization/55789
28348 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
28349 the dead call anyway.
28351 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
28354 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
28356 2013-02-04 Alexander Potapenko <glider@google.com>
28357 Jack Howarth <howarth@bromo.med.uc.edu>
28358 Jakub Jelinek <jakub@redhat.com>
28361 * config/darwin.c (sort_ctor_records): Stabilized qsort
28362 on constructor priority by using original position.
28363 (finalize_ctors): New routine to sort constructors by
28364 priority before use in assemble_integer.
28365 (machopic_asm_out_constructor): Use finalize_ctors if needed.
28367 2013-02-04 Jakub Jelinek <jakub@redhat.com>
28370 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
28371 about visibility on artificial decls.
28372 * config/sol2.c (solaris_assemble_visibility): Likewise.
28374 2013-02-04 Kai Tietz <ktietz@redhat.com>
28377 * config/i386/i386.c (function_value_ms_64): Add additional valtype
28378 argument and improve checking of return-argument types for 16-byte
28380 (ix86_function_value_1): Add additional valtype argument on call
28381 of function_value_64.
28382 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
28383 handling infunction_value_64 function.
28385 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
28387 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
28389 2013-02-04 Richard Biener <rguenther@suse.de>
28391 PR tree-optimization/56188
28392 * tree-ssa-structalias.c (label_visit): Consider case with
28393 initially non-empty points-to set.
28394 (perform_var_substitution): Dump node mapping and clean up.
28396 2013-02-04 Richard Guenther <rguenther@suse.de>
28399 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
28400 node prevail as last resort.
28401 (lto_symtab_merge_decls): Remove guard on LTRANS here.
28402 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
28404 2013-02-04 Richard Biener <rguenther@suse.de>
28406 PR tree-optimization/56113
28407 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
28409 (equiv_class_lookup_or_add): ... this.
28410 (label_visit): Adjust and fix error in previous patch.
28411 (perform_var_substitution): Adjust.
28413 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28415 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28416 * config/sh/sh4-300.md: Likewise.
28417 * config/sh/sh4a.md: Likewise.
28418 * config/sh/constraints.md: Likewise.
28419 * config/sh/sh.md: Likewise.
28420 * config/sh/netbsd-elf.h: Likewise.
28421 * config/sh/predicates.md: Likewise.
28422 * config/sh/sh-protos.h: Likewise.
28423 * config/sh/ushmedia.h: Likewise.
28424 * config/sh/linux.h: Likewise.
28425 * config/sh/sh.c: Likewise.
28426 * config/sh/superh.h: Likewise.
28427 * config/sh/elf.h: Likewise.
28428 * config/sh/sh4.md: Likewise.
28429 * config/sh/sh.h: Likewise.
28431 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28433 * config/pa/constraints.md: Adjust unused letters. Change "T"
28434 constraint to match_test floating_point_store_memory_operand().
28435 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28436 (base14_operand): New.
28437 (floating_point_store_memory_operand): New.
28438 (integer_store_memory_operand): Revise to use base14_operand and
28439 reg_plus_base_memory_operand.
28440 (move_dest_operand): Allow symbolic_memory_operands.
28441 (symbolic_memory_operand): Check for LO_SOM.
28442 (symbolic_operand): Change default case to break.
28443 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28444 CONST_DOUBLE values to be reloaded by putting them into memory when
28445 the destination is a floating point register.
28446 (movdf): Remove code to handle CONST_DOUBLE.
28448 (reload_indf_r1): New.
28449 (reload_insf_r1): New.
28450 Consistently use "Q" and "T" constraints with integer and floating
28451 point move instructions, respectively.
28452 (movdi): Remove FAIL.
28453 Change predicate for source operand unamed DImode move from
28454 general_operand to move_src_operand.
28455 (umulsidi3): Change predicate for destination operand to
28457 Likewise for similar unamed patterns.
28458 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28459 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28460 (hppa_legitimize_address): Simplify mask calculation.
28461 (pa_emit_move_sequence): Revised handling of secondary reloads from
28462 REG+D addresses for floating point loads and stores. Directly handle
28463 loading CONST0_RTX (mode) to a floating point register.
28464 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28465 to floating point registers. Don't restrict secondary reloads to
28466 floating point registers to integer modes. Revise some comments and
28468 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28469 (pa_legitimate_address_p): New.
28470 (pa_legitimize_reload_address): New.
28471 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28472 (STRICT_REG_OK_FOR_BASE_P): New.
28473 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28474 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28476 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28477 Andrew Dixie <andrewd@gentrack.com>
28479 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28482 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28484 * expmed.c (extract_bit_field_1): Pass the full width of the
28485 structure to get_best_reg_extraction_insn.
28487 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28490 * configure.ac (use_cxa_atexit): Add AIX.
28491 * configure: Regenerate.
28493 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28495 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28498 * final.c (need_profile_function): New variable.
28499 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28500 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28501 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28502 notes, targetm.asm_out.function_prologue doesn't emit anything,
28503 HAVE_prologue and profiler should be emitted before prologue,
28504 set need_profile_function instead of emitting it.
28505 (final_scan_insn): If need_profile_function, emit
28506 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28507 NOTE_INSN_FUNCTION_BEG note.
28509 2013-02-01 Richard Henderson <rth@redhat.com>
28511 * config/rs6000/rs6000.md (smulditi3): New.
28514 * config/alpha/alpha.md (umulditi3): New.
28516 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28518 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28519 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28521 2013-02-01 Richard Biener <rguenther@suse.de>
28523 PR tree-optimization/56113
28524 * tree-ssa-structalias.c (label_visit): Reduce work for
28525 single-predecessor nodes.
28527 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28529 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28530 range isn't testing for zero.
28532 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28534 PR middle-end/56113
28535 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28537 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28538 Nick Clifton <nickc@redhat.com>
28540 * config/v850/constraints.md (Q): Define as a memory constraint.
28541 * config/v850/predicates.md (label_ref_operand): New predicate.
28542 (e3v5_shift_operand): New predicate.
28543 (ior_operator): New predicate.
28544 * config/v850/t-v850: Add e3v5 multilib.
28545 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28546 (v850_gen_movdi): Prototype.
28547 * config/v850/v850.c: Add support for e3v5 architecture.
28548 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28550 (construct_save_jarl): Add e3v5 long JARL support.
28551 (v850_adjust_insn_length): New function. Adjust length of call
28552 insns when using e3v5 instructions.
28553 (v850_gen_movdi): New function: Generate instructions to move a
28555 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
28556 (CPP_SPEC): Define __v850e3v5__ as appropriate.
28557 (TARGET_USE_FPU): Enable for e3v5.
28558 (CONST_OK_FOR_W): New macro.
28559 (ADJUST_INSN_LENGTH): Define.
28560 * config/v850/v850.md (UNSPEC_LOOP): Define.
28561 (attr cpu): Add v850e3v5.
28562 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
28563 (movdi): New pattern.
28564 (movdi_internal): New pattern.
28565 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
28566 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
28567 (cstoresf4): Likewise.
28568 (cstoredf4): Likewise.
28569 (insv): New pattern.
28570 (rotlso3_a): New pattern.
28571 (rotlsi3_b): New pattern
28572 (rotlsi3_v850e3v5): New pattern.
28573 (doloop_begin): New pattern.
28574 (fix_loop_counter): New pattern.
28575 (doloop_end): New pattern.
28576 (branch_normal): Add e3v5 long branch support.
28577 (branch_invert): Likewise.
28578 (branch_z_normal): Likewise.
28579 (branch_z_invert): Likewise.
28580 (branch_nz_normal): Likewise.
28581 (branch_nz_invert): Likewise.
28582 (call_internal_short): Add e3v5 register-indirect JARL support.
28583 (call_internal_long): Likewise.
28584 (call_value_internal_short): Likewise.
28585 (call_value_internal_long): Likewise.
28586 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
28587 (mloop): New option.
28588 * config.gcc: Add support for configuring v840e3v5 target.
28589 * doc/invoke.texi: Document new v850 specific command line options.
28591 2013-01-31 Paul Koning <ni1d@arrl.net>
28595 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
28596 children if parent is a class.
28597 (prune_unused_types_prune): Don't add DW_AT_declaration.
28599 2013-01-31 Richard Biener <rguenther@suse.de>
28601 PR tree-optimization/56157
28602 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28603 match up operand with SLP child.
28605 2013-01-31 Jason Merrill <jason@redhat.com>
28608 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28609 parameters the first time.
28610 (gen_scheduled_generic_parms_dies): Check completeness here.
28612 2013-01-31 Richard Biener <rguenther@suse.de>
28614 PR middle-end/53073
28615 * common.opt (faggressive-loop-optimizations): New flag,
28616 enabled by default.
28617 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28618 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28619 infer_loop_bounds_from_undefined by it.
28621 2013-01-31 Richard Biener <rguenther@suse.de>
28623 PR tree-optimization/56150
28624 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28625 visit virtual operands.
28626 (find_uses_to_rename_bb): Likewise.
28628 2013-01-31 Richard Biener <rguenther@suse.de>
28630 PR tree-optimization/56150
28631 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28632 mixed store non-store stmts.
28634 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28637 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28638 LIBASAN_EARLY_SPEC is defined.
28639 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28640 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28642 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28645 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28646 invalid args instead of ICEing on it.
28647 (ix86_valid_target_attribute_tree): Return error_mark_node if
28648 ix86_valid_target_attribute_inner_p failed.
28649 (ix86_valid_target_attribute_p): Return false only if
28650 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28651 target("default") attribute.
28652 (sorted_attr_string): Change argument from const char * to tree,
28653 merge in all target attribute arguments rather than just one.
28654 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28655 instead of free. Avoid using strcat.
28656 (ix86_mangle_function_version_assembler_name): Mangle
28657 target("default") as if no target attribute is present. Adjust
28658 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28659 instead of xmalloc and XDELETEVEC instead of free.
28660 (ix86_function_versions): Don't return true if one of the decls
28661 doesn't have target attribute. If they don't and one of the decls
28662 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28663 sorted_attr_string caller. Use XDELETEVEC instead of free.
28664 (ix86_supports_function_versions): Remove.
28665 (make_name): Fix up formatting.
28666 (make_dispatcher_decl): Remove resolver_name and its initialization.
28667 Avoid leaking memory.
28668 (is_function_default_version): Return true if there is
28669 target("default") attribute rather than no target attribute at all.
28670 (make_resolver_func): Avoid leaking memory.
28671 (ix86_generate_version_dispatcher_body): Likewise.
28672 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28673 * target.def (supports_function_versions): Remove.
28674 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28675 * doc/tm.texi: Regenerated.
28677 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28679 PR rtl-optimization/56144
28680 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28681 for values with side effects.
28683 2013-01-30 Richard Biener <rguenther@suse.de>
28685 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28686 (sparseset_pop): Likewise.
28687 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28688 to be able to use quick_push in the worker loop.
28690 2013-01-30 Marek Polacek <polacek@redhat.com>
28692 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28694 2013-01-30 Richard Biener <rguenther@suse.de>
28697 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28699 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28701 PR tree-optimization/56064
28702 * fixed-value.c (fixed_from_double_int): New function.
28703 * fixed-value.h (fixed_from_double_int): New prototype.
28704 (const_fixed_from_double_int): New static inline function.
28705 * fold-const.c (native_interpret_fixed): New static function.
28706 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28707 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28708 (native_encode_fixed): New static function.
28709 (native_encode_expr) <FIXED_CST>: Use it.
28710 (native_interpret_int): Move double_int worker code to...
28711 * double-int.c (double_int::from_buffer): ...this new static method.
28712 * double-int.h (double_int::from_buffer): Prototype it.
28714 2013-01-30 Richard Biener <rguenther@suse.de>
28716 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28717 New pointer-map and obstack.
28718 (init_alias_vars): Allocate pointer-map and obstack.
28719 (delete_points_to_sets): Free them.
28720 (find_what_var_points_to): Cache result.
28721 (find_what_p_points_to): Adjust for changed interface of
28722 find_what_var_points_to.
28723 (compute_points_to_sets): Likewise.
28724 (ipa_pta_execute): Likewise.
28726 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28728 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28729 * configure: Regenerate.
28730 * config.in: Regenerate.
28731 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28732 #nobits/#progbits if supported.
28734 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28737 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28738 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28739 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28741 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28743 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28744 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28746 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28748 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28749 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28751 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28753 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28755 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28756 * config/arm/cortex-a7.md: New bypasses using
28757 arm_mac_accumulator_is_result.
28759 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28761 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28762 (cortex_a7_neon_mla): Likewise.
28763 (cortex_a7_fpfmad): New reservation.
28764 (cortex_a7_fpmacs): Use ffmas and update required units.
28765 (cortex_a7_fpmuld): Update required units and latency.
28766 (cortex_a7_fpmacd): Likewise.
28767 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28768 (cortex_a7_neon). Likewise.
28769 (bypass) Update participating units.
28771 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28773 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28774 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28776 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28777 (vfp_fmul): Use ffmad.
28778 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28779 (cortex_r4_fmacd): Use ffmad.
28780 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28781 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28782 (cortex_a9_fmacd): Use ffmad.
28783 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28784 (cortex_a8_vfp_macd): Use ffmad.
28785 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28786 (cortex_a5_fpmacd): Use ffmad.
28787 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28788 (cortex_a15_vfp_macd): Use ffmad.
28789 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28791 2013-01-29 Jason Merrill <jason@redhat.com>
28794 * varasm.c (default_assemble_visibility): Don't warn about
28795 visibility on artificial decls.
28797 2013-01-29 Richard Biener <rguenther@suse.de>
28799 PR tree-optimization/56113
28800 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28802 (label_visit): Free duplicate bitmaps and record the leader instead.
28803 (perform_var_substitution): Adjust.
28805 2013-01-29 Richard Biener <rguenther@suse.de>
28807 PR tree-optimization/55270
28808 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28809 the CFG, schedule loops for fixup.
28811 2013-01-29 Nick Clifton <nickc@redhat.com>
28813 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28816 2013-01-28 Leif Ekblad <leif@rdos.net>
28818 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28819 * config/i386/i386.h (TARGET_RDOS): New macro.
28820 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28821 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28822 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28823 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28824 DEFAULT_LARGE_SECTION_THRESHOLD.
28825 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28826 * config/i386/rdos.h: New file.
28827 * config/i386/rdos64.h: New file.
28829 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28832 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28835 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28837 PR rtl-optimization/56117
28838 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28839 call cselib_lookup_from_insn on the MEM before calling
28840 add_insn_mem_dependence.
28842 2013-01-28 Richard Biener <rguenther@suse.de>
28844 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28845 to a stmt that didn't have one.
28846 (copy_phis_for_bb): Likewise for PHI arguments.
28847 (copy_debug_stmt): Likewise for debug stmts.
28849 2013-01-28 Richard Biener <rguenther@suse.de>
28851 PR tree-optimization/56034
28852 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28853 (partition_builtin_p): Adjust.
28854 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28855 it is the last partition.
28856 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28857 up the vertex for the definition.
28858 (classify_partition): Classify whether a partition is a
28859 PKIND_REDUCTION, thus has uses outside of the loop.
28860 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28861 Merge all PKIND_REDUCTION partitions into the last partition.
28862 (tree_loop_distribution): Seed partitions from reductions as well.
28864 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28866 PR tree-optimization/56125
28867 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28868 pow(x,c) into sqrt(x) * powi(x, n/2) or
28869 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28870 optimizing for size.
28871 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28872 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28875 PR tree-optimization/56094
28876 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28877 to UNKNOWN_LOCATION while gimplifying expr.
28879 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28882 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28883 operand 0 in movabs insn template for -masm=intel asm alternative.
28884 (*movabs<mode>_2): Ditto for operand 1.
28886 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28889 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28892 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28894 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28897 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28899 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28900 Undef to avoid warning.
28902 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28904 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28905 * configure: Regenerate.
28907 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28909 PR tree-optimization/56098
28910 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28911 for stmts with volatile ops.
28912 (cond_store_replacement): Don't optimize if assign has volatile ops.
28913 (cond_if_else_store_replacement_1): Don't optimize if either
28914 then_assign or else_assign have volatile ops.
28915 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28918 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28920 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28922 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28924 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28925 missing ':' in asm example.
28927 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28929 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28930 entries into lane and laneq entries.
28931 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28932 Remove AdvSIMD scalar modes.
28933 (aarch64_sq<r>dmulh_laneq<mode>): New.
28934 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28936 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28937 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28938 * config/aarch64/iterators.md (VCOND): New.
28941 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28944 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28945 Add NULL LIBNAME argument to existing definitions.
28946 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28947 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28948 * config/avr/avr.c (DEF_BUILTIN): Same.
28949 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28950 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28951 implementation is available (DECL_ASSEMBLER_NAME is set).
28952 (avr_fold_absfx): New static function.
28953 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28954 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28955 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28956 AVR_BUILTIN_ABSLLK.
28957 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28958 (abshk, absk, abslk, absllk): Provide as static inline functions.
28960 2013-01-25 Marek Polacek <polacek@redhat.com>
28962 PR tree-optimization/56035
28963 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28965 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28967 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28968 (*movtf_internal_rex64): Add (!o,C) alternative
28969 (*movxf_internal_rex64): Ditto.
28970 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28972 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28974 * doc/invoke.texi: fix typo.
28975 * doc/objc.texi: fix typo.
28977 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28979 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28980 for the first two alternatives.
28982 2013-01-24 Diego Novillo <dnovillo@google.com>
28984 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28985 (ggc-zone.o): Remove.
28986 * configure.ac: Remove option --with-gc.
28987 * configure: Re-generate.
28988 * doc/install.texi: Remove documentation for --with-gc.
28989 * gengtype.c (write_enum_defn): Remove. Update all users.
28990 (write_Types_process_field): Remove generation of gt_e_* argument.
28991 (output_type_enum): Remove. Update all users.
28992 (write_enum_defn): Remove. Update all users.
28993 (enum alloc_zone): Remove. Update all users.
28994 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28995 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28996 Update all callers.
28997 (struct ptr_data): Remove field TYPE. Update all users.
28998 (gt_pch_note_object): Remove argument TYPE. Update all users.
28999 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
29001 * ggc-none.c (ggc_alloc_typed_stat): Remove.
29002 (struct alloc_zone): Remove.
29003 (ggc_internal_alloc_zone_stat): Remove.
29004 (ggc_internal_cleared_alloc_zone_stat): Remove.
29005 * ggc-page.c (ggc_alloc_typed_stat): Remove.
29006 (ggc_pch_count_object): Remove last argument. Update all users.
29007 (ggc_pch_alloc_object): Remove last argument. Update all users.
29008 (struct alloc_zone): Remove.
29009 * ggc-zone.c: Remove.
29010 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
29011 (struct alloc_zone): Remove.
29012 (ggc_alloc_typed_stat): Remove.
29013 (ggc_alloc_typed): Remove.
29014 (ggc_splay_alloc): Remove first argument.
29015 (rtl_zone): Remove. Update all users.
29016 (tree_zone): Remove. Update all users.
29017 (tree_id_zone): Remove. Update all users.
29018 (ggc_internal_zone_alloc_stat): Remove. Update all users.
29019 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
29020 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
29021 * tree-ssanames.c: Remove references to zone allocator in comments.
29023 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
29025 * config/avr/avr.c (avr_out_fract): Make register numbers that
29026 might be outside of source operand signed.
29028 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
29030 * config/i386/constraints.md (Yf): New constraint.
29031 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
29032 of f constraint to conditionaly disable x87 register preferences.
29033 (*movdf_internal): Ditto.
29034 (*movsf_internal): Ditto.
29036 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
29038 PR inline-asm/55934
29039 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
29040 that have operands with impossible constraints.
29041 Add a FIXME for a speed-up opportunity.
29042 * lra-constraints.c (process_alt_operands): Verify that a class
29043 selected from constraints on asms is valid for the operand mode.
29044 (curr_insn_transform): Remove incorrect comment.
29046 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
29048 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
29049 TOC operand is a valid symbol ref in the constant pool.
29051 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29053 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
29055 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
29058 * config/avr/stdfix.h: New file.
29059 * t-avr (stdfix-gcc.h): New rule to build it.
29060 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
29062 2013-01-23 Kostya Serebryany <kcc@google.com>
29064 * config/darwin.h: remove dependency on
29065 CoreFoundation (asan on Mac OS).
29067 2013-01-23 Jakub Jelinek <jakub@redhat.com>
29070 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
29071 instead of cmpdi_operand for first comparison operand.
29072 Don't assert that comparison operands aren't both constants.
29074 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29076 * doc/install.texi (Downloading the Source): Update references to
29077 downloading separate components.
29079 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29081 * doc/extend.texi (__int128): Improve grammar.
29083 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
29086 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
29087 alternative to (o,r).
29088 (*movdi_internal_rex64): Remove (!o,n) alternative.
29089 (DImode immediate->memory splitter): Remove.
29090 (DImode immediate->memory peephole2): Remove.
29091 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
29092 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
29093 alternative to (!o,*r).
29094 (*movtf_internal_sse): New pattern.
29095 (*movxf_internal_rex64): New pattern.
29096 (*movxf_internal): Disable for TARGET_64BIT.
29097 (*movdf_internal_rex64): Remove (!o,F) alternative.
29099 2013-01-22 Jakub Jelinek <jakub@redhat.com>
29101 PR middle-end/56074
29102 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
29103 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
29104 * tree-vect-loop-manip.c (find_loop_location): Also ignore
29105 stmt locations where LOCATION_LOCUS of the stmt location is
29106 UNKNOWN_LOCATION or BUILTINS_LOCATION.
29109 * config/i386/i386.md (UNSPEC_STOS): New.
29110 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
29111 *strsetqi_1): Add UNSPEC_STOS.
29113 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
29116 * doc/invoke.texi: Remove left over -Wsynth example.
29118 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29120 PR tree-optimization/56051
29121 * fold-const.c (fold_binary_loc): Don't fold
29122 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
29123 a narrowing conversion, or widening conversion from signed
29126 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29128 PR rtl-optimization/56023
29129 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
29130 dependent on debug instruction.
29132 2013-01-21 Martin Jambor <mjambor@suse.cz>
29134 PR middle-end/56022
29135 * function.c (allocate_struct_function): Call
29136 invoke_set_current_function_hook earlier.
29138 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29140 * reload1.c (init_reload): Only initialize reload_obstack
29141 during the first call.
29143 2013-01-21 Marek Polacek <polacek@redhat.com>
29145 * cfgloop.c (verify_loop_structure): Fix up grammar.
29147 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
29149 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
29150 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
29152 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29155 * config/arm/marvell-pj4.md: Update copyright year.
29156 Fix up use of alu to alu_reg and simple_alu_imm.
29158 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29160 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
29162 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
29165 * lra-constraints.c (curr_insn_transform): Don't reuse original
29166 insn for secondary memory move when memory mode should be different.
29168 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29170 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
29171 atomic_storedi_1): New patterns.
29173 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29175 btver2 pipeline descriptions.
29176 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
29178 * config/i386/i386.md (btver2_decode): New type attributes.
29179 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
29181 * config/i386/btver2.md: New file describing btver2 pipelines.
29183 2013-01-19 Andrew Pinski <apinski@cavium.com>
29185 PR tree-optimization/52631
29186 * tree-ssa-sccvn (visit_use): Before looking up the original
29187 statement, try looking up the simplified expression.
29189 2013-01-19 Anthony Green <green@moxielogic.com>
29191 * config/moxie/moxie.c (moxie_expand_prologue): Set
29192 current_function_static_stack_size.
29194 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29196 PR tree-optimization/56029
29197 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
29198 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
29200 2013-01-18 Sharad Singhai <singhai@google.com>
29202 PR tree-optimization/55995
29203 * dumpfile.c (dump_loc): Print location only if available.
29204 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
29206 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
29209 * lra-constraints.c (curr_insn_transform): Reuse original insn for
29210 secondary memory move.
29211 (inherit_reload_reg): Use rclass instead of cl for
29212 check_secondary_memory_needed_p.
29214 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29216 PR middle-end/56015
29217 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
29218 the case where writing real complex part of target modifies op1.
29220 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
29222 * config/aarch64/aarch64-simd.md
29223 (aarch64_vcond_internal<mode>): Handle unordered cases.
29224 * config/aarch64/iterators.md (v_cmp_result): New.
29226 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
29227 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29229 * config/arm/marvell-pj4.md: New file.
29230 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
29231 * config/arm/arm.md (generic_sched): Add marvell_pj4.
29232 (generic_vfp): Likewise.
29233 * config/arm/arm-cores.def: Add marvell-pj4.
29234 * config/arm/arm-tune.md: Regenerate.
29235 * config/arm/arm-tables.opt: Regenerate.
29236 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
29237 * doc/invoke.texi: Document marvell-pj4.
29239 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
29241 * config/aarch64/arm_neon.h: Map scalar types to standard types.
29243 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29248 * var-tracking.c (negative_power_of_two_p): New.
29249 (global_get_addr_cache, local_get_addr_cache): New.
29250 (get_addr_from_global_cache, get_addr_from_local_cache): New.
29251 (vt_canonicalize_addr): Rewrite using the above. Adjust the
29253 (vt_stack_offset_p): Remove.
29254 (vt_canon_true_dep): Always canonicalize loc's address.
29255 (clobber_overlapping_mems): Make sure we have a MEM.
29256 (local_get_addr_clear_given_value): New.
29257 (val_reset): Clear local cached entries.
29258 (compute_bb_dataflow): Create and release the local cache.
29259 Disable duplicate MEMs clobbering.
29260 (emit_notes_in_bb): Clobber MEMs likewise.
29261 (vt_emit_notes): Create and release the local cache.
29262 (vt_initialize, vt_finalize): Create and release the global
29263 cache, respectively.
29264 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
29266 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29268 PR libmudflap/53359
29269 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
29270 not found in the symtab.
29272 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29275 PR rtl-optimization/55547
29276 PR rtl-optimization/53827
29279 * alias.c (offset_overlap_p): New, factored out of...
29280 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
29281 the conservative special case for symbolic constants. Don't
29282 adjust zero sizes on alignment.
29284 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
29286 PR rtl-optimization/52573
29287 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
29288 REG_UNUSED for the same register.
29290 2013-01-17 Richard Biener <rguenther@suse.de>
29291 Marek Polacek <polacek@redhat.com>
29293 PR rtl-optimization/55833
29294 * loop-unswitch.c (unswitch_loops): Move loop verification...
29295 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
29296 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
29297 Set it to true when we're removing a loop from hierarchy tree in
29298 an irreducible region.
29299 (fix_bb_placements): Adjust caller.
29300 (fix_loop_placements): Likewise.
29302 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
29304 * config/avr/builtins.def (DEF_BUILTIN): Factor out
29305 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
29306 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
29307 Remove ID. Adjust comments.
29308 * config/avr/avr-c.c (avr_builtin_name): Remove.
29309 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
29310 * config/avr/avr.c (avr_tolower): New static function.
29311 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
29312 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
29313 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
29316 2013-01-17 Jan Hubicka <jh@suse.cz>
29318 PR tree-optimization/55273
29319 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
29321 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
29324 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
29325 store through atomic_store<mode>_1.
29326 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
29328 2013-01-17 Martin Jambor <mjambor@suse.cz>
29330 PR tree-optimizations/55264
29331 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
29332 for virtual methods.
29333 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
29334 virtual methods before inlining is over.
29335 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
29337 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
29340 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
29342 PR rtl-optimization/56005
29343 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
29344 pending reads for prefetch.
29346 2013-01-16 Ian Bolton <ian.bolton@arm.com>
29348 * config/aarch64/aarch64.md
29349 (*cstoresi_neg_uxtw): New pattern.
29350 (*cmovsi_insn_uxtw): New pattern.
29351 (*<optab>si3_uxtw): New pattern.
29352 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
29353 (*<optab>si3_insn_uxtw): New pattern.
29354 (*bswapsi2_uxtw): New pattern.
29356 2013-01-16 Richard Biener <rguenther@suse.de>
29358 * tree-inline.c (tree_function_versioning): Remove set but
29359 never used variable.
29361 2013-01-16 Richard Biener <rguenther@suse.de>
29363 PR tree-optimization/55964
29364 * tree-flow.h (rename_variables_in_loop): Remove.
29365 (rename_variables_in_bb): Likewise.
29366 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
29367 (copy_loop_before): Adjust and delete update-ssa status.
29368 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
29369 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
29370 (rename_variables_in_loop): Remove.
29371 (slpeel_update_phis_for_duplicate_loop): Likewise.
29372 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
29373 use available cfg machinery instead of duplicating it.
29374 Update PHI nodes and perform poor-mans SSA update here.
29375 (slpeel_tree_peel_loop_to_edge): Adjust.
29377 2013-01-16 Richard Biener <rguenther@suse.de>
29379 PR tree-optimization/54767
29380 PR tree-optimization/53465
29381 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
29382 (vrp_visit_phi_node): For PHI arguments coming via backedges
29383 drop all symbolical range information.
29384 (execute_vrp): Compute backedges.
29386 2013-01-16 Richard Biener <rguenther@suse.de>
29388 * doc/install.texi: Update CLooG and ISL requirements to
29391 2013-01-16 Christian Bruel <christian.bruel@st.com>
29394 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
29395 (broken_move): Handle UNSPECV_SP_SWITCH_B.
29396 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
29398 2013-01-16 DJ Delorie <dj@redhat.com>
29400 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
29401 (UNSPECV_SP_SWITCH_E): New.
29402 (sp_switch_1): Change to an unspec.
29403 (sp_switch_2): Change to an unspec. Don't use post-inc when we
29406 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
29408 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29409 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29410 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29411 (expand_mem_thread_fence): Ditto.
29412 (expand_mem_signal_fence): Ditto.
29413 (expand_atomic_load): Ditto.
29414 (expand_atomic_store): Ditto.
29416 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29418 PR rtl-optimization/55547
29419 PR rtl-optimization/53827
29422 * alias.c (memrefs_conflict_p): Set sizes to negative after
29425 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29428 * function.c (thread_prologue_and_epilogue_insns): Always
29429 add crtl->drap_reg to set_up_by_prologue.set, even if
29430 stack_realign_drap is false.
29432 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29434 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29435 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29436 *call): Fix indention.
29438 2013-01-15 Tom de Vries <tom@codesourcery.com>
29441 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29444 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29446 PR rtl-optimization/55153
29447 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29449 2013-01-15 Martin Jambor <mjambor@suse.cz>
29451 PR tree-optimization/55920
29452 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29453 accesses as grp_to_be_debug_replaced.
29455 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29457 PR tree-optimization/55920
29458 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29459 there is non-useless type conversion needed from debug rhs to lhs,
29460 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29462 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29463 Mikael Pettersson <mikpe@it.uu.se>
29466 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29468 (ASM_OUTPUT_CASE_LABEL): Remove.
29469 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29470 * final.c (shorten_branches): Update alignment of labels before
29471 jump tables if CASE_VECTOR_SHORTEN_MODE.
29473 2013-01-15 Richard Biener <rguenther@suse.de>
29476 * system.h: Do not include gmp.h for building host tools.
29478 2013-01-15 Richard Biener <rguenther@suse.de>
29480 PR middle-end/55882
29481 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29482 account for bitpos when computing alignment.
29484 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29486 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29487 (ix86_target_macros_internal): Likewise.
29489 * config/i386/i386.c (m_CORE2I7): Removed.
29490 (m_CORE_HASWELL): New macro.
29491 (m_CORE_ALL): Likewise.
29492 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29493 (initial_ix86_arch_features): Likewise.
29494 (processor_target_table): Initializations for Core avx2.
29495 (cpu_names): New names "core-avx2".
29496 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29497 PROCESSOR_CORE_HASWELL.
29498 (ix86_issue_rate): New case.
29499 (ia32_multipass_dfa_lookahead): Likewise.
29500 (ix86_sched_init_global): Likewise.
29502 * config/i386/i386.h (TARGET_HASWELL): New macro.
29503 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29504 (processor_type): New PROCESSOR_HASWELL.
29506 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29508 PR tree-optimization/55955
29509 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29510 *SHIFT_EXPR and *ROTATE_EXPR codes.
29512 PR tree-optimization/48766
29513 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29514 -ftrapv disable -fwrapv.
29516 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29519 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29520 etc. to 1 and not to __flash.
29521 Use LL suffix for __INT24_MAX__ with -mint8.
29522 Use ULL suffix for __UINT24_MAX__ with -mint8.
29524 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29526 * config/avr/avr-arch.h
29527 (struct base_arch_s): Use typedef avr_arch_t instead.
29528 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29529 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29530 * config/avr/avr.c: Same.
29531 * config/avr/avr-devices.c: Same.
29532 * config/avr/driver-avr.c: Same.
29533 * config/avr/gen-avr-mmcu-texi.c: Same.
29534 * config/avr/avr-mcus.def: Adjust comment.
29536 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29538 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29539 * config/aarch64/iterators.md (VALLDI): New.
29541 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29542 Andi Kleen <ak@linux.intel.com>
29545 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29546 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29549 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29551 * config/avr/avr-stdint.h: Remove trailing blanks.
29552 * config/avr/avr-log.h: Same.
29553 * config/avr/avr-arch.h: Same.
29554 * config/avr/avr-devices.c: Same.
29555 * config/avr/avr-dimode.md: Same.
29556 * config/avr/predicates.md: Same.
29557 * config/avr/avr-c.c: Same. And fix typo.
29559 * config/avr/avr-protos.h: Same. And:
29560 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
29561 (init_cumulative_args): Rename to avr_init_cumulative_args.
29562 (expand_prologue): Rename to avr_expand_prologue.
29563 (expand_epilogue): Rename to avr_expand_epilogue.
29564 (adjust_insn_length): Rename to avr_adjust_insn_length.
29565 (notice_update_cc): Rename to avr_notice_update_cc.
29566 (final_prescan_insn): Rename to avr_final_prescan_insn.
29567 * config/avr/avr.c: Same.
29568 * config/avr/avr.h: Same.
29569 * config/avr/avr.md: Remove trailing blanks.
29570 (prologue): Use avr_expand_prologue.
29571 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
29573 2013-01-14 Richard Biener <rguenther@suse.de>
29575 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
29576 verify_location, collect_subblocks): New functions.
29577 (verify_gimple_in_cfg): Verify that locations only reference
29578 BLOCKs in the functions BLOCK tree.
29580 2013-01-14 Richard Biener <rguenther@suse.de>
29582 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
29584 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
29586 (insert_out_of_ssa_copy_on_edge): Likewise.
29587 (rewrite_close_phi_out_of_ssa): Likewise.
29588 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
29590 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
29591 propagated constants.
29592 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
29595 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29597 * config/avr/avr-modes.def: Add GPL copyright notice.
29599 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
29601 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29602 MEMMODEL_MASK to determine memory model.
29603 (atomic_store<mode>): Ditto from operands[2].
29604 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29606 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29609 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29610 (fold_gimple_assign): Don't call unshare_expr here.
29611 (fold_ctor_reference): Call unshare_expr.
29613 2013-01-13 Terry Guo <terry.guo@arm.com>
29615 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29616 * doc/fragments.texi: Document MULTILIB_REUSE.
29617 * gcc.c (multilib_reuse): New internal spec.
29618 (set_multilib_dir): Also search multilib from multilib_reuse.
29619 * genmultilib (tmpmultilib3): Refactor code.
29620 (tmpmultilib4): Ditto.
29621 (multilib_reuse): New multilib argument.
29623 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29625 * Makefile.in: Update copyright.
29627 2013-01-12 Tom de Vries <tom@codesourcery.com>
29629 PR middle-end/55890
29630 * calls.c (expand_call): Check if arg_nr is valid.
29632 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29634 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29635 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29636 documentation. Add missing '__' in front of
29637 __builtin_ia32_packssdw256.
29639 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29642 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29643 NO_REGS for larl operands.
29644 (s390_reload_larl_operand): Use s390_load_address instead of
29647 2013-01-11 Richard Biener <rguenther@suse.de>
29649 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29650 (verify_node_sharing): ... here.
29651 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29653 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29655 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29656 Substitute TREECHECKING.
29657 * configure: Regenerate.
29658 * Makefile.in (TREECHECKING): New.
29660 2013-01-11 Richard Guenther <rguenther@suse.de>
29662 PR tree-optimization/44061
29663 * tree-vrp.c (extract_range_basic): Compute zero as
29664 value-range for __builtin_constant_p of function parameters.
29666 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29668 Update copyright years.
29670 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29672 PR rtl-optimization/55672
29673 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29674 const to be eliminable.
29676 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29678 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29679 * configure: Regenerate.
29681 2013-01-10 Richard Biener <rguenther@suse.de>
29683 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29685 2013-01-10 Richard Biener <rguenther@suse.de>
29688 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29689 locations for virtual PHI arguments.
29690 (rewrite_update_phi_arguments): Likewise.
29692 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29694 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29697 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29699 PR tree-optimization/55921
29700 * tree-complex.c (expand_complex_asm): New function.
29701 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29703 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29706 * config/s390/s390.c (s390_symref_operand_p)
29707 (s390_loadrelative_operand_p): Merge the two functions.
29708 (s390_check_qrst_address, print_operand_address): Add parameters
29709 to s390_loadrelative_operand_p invokation.
29710 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29711 (s390_reload_larl_operand, s390_secondary_reload): Use
29712 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29713 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29715 2013-01-09 Mike Stump <mikestump@comcast.net>
29717 * dse.c (record_store): Remove unnecessary assert.
29719 2013-01-09 Jan Hubicka <jh@suse.cz>
29721 PR tree-optimization/55569
29722 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29723 * cfgloop.h (scale_loop_profile): Likewise.
29725 2013-01-09 Jan Hubicka <jh@suse.cz>
29728 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29730 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29732 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29734 PR middle-end/55114
29735 * expr.h (maybe_emit_group_store): Declare.
29736 * expr.c (maybe_emit_group_store): New function.
29737 * builtins.c (expand_builtin_int_roundingfn): Call it.
29738 (expand_builtin_int_roundingfn_2): Likewise.
29740 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29742 PR rtl-optimization/55829
29743 * lra-constraints.c (match_reload): Add code for absent output.
29744 (curr_insn_transform): Add code for reloads of matched inputs
29747 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29749 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29750 attribute of movddup insn to DF.
29751 (*vec_interleave_lowv2df): Ditto.
29752 (vec_dupv2df): Ditto.
29754 2013-01-09 Jan Hubicka <jh@suse.cz>
29756 PR tree-optimiation/55875
29757 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29758 EVERY_ITERATION parameter.
29759 (number_of_iterations_exit): Check if exit is executed every iteration.
29760 (idx_infer_loop_bounds): Similarly here.
29761 (n_of_executions_at_most): Simplify
29762 to only test for cases where statement is dominated by the
29763 particular bound; handle correctly the "postdominance" test.
29764 (scev_probably_wraps_p): Use max loop iterations info
29765 as a global bound first.
29767 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29768 Nick Clifton <nickc@redhat.com>
29770 * config/v850/v850.md (cbranchsf4): New pattern.
29771 (cstoresf4): New pattern.
29772 (cbranchdf4): New pattern.
29773 (cstoredf4): New pattern.
29774 (movsicc): Disallow floating point comparisons.
29775 (cmpsf_le_insn): Fix order of operators.
29776 (cmpsf_lt_insn): Likewise.
29777 (cmpsf_eq_insn): Likewise.
29778 (cmpdf_le_insn): Likewise.
29779 (cmpdf_lt_insn): Likewise.
29780 (cmpdf_eq_insn): Likewise.
29781 (cmpsf_ge_insn): Use LE comparison.
29782 (cmpdf_ge_insn): Likewise.
29783 (cmpsf_gt_insn): Use LT comparison.
29784 (cmpdf_gt_insn): Likewise.
29785 (cmpsf_ne_insn): Delete pattern.
29786 (cmpdf_ne_insn): Delete pattern.
29787 * config/v850/v850.c (v850_gen_float_compare): Use
29788 gen_cmpdf_eq_insn for NE comparison.
29789 (v850_float_z_comparison_operator)
29790 (v850_float_nz_comparison_operator): Move from here ...
29791 * config/v850/predicates.md: ... to here. Move GT and GE
29792 comparisons into v850_float_z_comparison_operator.
29793 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29795 (v850_float_nz_comparison_operator): Likewise.
29797 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29799 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29800 with calls to gen_insvsi/gen_insvdi.
29802 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29804 * config/i386/i386.c (initial_ix86_tune_features): Set up
29805 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29807 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29808 Jakub Jelinek <jakub@redhat.com>
29810 PR tree-optimization/48189
29811 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29812 If nitercst is 0, don't predict the exit edge.
29814 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29816 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29817 in asm_fprintf with reg_names.
29818 (aarch64_print_operand_address): Likewise.
29819 (aarch64_return_addr): Likewise.
29820 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29822 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29824 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29825 (INT_U6_BITS): Likewise.
29826 * config/pa/predicates.md (uint6_operand): New predicate.
29827 (shift5_operand, shift6_operand): Likewise.
29828 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29830 (lshrdi3): Use shift6_operand.
29831 (shrpsi4, shrpdi4): New insn patterns.
29832 (extzv): Delete expander.
29833 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29834 predicates in unamed zero extract patterns. Tighten common constraint.
29835 (extv): Delete expander.
29836 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29837 predicates in unamed sign extract patterns. Tighten common constraint.
29838 (insv): Delete expander.
29839 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29840 predicates in unamed insert patterns. Tighten common constraint.
29841 Change uint32_operand predicate to uint6_operand predicate in unamed
29842 DImode pattern to insert constant values of type 1...1xxxx.
29844 2013-01-04 Jan Hubicka <jh@suse.cz>
29846 PR tree-optimization/55823
29847 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29850 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29851 Uros Bizjak <ubizjak@gmail.com>
29853 PR rtl-optimization/55845
29854 * df-problems.c (can_move_insns_across): Stop scanning at
29855 volatile_insn_p source instruction or give up if
29856 across_from .. across_to range contains any volatile_insn_p
29859 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29861 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29862 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29864 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29865 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29867 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29870 * asan.c (asan_clear_shadow): New function.
29871 (asan_emit_stack_protection): Use it.
29873 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29875 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29876 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29877 with tab instead of space.
29879 2013-01-08 Nick Clifton <nickc@redhat.com>
29881 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29882 register bank 0 at the start of an interrupt handler.
29883 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29886 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29888 * config/aarch64/aarch64-simd.md
29889 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29890 (aarch64_simd_bsl): Likewise.
29891 (aarch64_vcond_internal<mode>): Likewise.
29892 (vcond<mode><mode>): Likewise.
29893 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29894 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29896 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29898 * config/aarch64/aarch64-builtins.c
29899 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29901 2013-01-08 Martin Jambor <mjambor@suse.cz>
29904 * tree-sra.c (analyze_access_subtree): Return true also after
29905 potentially creating a debug-only replacement.
29907 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29909 PR middle-end/55890
29910 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29912 PR tree-optimization/54120
29913 * tree-vrp.c (range_fits_type_p): Don't allow
29914 src_precision < precision from signed vr to unsigned_p
29915 if vr->min or vr->max is negative.
29916 (simplify_float_conversion_using_ranges): Test can_float_p
29917 against CODE_FOR_nothing.
29919 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29920 Richard Biener <rguenther@suse.de>
29922 PR middle-end/55851
29923 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29924 types instead of just INTEGER_TYPE types.
29926 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29928 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29931 2013-01-07 Steve Ellcey <sellcey@mips.com>
29934 * config/mips/mips.opt: Change mad to mmad to match documentation.
29936 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29939 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29940 .progmemx.data now.
29942 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29945 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29946 (avr_addrspace_t): Add .section_name field.
29947 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29949 (avr_addrspace): Same. Initialize .section_name. Remove last
29950 NULL entry. Put __memx into .progmemx.data.
29951 (progmem_section_prefix): Remove.
29952 (avr_asm_init_sections): No need to initialize progmem_section.
29953 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29954 section name prefix.
29955 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29956 retrieve the progmem section.
29957 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29958 boundary to run over avr_addrspace[].
29959 (avr_register_target_pragmas): Ditto.
29961 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29963 * varasm.c (output_constant_def_contents): For asan_protect_global
29964 protected strings, adjust DECL_ALIGN if needed, before testing for
29966 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29967 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29969 (output_object_block): For asan protected decls, emit asan padding
29970 after their contents.
29971 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29972 (asan_finish_file): Test it here instead.
29974 2013-01-07 Nick Clifton <nickc@redhat.com>
29975 Matthias Klose <doko@debian.org>
29976 Doug Kwan <dougkwan@google.com>
29977 H.J. Lu <hongjiu.lu@intel.com>
29980 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29982 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29984 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29986 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29988 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29990 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29993 * doc/install.texi (Cross-Compiler-Specific Options): Document
29996 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29998 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29999 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
30000 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
30001 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
30002 vqmovun_high_s64): Fix source operand number and update copyright.
30004 2013-01-07 Richard Biener <rguenther@suse.de>
30006 PR middle-end/55890
30007 * gimple.h (gimple_call_builtin_p): New overload.
30008 * gimple.c (validate_call): New function.
30009 (gimple_call_builtin_p): Likewise.
30010 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
30011 Use gimple_call_builtin_p.
30012 (find_func_clobbers): Likewise.
30013 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
30014 (strlen_optimize_stmt): Likewise.
30016 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30018 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
30019 (vld1q_dup_*): Likewise.
30020 (vld1_*): Likewise.
30021 (vld1q_*): Likewise.
30022 (vld1_lane_*): Likewise.
30023 (vld1q_lane_*): Likewise.
30025 2013-01-07 Richard Biener <rguenther@suse.de>
30027 * lto-streamer.h (LTO_minor_version): Bump to 2.
30029 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30031 * config/aarch64/aarch64-protos.h
30032 (aarch64_const_double_zero_rtx_p): Rename to...
30033 (aarch64_float_const_zero_rtx_p): ...this.
30034 (aarch64_float_const_representable_p): New.
30035 (aarch64_output_simd_mov_immediate): Likewise.
30036 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
30037 move immediate case.
30038 * config/aarch64/aarch64.c
30039 (aarch64_const_double_zero_rtx_p): Rename to...
30040 (aarch64_float_const_zero_rtx_p): ...this.
30041 (aarch64_print_operand): Allow printing of new constants.
30042 (aarch64_valid_floating_const): New.
30043 (aarch64_legitimate_constant_p): Check for valid floating-point
30045 (aarch64_simd_valid_immediate): Likewise.
30046 (aarch64_vect_float_const_representable_p): New.
30047 (aarch64_float_const_representable_p): Likewise.
30048 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
30049 (aarch64_output_simd_mov_immediate): New.
30050 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
30051 (*movdf_aarch64): Likewise.
30052 * config/aarch64/constraints.md (Ufc): New.
30053 (Y): call aarch64_float_const_zero_rtx.
30054 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
30056 2013-01-07 Richard Biener <rguenther@suse.de>
30058 PR tree-optimization/55888
30059 PR tree-optimization/55862
30060 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
30061 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
30062 not if it is contained therein.
30064 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30066 * config/avr/t-avr: Typo.
30068 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30071 * config/avr/t-avr: Don't automatically rebuild
30072 $(srcdir)/config/avr/t-multilib
30073 $(srcdir)/config/avr/avr-tables.opt
30074 $(srcdir)/doc/avr-mmcu.texi
30075 (avr-mcus): New phony target to build them on request.
30076 (s-avr-mlib, s-avr-mmcu-texi): Remove.
30077 * avr/avr-mcus.def: Adjust comments.
30079 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
30081 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
30083 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30085 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
30087 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30089 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
30091 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
30093 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
30094 to generate profiling.
30095 * config/rs6000/aix64.h (LIB_SPEC): Same.
30097 2013-01-04 Andrew Pinski <apinski@cavium.com>
30099 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
30101 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
30103 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
30105 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
30107 (ix86_expand_move): Ditto.
30108 (ix86_zero_extend_to_Pmode): Ditto.
30109 (ix86_expand_call): Ditto.
30110 (ix86_expand_special_args_builtin): Ditto.
30111 (ix86_expand_builtin): Ditto.
30113 2013-01-04 Richard Biener <rguenther@suse.de>
30115 PR tree-optimization/55862
30116 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
30117 translating them through PHI nodes.
30119 2013-01-04 Martin Jambor <mjambor@suse.cz>
30121 PR tree-optimization/55755
30122 * tree-sra.c (sra_modify_assign): Do not check that an access has no
30123 children when trying to avoid producing a VIEW_CONVERT_EXPR.
30125 2013-01-04 Marek Polacek <polacek@redhat.com>
30127 PR middle-end/55859
30128 * opts.c (default_options_optimization): Clarify error message.
30130 2013-01-04 Richard Biener <rguenther@suse.de>
30132 PR middle-end/55863
30133 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
30136 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30139 * config/pa/pa.md (movsi): Revert previous change.
30140 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
30143 2013-01-03 Richard Henderson <rth@redhat.com>
30145 * config/i386/i386.c (ix86_expand_move): Always assign to op1
30146 after eliminating TLS symbols.
30148 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
30151 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
30152 * graphite-poly.c (debug_gmp_value): Likewise.
30154 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
30157 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
30158 selected code model, define __code_mode_small__, __code_model_medium__,
30159 __code_model_large__, __code_model_32__ or __code_model_kernel__.
30160 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
30161 xchg temporary register with %k. Declare temporary register as
30163 [__x86_64__]: For medium and large code models, preserve %rbx register.
30165 2013-01-03 Richard Biener <rguenther@suse.de>
30167 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
30168 (dump_subscript): Adjust.
30169 (finalize_ddr_dependent): Do not dump redundant info.
30170 (analyze_siv_subscript): Adjust.
30171 (subscript_dependence_tester): Likewise.
30172 (compute_affine_dependence): Likewise.
30174 2013-01-03 Richard Biener <rguenther@suse.de>
30177 2013-01-03 Richard Biener <rguenther@suse.de>
30179 PR tree-optimization/55857
30180 * tree-vect-stmts.c (vectorizable_load): Do not setup
30181 re-alignment for invariant loads.
30183 2013-01-02 Richard Biener <rguenther@suse.de>
30185 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30186 invariant load do not generate a vector load from the scalar location.
30188 2013-01-03 Richard Biener <rguenther@suse.de>
30190 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
30191 for not vectorizing.
30192 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
30193 not build INDIRECT_REFs, call get_name once only.
30194 (vect_create_data_ref_ptr): Likewise. Dump base object kind
30195 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
30197 2013-01-03 Richard Biener <rguenther@suse.de>
30199 PR tree-optimization/55857
30200 * tree-vect-stmts.c (vectorizable_load): Do not setup
30201 re-alignment for invariant loads.
30203 2013-01-03 Richard Biener <rguenther@suse.de>
30206 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
30207 prefer a built-in decl.
30209 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30211 * gcc.c (process_command): Update copyright notice dates.
30212 * gcov.c (print_version): Likewise.
30213 * gcov-dump.c (print_version): Likewise.
30215 PR rtl-optimization/55838
30216 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
30217 iv0.step, iv1.step and step.
30219 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30220 Marc Glisse <marc.glisse@inria.fr>
30222 PR tree-optimization/55832
30223 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
30224 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
30225 integer_{one,zero}_node.
30227 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30230 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
30231 * var-tracking.c (reverse_op): Don't add reverse ops to
30232 VALUEs that have already
30233 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
30235 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
30237 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
30239 2013-01-02 Teresa Johnson <tejohnson@google.com>
30241 * dumpfile.c (dump_loc): Print filename with location.
30242 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
30243 new location_t parameter to emit complete unroll message with
30244 new dump framework.
30245 (canonicalize_loop_induction_variables): Compute loops location
30246 and pass to try_unroll_loop_completely.
30247 * loop-unroll.c (report_unroll_peel): New function.
30248 (peel_loops_completely): Use new dump format with location
30249 for main dumpfile message, and invoke report_unroll_peel on success.
30250 (decide_unrolling_and_peeling): Ditto.
30251 (decide_peel_once_rolling): Remove old dumpfile message subsumed
30252 by report_unroll_peel.
30253 (decide_peel_completely): Ditto.
30254 (decide_unroll_constant_iterations): Ditto.
30255 (decide_unroll_runtime_iterations): Ditto.
30256 (decide_peel_simple): Ditto.
30257 (decide_unroll_stupid): Ditto.
30258 * cfgloop.c (get_loop_location): New function.
30259 * cfgloop.h (get_loop_location): Declare.
30261 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30263 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
30266 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30268 PR middle-end/55198
30269 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
30270 BLKmode objects when EXPAND_MEMORY is specified.
30272 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30274 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
30276 (fold_builtin_cpu): Do not share cpu model decls across statements.
30278 2013-01-02 Jason Merrill <jason@redhat.com>
30281 * tree.c (build_array_type_1): Revert earlier change.
30283 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
30285 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
30287 * config/aarch64/aarch64-tune.md: Re-generate.
30289 2013-01-02 Richard Biener <rguenther@suse.de>
30291 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30292 invariant load do not generate a vector load from the scalar location.
30294 2013-01-02 Richard Biener <rguenther@suse.de>
30297 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
30298 * configure: Regenerate.
30300 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30302 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
30303 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
30304 (expand_builtin_int_roundingfn_2): Keep the original target around
30305 for the fallback case.
30307 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30309 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
30310 to be clear for sign changes.
30312 2013-01-01 Jan Hubicka <jh@suse.cz>
30314 * ipa-inline-analysis.c: Fix formatting.
30316 2013-01-01 Jakub Jelinek <jakub@redhat.com>
30318 PR tree-optimization/55831
30319 * tree-vect-loop.c (get_initial_def_for_induction): Use
30320 gsi_after_labels instead of gsi_start_bb.
30322 Copyright (C) 2013 Free Software Foundation, Inc.
30324 Copying and distribution of this file, with or without modification,
30325 are permitted in any medium without royalty provided the copyright
30326 notice and this notice are preserved.