1 2013-11-12 Vladimir Makarov <vmakarov@redhat.com>
4 * ira-costs.c (record_operand_costs): Check operands number for
7 2013-11-12 Michael Meissner <meissner@linux.vnet.ibm.com>
10 * config/rs6000/rs6000.md (movdi_internal32): Eliminate
11 constraints that would allow DImode into the traditional Altivec
12 registers, but cause undesirable code generation when loading 0 as
14 (movdi_internal64): Likewise.
15 (cmp<mode>_fpr): Do not use %x for CR register output.
16 (extendsfdf2_fpr): Fix constraints when -mallow-upper-df and
17 -mallow-upper-sf debug switches are used.
19 2013-11-12 Andrew MacLeod <amacleod@redhat.com>
21 * gimple-expr.h (create_tmp_var_name, create_tmp_var_raw,
22 create_tmp_var, create_tmp_reg, mark_addressable, is_gimple_reg_rhs):
23 Relocate prototypes from gimple.h.
24 * gimplify.h: New File. Relocate some prototypes from gimple.h here.
25 (gimple_predicate, enum fallback, enum gimplify_status): Relocate
27 * gimple.h: Move some prototypes to gimplify.h.
28 (gimple_predicate, enum fallback, enum gimplify_status): Move to
30 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location):
31 Relocate from gimpify.c.
32 * gimple-expr.c (remove_suffix, tmp_var_id_num, create_tmp_var_name,
33 create_tmp_var_raw, create_tmp_var, create_tmp_reg, mark_addressable,
34 is_gimple_reg_rhs) Relocate from gimplify.c.
35 * gimplify.c (mark_addressable): Move to gimple-expr.c.
36 (gimple_seq_add_stmt_without_update): Move to gimple.c.
37 (remove_suffix, tmp_var_id_num, create_tmp_var_name, create_tmp_var_raw,
38 create_tmp_var, create_tmp_reg, is_gimple_reg_rhs): Move to
40 (should_carry_location_p): Move to gimple.c.
41 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location): Move
43 (annotate_one_with_location, annotate_all_with_location_after,
44 annotate_all_with_location): Move to gimple.c.
45 (compare_case_labels, sort_case_labels,
46 preprocess_case_label_vec_for_gimple): Move to gimple.c.
47 (rhs_predicate_for): Make static.
48 (gimplify_assign): Relocate from gimple.c.
49 * gimple.c (gimplify_assign): Move to gimplify.c.
50 (gimple_seq_add_stmt_without_update, should_carry_location_p,
51 annotate_one_with_location, annotate_all_with_location_after,
52 annotate_all_with_location, compare_case_labels, sort_case_labels,
53 preprocess_case_label_vec_for_gimple): Relocate from gimplify.c.
54 * tree.h (unshare_expr, unshare_expr_without_location,
55 mark_addressable): Move prototypes to gimplify.h.
56 * Makefile.in (GTFILES): gimple-expr.c now has the GTY tag for
58 * asan.c: Include gimplify.h rather than gimple.h.
59 * cfgloopmanip.c: Likewise.
60 * cgraphunit.c: Likewise.
61 * cilk-common.c: Likewise.
62 * dwarf2out.c: Dont include gimple.h.
63 * fold-const.c: Include gimplify.h rather than gimple.h.
64 * function.c: Likewise.
65 * gimple-fold.c: Likewise.
66 * gimple-ssa-strength-reduction.c: Likewise.
67 * graphite-clast-to-gimple.c: Likewise.
68 * graphite-sese-to-poly.c: Likewise.
69 * ipa-prop.c: Likewise.
70 * ipa-split.c: Likewise.
72 * langhooks.c: Dont include gimple.h.
73 * loop-init.c: Include gimplify.h rather than gimple.h.
74 * omp-low.c: Likewise.
76 * stor-layout.c: Likewise.
77 * targhooks.c: Likewise.
78 * trans-mem.c: Likewise.
79 * tree-affine.c: Likewise.
80 * tree-cfg.c: Likewise.
81 * tree-cfgcleanup.c: Likewise.
82 * tree-complex.c: Likewise.
83 * tree-if-conv.c: Likewise.
84 * tree-inline.c: Likewise.
85 * tree-iterator.c: Likewise.
86 * tree-loop-distribution.c: Likewise.
87 * tree-nested.c: Likewise.
88 * tree-parloops.c: Likewise.
89 * tree-predcom.c: Likewise.
90 * tree-profile.c: Likewise.
91 * tree-scalar-evolution.c: Likewise.
92 * tree-sra.c: Likewise.
93 * tree-ssa-address.c: Likewise.
94 * tree-ssa-ccp.c: Likewise.
95 * tree-ssa-dce.c: Likewise.
96 * tree-ssa-forwprop.c: Likewise.
97 * tree-ssa-ifcombine.c: Likewise.
98 * tree-ssa-loop-im.c: Likewise.
99 * tree-ssa-loop-ivopts.c: Likewise.
100 * tree-ssa-loop-manip.c: Likewise.
101 * tree-ssa-loop-niter.c: Likewise.
102 * tree-ssa-loop-prefetch.c: Likewise.
103 * tree-ssa-loop-unswitch.c: Likewise.
104 * tree-ssa-math-opts.c: Likewise.
105 * tree-ssa-phiopt.c: Likewise.
106 * tree-ssa-phiprop.c: Likewise.
107 * tree-ssa-pre.c: Likewise.
108 * tree-ssa-propagate.c: Likewise.
109 * tree-ssa-reassoc.c: Likewise.
110 * tree-ssa-sccvn.c: Likewise.
111 * tree-ssa-strlen.c: Likewise.
112 * tree-ssa.c: Likewise.
113 * tree-switch-conversio: Likewise.n.c
114 * tree-tailcall.c: Likewise.
115 * tree-vect-data-refs.c: Likewise.
116 * tree-vect-generic.c: Likewise.
117 * tree-vect-loop-manip.c: Likewise.
118 * tree-vect-loop.c: Likewise.
119 * tree-vect-patterns.c: Likewise.
120 * tree-vect-stmts.c: Likewise.
122 * value-prof.c: Likewise.
123 * config/aarch64/aarch64.c: Include gimplify.h instead of gimple.h.
124 * config/alpha/alpha.c: Likewise.
125 * config/darwin.c: Likewise.
126 * config/i386/i386.c: Likewise.
127 * config/ia64/ia64.c: Likewise.
128 * config/mep/mep.c: Likewise.
129 * config/mips/mips.c: Likewise.
130 * config/rs6000/rs6000.c: Likewise.
131 * config/s390/s390.c: Likewise.
132 * config/sh/sh.c: Likewise.
133 * config/sparc/sparc.c: Likewise.
134 * config/spu/spu.c: Likewise.
135 * config/stormy16/stormy16.c: Likewise.
136 * config/tilegx/tilegx.c: Likewise.
137 * config/tilepro/tilepro.c: Likewise.
138 * config/xtensa/xtensa.c: Likewise.
140 2013-11-12 Adam Butcher <adam@jessamine.co.uk>
142 * tree.c (grow_tree_vec_stat): New function ...
143 * tree.h (grow_tree_vec_stat) (grow_tree_vec): ... and its declaration
146 2013-11-12 Marek Polacek <polacek@redhat.com>
148 * final.c (update_alignments): Initialize label to NULL_RTX.
150 2013-11-12 Jeff Law <law@redhat.com>
152 * gimple-ssa-isolate-paths.c (check_loadstore): New function.
153 (insert_trap_and_remove_trailing_statements): New argument OP which
154 is the NULL pointer. Emit the trap after the load/store through
155 the NULL pointer. Simplify the RHS of a store through a NULL pointer
156 when trivial to do so.
157 (isolate_path): Corresponding changes.
158 (gimple_ssa_isolate_erroneous_path): Likewise.
160 2013-11-12 Teresa Johnson <tejohnson@google.com>
161 Jan Hubicka <jh@suse.cz>
163 * predict.c (drop_profile): New function.
164 (handle_missing_profiles): Ditto.
165 (counts_to_freqs): Don't overwrite estimated frequencies
166 when function has no profile counts.
167 * predict.h (handle_missing_profiles): Declare.
168 * tree-inline.c (freqs_to_counts): New function.
169 (copy_cfg_body): Invoke freqs_to_counts as needed.
170 * tree-profile.c (tree_profiling): Invoke handle_missing_profiles.
172 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
175 * config/i386/x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL):
177 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Set for m_HASWELL.
179 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
182 * config/i386/i386.c (ix86_option_override_internal): Check
183 X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL and
184 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL for
185 MASK_AVX256_SPLIT_UNALIGNED_LOAD and
186 MASK_AVX256_SPLIT_UNALIGNED_STORE.
188 * config/i386/x86-tune.def (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL):
189 Clear m_COREI7_AVX and update comments.
190 (X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): Likewise.
192 2013-11-12 Martin Jambor <mjambor@suse.cz>
194 PR rtl-optimization/10474
195 * ira.c (interesting_dest_for_shprep): New function.
196 (split_live_ranges_for_shrink_wrap): Likewise.
197 (find_moveable_pseudos): Move calculation of dominance info,
198 df_analysios and the final anlyses to...
199 (ira): ...here, call split_live_ranges_for_shrink_wrap.
201 2013-11-12 Bin Cheng <bin.cheng@arm.com>
203 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Refactor the code.
204 Handle type conversion.
206 2013-11-11 Martin Liska <marxin.liska@gmail.com>
207 Jan Hubicka <jh@suse.cz>
209 * cgraph.c (dump_cgraph_node): Profile dump added.
210 * cgraph.h (struct cgraph_node): New time profile variable added.
211 * cgraphclones.c (cgraph_clone_node): Time profile is cloned.
212 * gcov-io.h (gcov_type): New profiler type introduced.
213 * ipa-profile.c (lto_output_node): Streaming for time profile added.
214 (input_node): Time profiler is read from LTO stream.
215 * predict.c (maybe_hot_count_p): Hot prediction changed.
216 * profile.c (instrument_values): New case for time profiler added.
217 (compute_value_histograms): Read of time profile.
218 * tree-pretty-print.c (dump_function_header): Time profiler is dumped.
219 * tree-profile.c (init_ic_make_global_vars): Time profiler function added.
220 (gimple_init_edge_profiler): TP function instrumentation.
221 (gimple_gen_time_profiler): New.
222 * value-prof.c (gimple_add_histogram_value): Support for time profiler
224 (dump_histogram_value): TP type added to dumps.
225 (visit_hist): More sensitive check that takes TP into account.
226 (gimple_find_values_to_profile): TP instrumentation.
227 * value-prof.h (hist_type): New histogram type added.
228 (struct histogram_value_t): Pointer to struct function added.
229 * libgcc/Makefile.in: New GCOV merge function for TP added.
230 * libgcov.c: function_counter variable introduced.
231 (_gcov_merge_time_profile): New.
232 (_gcov_time_profiler): New.
234 2013-11-11 Marc Glisse <marc.glisse@inria.fr>
235 Jeff Law <law@redhat.com>
237 * tree-ssa-alias.c (stmt_kills_ref_p_1): Use
238 ao_ref_init_from_ptr_and_size for builtins.
240 2013-11-11 Uros Bizjak <ubizjak@gmail.com>
241 H.J. Lu <hongjiu.lu@intel.com>
244 * config/i386/x86-tune.def
245 (X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
246 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES.
248 (TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
249 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES. Update for renamed
250 X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES.
251 * config/i386/i386.c (ix86_expand_set_or_movmem): Use
252 TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES to calculate
253 misaligned_prologue_used. Check that
254 desired_aling <= epilogue_size_needed.
256 2013-11-11 Cong Hou <congh@google.com>
258 PR tree-optimization/59050
259 * tree-vect-data-refs.c (comp_dr_addr_with_seg_len_pair): Bug fix.
261 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
264 * expmed.c (emit_store_flag): Fail for const-const comparison.
266 2013-11-11 Tristan Gingold <gingold@adacore.com>
267 Eric Botcazou <ebotcazou@adacore.com>
269 * tree.h (CONSTRUCTOR_NO_CLEARING): Define.
270 * tree-core.h (CONSTRUCTOR_NO_CLEARING): Document it.
271 * tree.def (CONSTRUCTOR): Likewise.
272 * doc/generic.texi (CONSTRUCTOR): Likewise. Update description.
273 * gimplify.c (gimplify_init_constructor): Do not clear the object when
274 the constructor is incomplete and CONSTRUCTOR_NO_CLEARING is set.
276 2013-11-11 Basile Starynkevitch <basile@starynkevitch.net>
278 * toplev.c (toplev_main): Move PLUGIN_FINISH invocation before
281 2013-11-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
283 * config/arm/arm.c (arm_new_rtx_costs): Return after handling
286 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
288 * config/arc/arc.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
290 2013-11-08 Jeff Law <law@redhat.com>
292 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
293 threading paths first, then perform PHI node checks if applicable.
295 2013-11-10 Karlson2k <k2k@narod.ru>
296 Kai Tietz <ktietz@redhat.com>
299 * configure.ac: Adding for exported symbols check
300 and for rdynamic-check executable-extension.
301 * configure: Regenerated.
303 2013-11-10 Uros Bizjak <ubizjak@gmail.com>
305 * mode-switching.c (optimize_mode_switching): Mark block as
306 nontransparent, if last_mode at block exit is different from no_mode.
308 2013-11-09 Jan-Benedict Glaw <jbglaw@lug-owl.de>
310 * function.c (NAME__MAIN): Move to...
311 * cfgexpand.c (NAME__MAIN): ...here.
313 2013-11-09 Richard Sandiford <rdsandiford@googlemail.com>
315 * target.def (can_use_doloop_p): New hook.
316 * doc/tm.texi.in (TARGET_CAN_USE_DOLOOP_P): Add.
317 * doc/tm.texi: Regenerate.
318 * doc/md.texi (doloop_begin, doloop_end): Update documentation.
319 * hooks.h (hook_bool_dint_dint_uint_true): Declare.
320 * hooks.c (hook_bool_dint_dint_uint_true): New function.
321 * targhooks.h (can_use_doloop_if_innermost): Declare.
322 * targhooks.c (can_use_doloop_if_innermost): New function.
323 * target.h: Include double-int.h.
324 * loop-doloop.c (doloop_optimize): Call targetm.can_use_doloop_p.
325 Remove iteration count, maximum iteration count, loop depth and
326 enter-at-top inputs from doloop_begin and doloop_end.
327 * config/arc/arc.md (doloop_begin, doloop_end): Update for new
329 * config/arc/arc.c (arc_can_use_doloop_p): New function.
330 (TARGET_CAN_USE_DOLOOP_P): Define.
331 * config/arm/thumb2.md (doloop_end): Update for new interface.
332 * config/arm/arm.c (TARGET_CAN_USE_DOLOOP_P): Define.
333 * config/bfin/bfin.md (doloop_end): Update for new interface.
334 * config/bfin/bfin.c (bfin_can_use_doloop_p): New function.
335 (TARGET_CAN_USE_DOLOOP_P): Define.
336 * config/c6x/c6x.md (doloop_end): Update for new interface.
337 * config/ia64/ia64.md (doloop_end): Update for new interface.
338 * config/ia64/ia64.c (TARGET_CAN_USE_DOLOOP_P): Define.
339 * config/mep/mep.md (doloop_begin, doloop_end): Update for new
341 * config/mep/mep.c (mep_emit_doloop): Likewise.
342 (TARGET_CAN_USE_DOLOOP_P): Define.
343 * config/rs6000/rs6000.md (doloop_end): Update for new interface.
344 * config/rs6000/rs6000.c (TARGET_CAN_USE_DOLOOP_P): Define.
345 * config/s390/s390.md (doloop_end): Update for new interface.
346 * config/sh/sh.md (doloop_end): Likewise.
347 * config/spu/spu.md (doloop_end): Likewise.
348 * config/spu/spu.c (TARGET_CAN_USE_DOLOOP_P): Define.
349 * config/tilegx/tilegx.md (doloop_end): Update for new interface.
350 * config/tilegx/tilegx.c (TARGET_CAN_USE_DOLOOP_P): Define.
351 * config/tilepro/tilepro.md (doloop_end): Update for new interface.
352 * config/tilepro/tilepro.c (TARGET_CAN_USE_DOLOOP_P): Define.
353 * config/v850/v850.md (doloop_begin, doloop_end): Update for new
355 * config/v850/v850.c (TARGET_CAN_USE_DOLOOP_P): Define.
357 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
360 * doc/extend.texi: Move Cilk Plus Builtins node before Other
363 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
364 Joseph Myers <joseph@codesourcery.com>
366 * ginclude/stdatomic.h: New file.
367 * Makefile.in (USER_H): Add stdatomic.h.
369 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
371 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
374 2013-11-08 Jeff Law <law@redhat.com>
376 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
377 * tree-ssa-threadupdate.c (delete_thread_path): New function.
378 (ssa_redirect_edges, thread_block_1): Use it.
379 (thread_through_loop_header, mark_threaded_blocks): Likewise.
380 (thread_through_all_blocks, register_jump_thread): Likewise.
381 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
383 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
385 * config/arm/aarch-common.c
386 (search_term): New typedef.
387 (shift_rtx_costs): New array.
388 (arm_rtx_shift_left_p): New.
389 (arm_find_sub_rtx_with_search_term): Likewise.
390 (arm_find_sub_rtx_with_code): Likewise.
391 (arm_early_load_addr_dep): Add sanity checking.
392 (arm_no_early_alu_shift_dep): Likewise.
393 (arm_no_early_alu_shift_value_dep): Likewise.
394 (arm_no_early_mul_dep): Likewise.
395 (arm_no_early_store_addr_dep): Likewise.
397 2013-11-08 Richard Biener <rguenther@suse.de>
399 PR tree-optimization/59047
400 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses properly.
402 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
404 * common.opt (fcheck-pointer-bounds): Move to ...
405 * c-family/c.opt: ... here.
406 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
407 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
408 * langhooks.h (lang_hooks): Remove chkp_supported field.
409 * toplev.c (process_options): Remove chkp_supported check.
411 2013-11-08 Richard Biener <rguenther@suse.de>
413 PR tree-optimization/59038
414 PR tree-optimization/58955
415 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
416 previous change. Handle known dependences correctly.
418 2013-11-08 Tom de Vries <tom@codesourcery.com>
420 * config/rs6000/t-xilinx: Remove duplicate contents.
422 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
423 Joseph Myers <joseph@codesourcery.com>
425 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
426 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
427 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
428 (struct tree_base): Add atomic_flag field.
429 * tree.h (TYPE_ATOMIC): New accessor macro.
430 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
431 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
432 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
433 (atomicDI_type_node, atomicTI_type_node): New macros for type nodes.
434 * tree.c (set_type_quals): Set TYPE_ATOMIC.
435 (find_atomic_core_type): New function.
436 (build_qualified_type): Adjust alignment for qualified types.
437 (build_atomic_base): New function
438 (build_common_tree_nodes): Build atomicQI_type_node,
439 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
441 * print-tree.c (print_node): Print atomic qualifier.
442 * tree-pretty-print.c (dump_generic_node): Print atomic type attribute.
443 * target.def (atomic_assign_expand_fenv): New hook.
444 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
445 * doc/tm.texi: Regenerate.
446 * targhooks.c (default_atomic_assign_expand_fenv): New function.
447 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
448 * sync-builtins.def (__atomic_feraiseexcept): New built-in function.
449 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
451 * config/i386/i386.c (enum ix86_builtins): Add
452 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
454 (bdesc_special_args): Add __builtin_ia32_fnstenv,
455 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and __builtin_ia32_fnclex.
456 (ix86_expand_builtin): Handle the new built-in functions.
457 (ix86_atomic_assign_expand_fenv): New function.
458 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
459 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
460 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
461 (fnstenv, fldenv, fnstsw, fnclex): New insns.
463 2013-11-07 Steve Ellcey <sellcey@mips.com>
465 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
466 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
467 (MULTILIB_DIRNAMES): Add fp64 directory.
468 (MULTILIB_EXCEPTIONS): Add new exclusions.
470 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
472 * gimplify.c (gimple_regimplify_operands): Do not set
474 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
475 (rewrite_close_phi_out_of_ssa): Same.
476 (rewrite_phi_out_of_ssa): Same.
477 (rewrite_degenerate_phi): Same.
478 (handle_scalar_deps_crossing_scop_limits): Same.
479 * tree-if-conv.c (predicate_scalar_phi): Same.
480 * tree-parloops.c (create_loads_for_reductions): Same.
481 (create_final_loads_for_reduction): Same.
482 (create_loads_and_stores_for_name): Same.
483 (transform_to_exit_first_loop): Same.
484 (create_parallel_loop): Same.
486 (move_computations_dom_walker::before_dom_children): Same.
487 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
488 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Same.
489 * tree-ssa-propagate.c (substitute_and_fold): Same.
490 * tree-vect-loop.c (vect_finalize_reduction): Same.
491 * tree-vect-stmts.c (vectorizable_call): Same.
493 2013-11-07 Mike Stump <mikestump@comcast.net>
495 * config/pdp11/pdp11.c: Include dbxout.h.
496 * config/picochip/picochip.c: Likewise.
498 2013-11-07 Cong Hou <congh@google.com>
500 PR tree-optimization/56764
501 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
502 Combine alias checks if it is possible to amortize the runtime
503 overhead. Return the number of alias checks after merging.
504 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
505 Use the function vect_create_cond_for_alias_checks () to check
506 the number of alias checks.
508 2013-11-07 Jeff Law <law@redhat.com>
510 * varpool.c (ctor_for_folding): Fix typo in comment.
512 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
514 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
515 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
517 2013-11-07 Jeff Law <law@redhat.com>
519 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
521 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
522 No longer check if we have __builtin_trap, assume it's available.
524 2013-11-07 Diego Novillo <dnovillo@google.com>
526 * attribs.c (lookup_scoped_attribute_spec): Make static.
527 (get_attribute_namespace): Likewise.
528 * builtins.c (more_const_call_expr_args_p): Move from tree.h.
529 (validate_arglist): Move earlier in the file. Make static.
530 (expand_stack_restore): Move from stmt.c
531 (expand_stack_save): Move from stmt.c
532 (rewrite_call_expr_array): Move earlier in the file.
533 (rewrite_call_expr_valist): Likewise.
534 * cfgexpand.c: Include hard-reg-set.h before tree.h
537 (expand_asm_loc): Move from stmt.c.
538 (n_occurrences): Move from stmt.c.
539 (check_operand_nalternatives): Move from stmt.c.
540 (tree_conflicts_with_clobbers_p): Move from stmt.c.
541 (expand_asm_operands): Move from stmt.c
542 (expand_asm_stmt): Move from stmt.c
543 (expand_computed_goto): Move from stmt.c
544 (expand_goto): Move from stmt.c
545 (expand_null_return_1): Move from stmt.c
546 (expand_null_return): Move from stmt.c
547 (expand_value_return): Move from stmt.c
548 (expand_return): Move from stmt.c
549 (expand_main_function): Move from function.c
550 (stack_protect_prologue): Move from function.c
551 * cgraphclones.c (build_function_type_skip_args): Move from tree.c.
552 (build_function_decl_skip_args): Move from tree.c.
553 * explow.c (tree_expr_size): Move from tree.c.
554 * expr.c (addr_expr_of_non_mem_decl_p): Remove.
555 (fields_length): Move from tree.c.
556 * fold-const.c (size_low_cst): Move from tree.c.
557 (tree_expr_nonzero_warnv_p): Make static. Move earlier in the file.
558 (tree_expr_nonzero_p): Make static. Move earlier in the file.
559 (fold_build3_initializer_loc): Remove.
560 (tree_invalid_nonnegative_warnv_p): Make static.
561 * function.c (expand_main_function): Move to cfgexpand.c.
562 (stack_protect_prologue): Move to cfgexpand.c.
563 (set_insn_locations): Move earlier in the file.
564 * gimple-fold.c: Include langhooks.h.
565 (truth_type_for): Move from tree.c.
566 * print-tree.c (print_vec_tree): Remove.
567 * stmt.c (expand_computed_goto): Move to cfgexpand.c.
568 (expand_goto): Move to cfgexpand.c.
569 (n_occurrences): Move to cfgexpand.c.
570 (expand_asm_loc): Move to cfgexpand.c
571 (tree_conflicts_with_clobbers_p): Move to cfgexpand.c.
572 (expand_asm_operands): Move to cfgexpand.c.
573 (expand_asm_stmt): Move to cfgexpand.c.
574 (check_operand_nalternatives): Move to cfgexpand.c
575 (expand_null_return): Move to cfgexpand.c.
576 (expand_value_return): Move to cfgexpand.c.
577 (expand_null_return_1): Move to cfgexpand.c.
578 (expand_return): Move to cfgexpand.c.
579 (expand_stack_save): Move to builtins.c.
580 (expand_stack_restore): Move to builtins.c
581 * symtab.c: Include output.h.
582 (decl_assembler_name_hash): Move from tree.c.
583 (decl_assembler_name_equal): Move from tree.c.
584 * trans-mem.c (is_tm_safe_or_pure): Move from tree.h.
585 * tree-eh.c (in_array_bounds_p): Move from tree.c.
586 (range_in_array_bounds_p): Move from tree.c.
587 * tree-object-size.c (fini_object_sizes): Make static.
588 * tree-ssa-dom.c (iterative_hash_exprs_commutative): Move from tree.h.
589 * tree-vrp.c (ssa_name_nonnegative_p): Remove.
590 * tree.c (decl_assembler_name_equal): Move to symtab.c.
591 (tree_expr_size): Move to explow.c.
592 (decl_assembler_name_hash): Move to symtab.c.
594 (tree_expr_size): Move to explow.c.
595 (stabilize_reference_1): Move earlier in the file. Make static.
596 (omp_remove_redundant_declare_simd_attrs): Remove.
597 (simple_cst_list_equal): Move earlier in the file. Make static.
598 (size_low_cst): Move to fold-const.c.
599 (build_type_no_quals): Remove.
600 (build_function_type_skip_args): Move to cgraphclones.c.
601 (build_function_decl_skip_args): Move to cgraphclones.c.
602 (in_array_bounds_p): Move to tree-eh.c.
603 (range_in_array_bounds_p): Move to tree-eh.c.
604 (truth_type_for): Move to gimple-fold.c.
605 (list_equal_p): Remove.
606 * tree.h (decl_assembler_name_equal): Remove.
607 (decl_assembler_name_hash): Remove.
608 (truth_type_for): Remove.
609 (build_type_no_quals): Remove.
610 (build_function_decl_skip_args): Remove.
611 (in_array_bounds_p): Remove.
612 (range_in_array_bounds_p): Remove.
613 (size_low_cst): Remove.
614 (omp_remove_redundant_declare_simd_attrs): Remove.
615 (tree_expr_size): Remove.
616 (fields_length): Remove.
617 (stabilize_reference_1): Remove.
618 (expand_goto): Remove.
619 (expand_stack_save): Remove.
620 (expand_stack_restore): Remove.
621 (expand_return): Remove.
622 (fold_build3_initializer_loc): Remove.
623 (tree_expr_nonzero_p): Remove.
624 (tree_invalid_nonnegative_warnv_p): Remove.
625 (tree_expr_nonzero_warnv_p): Remove.
626 (fold_builtin_snprintf_chk): Remove.
627 (validate_arglist): Remove.
628 (iterative_hash_exprs_commutative): Move to tree-ssa-dom.c.
629 (simple_cst_list_equal): Remove.
631 (expand_main_function): Remove.
632 (stack_protect_prologue): Remove.
633 (print_vec_tree): Remove.
634 (lookup_scoped_attribute_spec): Remove.
635 (get_attribute_namespace): Remove.
636 (expand_computed_goto): Remove.
637 (expand_asm_stmt): Remove.
638 (list_equal_p): Remove.
639 (ssa_name_nonnegative_p): Remove.
640 (fini_object_sizes): Remove.
641 (addr_expr_of_non_mem_decl_p): Remove.
642 (is_tm_safe_or_pure): Move to trans-mem.c.
643 (more_const_call_expr_args_p): Remove.
644 (save_vtable_map_decl): Remove.
646 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
648 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
651 * config/arc/arc.h (LINK_COMMAND_SPEC): For -ftree-parallelize-loops=*,
652 link to libgomp and its dependencies.
653 * config/ia64/hpux.h (LIB_SPEC): Likewise.
654 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
655 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
656 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's dependencies.
658 2013-11-07 Jakub Jelinek <jakub@redhat.com>
660 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
661 (determine_value_range): Add loop argument. Use get_range_info to
663 (bound_difference): Adjust caller.
665 2013-11-07 Richard Biener <rguenther@suse.de>
666 Jakub Jelinek <jakub@redhat.com>
668 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
669 latches from header PHI arguments from the latch edge.
671 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
674 * varasm.c (output_constant): Handle NULLPTR_TYPE.
676 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
678 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
679 misaligned_prologue_used when it has been set.
681 2013-11-07 Yury Gribov <y.gribov@samsung.com>
682 Jakub Jelinek <jakub@redhat.com>
685 * asan.c (get_mem_refs_of_builtin_call): Allow
686 integer literals as addresses in instrumented builtins.
688 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
690 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
691 Explain why plus_constant is not used.
693 2013-11-07 Richard Biener <rguenther@suse.de>
695 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
696 (canonicalize_value): ... this. Also handle stripping of
698 (get_value, set_lattice_value, get_value_for_expr): Adjust.
699 * gimple-fold.c (canonicalize_constructor_val): Strip TREE_OVERFLOW.
700 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
702 2013-11-07 Richard Biener <rguenther@suse.de>
704 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
706 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
709 * config/i386/i386.md (push peepholer/splitter): Use Pmode
710 with stack_pointer_rtx.
712 2013-11-07 Bin Cheng <bin.cheng@arm.com>
714 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
715 using operand_equal_p.
717 2013-11-07 Bin Cheng <bin.cheng@arm.com>
719 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
720 * tree-affine.c (get_inner_reference_aff): Return base.
721 * tree-affine.h (get_inner_reference_aff): Change prototype.
723 2013-11-06 Tobias Burnus <burnus@net-b.de>
725 * doc/invoke.texi (Wdate-time): Fix typo.
727 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
729 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
730 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
731 fp_arith_reg_operand instead of arith_reg_operand.
733 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
735 * config/sh/sh.md (adddi3): Remove empty constraints.
736 Remove can_create_pseudo_p and arith_reg_operand check.
737 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
740 2013-11-06 Jeff Law <law@redhat.com>
741 Tom Tromey <tromey@redhat.com>
743 * gdbinit.in: Disable strict type checking.
745 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
747 * tree-pass.h (make_pass_live_range_shrinkage): New external.
748 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
749 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
750 (rest_of_handle_live_range_shrinkage): Ditto
751 (class pass_live_range_shrinkage): Ditto.
752 (pass_data_live_range_shrinkage): Ditto.
753 (make_pass_live_range_shrinkage): Ditto.
754 * sched-int.h (initialize_live_range_shrinkage): New prototype.
755 (finish_live_range_shrinkage): Ditto.
756 * sched-deps.c (create_insn_reg_set): Make void return value.
757 * passes.def: Add pass_live_range_shrinkage.
758 * ira.c (update_equiv_regs): Don't move if flag_live_range_shrinkage.
759 * haifa-sched.c (live_range_shrinkage_p): New.
760 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
762 (rank_for_schedule): Add code for pressure relief through live
764 (schedule_insn): Print more debug info.
765 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
766 through live range shrinkage.
767 * doc/invoke.texi (-flive-range-shrinkage): New.
768 * common.opt (flive-range-shrinkage): New.
770 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
773 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
774 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
775 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
776 RTXes that return in AVX256 register.
778 2013-11-06 Richard Biener <rguenther@suse.de>
780 PR tree-optimization/58653
781 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
782 (prepare_initializers_chain): Adjust.
784 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
786 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
787 (gimple_in_transaction): Use bb_in_transaction.
788 * basic-block.h (bb_in_transaction): Relocate here and rename.
789 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
791 2013-11-06 Richard Biener <rguenther@suse.de>
793 * tree.c (drop_tree_overflow): New function.
794 * tree.h (drop_tree_overflow): Declare.
795 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
796 * tree-vrp.c (range_int_cst_singleton_p): Use
797 is_overflow_infinity instead of testing TREE_OVERFLOW.
798 (extract_range_from_assert): Likewise.
799 (zero_nonzero_bits_from_vr): Likewise.
800 (extract_range_basic): Likewise.
801 (register_new_assert_for): Use drop_tree_overflow.
802 (vrp_visit_phi_node): Likewise.
804 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
806 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
807 checking for leaf functions without dynamic stack allocation.
808 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
809 (ia64_expand_prologue): Likewise.
810 * config/mips/mips.c (mips_expand_prologue): Likewise.
811 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
812 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
813 (sparc_flat_expand_prologue): Likewise.
815 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
817 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
818 (__ST3_LANE_FUNC): Likewise.
819 (__ST4_LANE_FUNC): Likewise.
821 2013-11-06 Nick Clifton <nickc@redhat.com>
823 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
824 name returned by msp430_mcu_name.
825 (LIB_SPEC): If a -T option has not been specified then set a
826 default, mcu-specific, linker script.
827 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
828 * config/msp430/msp430.c (msp430x_names): Likewise.
829 Alpha sort the names for ease of comparison.
830 (msp430_mcu_name): New function: Returns a string suitable for
831 use as a C preprocessor symbol based upon the name of the MCU
833 (msp430_option_override): Accept msp430x and msp430xv2 as generic
835 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
837 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
838 unless it is the last part of a spec substring.
839 * doc/invoke.texi (Spec Files): Document space insertion
842 2013-11-06 Christian Bruel <christian.bruel@st.com>
844 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
845 Factorize probabilities, Use adjust_address instead of
846 adjust_automodify_address when possible. Enable for optimize.
847 (sh_expand_strlen): New function.
848 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
849 * config/sh/sh.md (strlensi): New pattern.
850 (UNSPEC_BUILTIN_STRLEN): Define.
852 2013-11-06 Jakub Jelinek <jakub@redhat.com>
855 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
856 (expand_assignment): If *bitpos is negative, set *offset
857 and adjust *bitpos, so that it is not negative.
859 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
861 * config/i386/bdver3.md : Added two additional decoder units
862 to support issue rate of 4 and remodeled vector unit.
863 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
864 architectures is set to 4.
865 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
866 lookahead is set to 4 for BD architectures.
868 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
870 * config/rs6000/rs6000.c (rs6000_option_override_internal):
871 Remove restriction against use of VSX instructions when generating
872 code for little endian mode.
874 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
876 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
877 for both big and little endian.
878 (mulv8hi3): Swap input operands for merge high and merge low
879 instructions for little endian.
881 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
883 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
884 define_insn to define_expand that uses even patterns for big
885 endian and odd patterns for little endian.
886 (vec_widen_smult_even_v16qi): Likewise.
887 (vec_widen_umult_even_v8hi): Likewise.
888 (vec_widen_smult_even_v8hi): Likewise.
889 (vec_widen_umult_odd_v16qi): Likewise.
890 (vec_widen_smult_odd_v16qi): Likewise.
891 (vec_widen_umult_odd_v8hi): Likewise.
892 (vec_widen_smult_odd_v8hi): Likewise.
893 (altivec_vmuleub): New define_insn.
894 (altivec_vmuloub): Likewise.
895 (altivec_vmulesb): Likewise.
896 (altivec_vmulosb): Likewise.
897 (altivec_vmuleuh): Likewise.
898 (altivec_vmulouh): Likewise.
899 (altivec_vmulesh): Likewise.
900 (altivec_vmulosh): Likewise.
902 2013-11-05 Mike Stump <mikestump@comcast.net>
904 * Makefile.in (mostlyclean): Remove c-family objects.
906 2013-11-05 Ian Lance Taylor <iant@google.com>
908 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
909 If possible, add .cfi directives to record change to bx.
910 * config/i386/i386.c (ix86_emit_cfi): New function.
911 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
913 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
915 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
916 immediately after a label for a tablejump pattern.
918 * config/arm/arm.c (is_jump_table): Remove.
919 (create_fix_barrier): Use tablejump_p instead.
920 (arm_reorg): Likewise.
921 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
922 (thumb2_output_casesi): Likewise.
923 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
924 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
925 casesi_shift_media, casesi_load_media): Likewise.
926 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
927 * config/microblaze/microblaze.md: Likewise.
929 2013-11-05 Tobias Burnus <burnus@net-b.de>
931 * doc/invoke.texi (-Wdate-time): Document.
933 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
935 * double-int.c (lshift_double, rshift_double): Remove
936 SHIFT_COUNT_TRUNCATED handling.
938 2013-11-05 Jeff Law <law@redhat.com>
940 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
941 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
942 * gimple-ssa-isolate-paths.c: New file.
943 * gimple.c (check_loadstore): New function.
944 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
945 Verify OP is in the argument list and the argument corresponding
946 to OP is a pointer type. Use operand_equal_p rather than
947 pointer equality when testing if OP is on the nonnull list.
948 Use check_loadstore rather than count_ptr_derefs. Handle
949 GIMPLE_RETURN statements.
950 * tree-vrp.c (infer_nonnull_range): Remove.
951 * gimple.h (infer_nonnull_range): Declare.
952 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
953 * passes.def: Add pass_isolate_erroneous_paths.
954 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
955 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
956 * tree-ssa.c (struct count_ptr_d): Remove.
957 (count_ptr_derefs, count_uses_and_derefs): Remove.
958 * tree-ssa.h (count_uses_and_derefs): Remove.
960 2013-11-05 Jakub Jelinek <jakub@redhat.com>
962 PR rtl-optimization/58997
963 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
964 get_iv_value to be in iv->mode rather than iv->extend_mode.
965 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
966 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
967 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
970 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
972 * gimple.h: Move some prototypes to gimple-expr.h and add to include
974 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
975 Move to gimple-expr.h.
976 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
977 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
978 is_gimple_id, virtual_operand_p, is_gimple_addressable,
979 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
981 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
982 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
983 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
984 is_gimple_address, is_gimple_invariant_address,
985 is_gimple_ip_invariant_address, is_gimple_min_invariant,
986 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
987 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
988 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
989 gimple_decl_printable_name, useless_type_conversion_p,
990 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
992 * gimple-expr.c: New File.
993 (useless_type_conversion_p, gimple_set_body, gimple_body,
994 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
995 gimple_can_coalesce_p, extract_ops_from_tree_1,
996 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
997 is_gimple_address, is_gimple_invariant_address,
998 is_gimple_ip_invariant_address, is_gimple_min_invariant,
999 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
1000 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
1001 is_gimple_mem_ref_addr): Relocate here.
1002 * Makefile.in (OBJS): Add gimple-expr.o.
1004 2013-11-05 David Malcolm <dmalcolm@redhat.com>
1006 * gengtype-parse.c (struct_field_seq): Support empty structs.
1008 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1010 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
1012 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1014 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
1016 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
1018 (LONG_DOUBLE_TYPE_SIZE): New define.
1019 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
1021 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
1024 * doc/extend.texi [visibility ("visibility_type")]: Add example
1025 about visibility attribute on namespace declaration.
1027 2013-11-05 Richard Biener <rguenther@suse.de>
1030 * passes.def (all_passes): Start with pass_fixup_cfg again.
1032 2013-11-05 Richard Biener <rguenther@suse.de>
1034 PR tree-optimization/58955
1035 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
1038 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1040 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
1042 (vec_pack_ufix_trunc_v2df): Likewise.
1044 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
1047 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
1048 pattern, instead of word_mode.
1050 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
1052 (set_storage_via_setmem): Likewise.
1054 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
1056 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
1057 gimple.h and the rest of the condition in eliminate_build.
1058 (eliminate_build): Call new routine.
1059 * gimple.h (phi_ssa_name_p): Delete.
1061 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
1063 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
1064 case of no prefix and reserving zero slots, because when that's the
1065 case we'll never get here.
1066 * vec.h (va_heap::reserve): Don't try and handle
1067 vec_prefix::calculate_allocation returning zero because that should
1070 2013-11-05 Richard Biener <rguenther@suse.de>
1073 * tree-dfa.c (get_ref_base_and_extent): Merge common code
1074 in MEM_REF and TARGET_MEM_REF handling. Make sure to
1075 process trailing array detection before diving into the
1076 view-converted object (and possibly apply some extra offset).
1078 2013-11-05 Joseph Myers <joseph@codesourcery.com>
1080 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
1082 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
1084 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
1086 PR tree-optimization/58958
1087 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
1088 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
1090 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
1092 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
1093 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
1095 2013-11-05 Jakub Jelinek <jakub@redhat.com>
1097 PR tree-optimization/58984
1098 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
1099 set *SIZE_P if non-NULL on success.
1100 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
1102 (ipcp_transform_function): Likewise. Punt if size of access
1103 is different from TYPE_SIZE on v->value's type.
1105 2013-11-05 Tobias Burnus <burnus@net-b.de>
1107 * doc/invoke.texi (-fopenmp-simd): Document new option.
1108 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
1109 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
1110 * tree.c (attribute_value_equal): Ditto.
1112 2013-11-04 Wei Mi <wmi@google.com>
1114 * sched-rgn.c (add_branch_dependences): Keep insns in
1115 a SCHED_GROUP at the end of BB to remain their location.
1117 2013-11-04 Wei Mi <wmi@google.com>
1119 * config/i386/i386.c (memory_address_length): Extract a part
1120 of code to rip_relative_addr_p.
1121 (rip_relative_addr_p): New Function.
1122 (ix86_macro_fusion_p): Ditto.
1123 (ix86_macro_fusion_pair_p): Ditto.
1124 * config/i386/i386.h: Add new tune features about macro-fusion.
1125 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
1126 * doc/tm.texi: Generated.
1127 * doc/tm.texi.in: Ditto.
1128 * haifa-sched.c (try_group_insn): New Function.
1129 (group_insns_for_macro_fusion): Ditto.
1130 (sched_init): Call group_insns_for_macro_fusion.
1131 * target.def: Add two hooks: macro_fusion_p and
1132 macro_fusion_pair_p.
1134 2013-11-04 Kostya Serebryany <kcc@google.com>
1136 Update to match the changed asan API.
1137 * asan.c (asan_function_start): New function.
1138 (asan_emit_stack_protection): Update the string stored in the
1139 stack red zone to match new API. Store the PC of the current
1140 function in the red zone.
1141 (asan_global_struct): Update the __asan_global definition to match
1143 (asan_add_global): Ditto.
1144 * asan.h (asan_function_start): New prototype.
1145 * final.c (final_start_function): Call asan_function_start.
1146 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
1149 2013-11-04 Wei Mi <wmi@google.com>
1151 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
1152 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
1153 * config/i386/i386.c (ix86_option_override_internal): Ditto.
1154 (ix86_issue_rate): Ditto.
1155 (ix86_adjust_cost): Ditto.
1156 (ia32_multipass_dfa_lookahead): Ditto.
1157 (ix86_sched_init_global): Ditto.
1158 (get_builtin_code_for_version): Ditto.
1159 * config/i386/i386.h (enum target_cpu_default): Ditto.
1160 (enum processor_type): Ditto.
1161 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
1163 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
1165 PR rtl-optimization/58967
1166 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
1167 !lra_in_progress for mode sizes bigger word.
1169 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1171 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
1172 arguments to merge instruction for little endian.
1173 (vec_widen_umult_lo_v16qi): Likewise.
1174 (vec_widen_smult_hi_v16qi): Likewise.
1175 (vec_widen_smult_lo_v16qi): Likewise.
1176 (vec_widen_umult_hi_v8hi): Likewise.
1177 (vec_widen_umult_lo_v8hi): Likewise.
1178 (vec_widen_smult_hi_v8hi): Likewise.
1179 (vec_widen_smult_lo_v8hi): Likewise.
1181 2013-11-04 Ian Lance Taylor <iant@google.com>
1183 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
1184 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
1185 builtins that take pointers.
1186 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
1187 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
1188 (run_gcc): Pass -fnon-call-exceptions.
1190 2013-11-04 Jakub Jelinek <jakub@redhat.com>
1192 * optabs.c (expand_vec_perm): Revert one incorrect line from
1195 PR tree-optimization/58978
1196 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
1197 use_stmt by single_imm_use directly. Only call single_imm_use
1200 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
1202 PR rtl-optimization/58968
1203 * lra-spills.c (return_regno_p): New function.
1204 (lra_final_code_change): Use it.
1206 2013-11-04 Joseph Myers <joseph@codesourcery.com>
1208 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
1209 * target.def (float_exceptions_rounding_supported_p): New hook.
1210 * targhooks.c (default_float_exceptions_rounding_supported_p): New
1212 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
1213 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
1215 * doc/tm.texi: Regenerate.
1216 * config.gcc (powerpc*-*-linux*): Set extra_objs.
1217 * config/rs6000/rs6000-linux.c: New file.
1218 * config/rs6000/rs6000-protos.h
1219 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
1220 * config/rs6000/linux.h
1221 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
1222 * config/rs6000/linux64.h
1223 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
1224 * config/rs6000/t-linux (rs6000-linux.o): New rule.
1225 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
1227 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1229 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
1230 Replace the define_insn_and_split with a define_insn and two
1231 define_splits, with the split after reload re-permuting the source
1232 register to its original value.
1233 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
1234 (*vsx_le_perm_store_v8hi): Likewise.
1235 (*vsx_le_perm_store_v16qi): Likewise.
1237 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1239 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
1242 2013-11-04 Jakub Jelinek <jakub@redhat.com>
1244 PR tree-optimization/58946
1245 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
1246 bbs with bbinfo[idx].op != NULL before all blocks with
1247 bbinfo[idx].op == NULL.
1249 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
1251 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
1252 (avr_dump_double_int_hex): Likewise.
1253 (avr_log_vadump): Remove %D and %X handling.
1254 * config/avr/avr.c (avr_double_int_push_digit): Delete.
1255 (avr_map_op_t): Change map from double_int to unsigned int.
1256 (avr_map_op): Update accordingly.
1257 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
1258 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
1259 unsigned ints rather than double_ints.
1261 2013-11-03 Marek Polacek <polacek@redhat.com>
1263 Implement -fsanitize=vla-bound.
1264 * opts.c (common_handle_option): Handle vla-bound.
1265 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
1266 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
1267 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
1269 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1271 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
1274 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
1276 * config/i386/constraints.md (Ts, Tv): New address constrains.
1277 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
1278 constraint for address_no_seg_operand.
1279 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
1280 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
1281 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
1282 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
1283 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
1284 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
1285 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
1286 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
1287 constraint for vsib_address_operand.
1289 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
1291 * gcse.c (pre_delete): Remove references to regmove from comments.
1292 * recog.c: (validate_replace_rtx_1): Likewise.
1293 * config/rl78/rl78.c: Likewise.
1294 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
1295 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
1296 * common/config/mmix/mmix-common.c: Likewise.
1298 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
1300 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
1301 * gimplify.c (gimplify_compound_lval): Likewise.
1302 * graphite-clast-to-gimple.c (gloog): Likewise.
1303 * graphite-dependences.c (loop_is_parallel_p): Likewise.
1304 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
1305 (limit_scop); Likewise.
1306 (build_scops): Likewise.
1307 (dot_scop): Likewise.
1308 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
1309 (build_scop_drs): Likewise.
1310 (insert_stmts): Likewise.
1311 (insert_out_of_ssa_copy): Likewise.
1312 (remove_phi): Likewise.
1313 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
1314 * hw-doloop.c (discover_loop): Likewise.
1315 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
1316 * tree-dfa.c (dump_enumerated_decls): Likewise.
1317 * tree-if-conv.c (if_convertable_loop_p): Likewise.
1318 * tree-inline.c (tree_function_versioning): Likewise.
1319 * tree-loop-distribution.c (build_rdg): Likewise.
1320 (rdg_flag_vertex_and_dependent): Likewise.
1321 (distribute_loop): Likewise.
1322 * tree-parloops.c (loop_parallel_p): Likewise.
1323 (eliminate_local_variables): Likewise.
1324 (separate_decls_in_region): Likewise.
1325 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
1326 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
1327 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
1328 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
1329 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
1330 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
1331 (vectorizable_condition): Likewise.
1333 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
1335 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
1336 * configure: Regenerate.
1337 * config/i386/i386.md (*movdi_internal): Change
1338 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
1339 (*movdf_internal): Ditto.
1340 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
1341 * config/i386/sse.md (vec_concatv2di): Output interunit movq
1342 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
1344 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
1346 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
1349 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1351 Automated part of renaming of symtab_node_base to symtab_node.
1353 Patch autogenerated by rename_symtab.py from
1354 https://github.com/davidmalcolm/gcc-refactoring-scripts
1355 revision 58bb219cc090b2f4516a9297d868c245495ee622
1356 with ChangeLog entry fixed up by hand.
1358 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
1360 (cgraph_node_for_asm): Likewise.
1361 * cgraph.h (symtab_node_base): Likewise.
1362 (cgraph_node): Likewise.
1363 (varpool_node): Likewise.
1364 (is_a_helper <cgraph_node>::test): Likewise.
1365 (is_a_helper <varpool_node>::test): Likewise.
1366 (symtab_nodes): Likewise.
1367 (symtab_register_node): Likewise.
1368 (symtab_unregister_node): Likewise.
1369 (symtab_remove_node): Likewise.
1370 (symtab_get_node): Likewise.
1371 (symtab_node_for_asm): Likewise.
1372 (symtab_node_asm_name): Likewise.
1373 (symtab_node_name): Likewise.
1374 (symtab_insert_node_to_hashtable): Likewise.
1375 (symtab_add_to_same_comdat_group): Likewise.
1376 (symtab_dissolve_same_comdat_group_list): Likewise.
1377 (dump_symtab_node): Likewise.
1378 (debug_symtab_node): Likewise.
1379 (dump_symtab_base): Likewise.
1380 (verify_symtab_node): Likewise.
1381 (verify_symtab_base): Likewise.
1382 (symtab_used_from_object_file_p): Likewise.
1383 (symtab_alias_ultimate_target): Likewise.
1384 (symtab_resolve_alias): Likewise.
1385 (fixup_same_cpp_alias_visibility): Likewise.
1386 (symtab_for_node_and_aliases): Likewise.
1387 (symtab_nonoverwritable_alias): Likewise.
1388 (availability symtab_node_availability): Likewise.
1389 (symtab_semantically_equivalent_p): Likewise.
1390 (fixup_same_cpp_alias_visibility): Likewise.
1391 (symtab_prevail_in_asm_name_hash): Likewise.
1393 (varpool): Likewise.
1394 (varpool_first_variable): Likewise.
1395 (varpool_next_variable): Likewise.
1396 (varpool_first_static_initializer): Likewise.
1397 (varpool_next_static_initializer): Likewise.
1398 (varpool_first_defined_variable): Likewise.
1399 (varpool_next_defined_variable): Likewise.
1400 (cgraph_first_defined_function): Likewise.
1401 (cgraph_next_defined_function): Likewise.
1402 (cgraph_first_function): Likewise.
1403 (cgraph_next_function): Likewise.
1404 (cgraph_first_function_with_gimple_body): Likewise.
1405 (cgraph_next_function_with_gimple_body): Likewise.
1406 (symtab_alias_target): Likewise.
1407 (symtab_real_symbol_p): Likewise.
1408 (symtab_can_be_discarded): Likewise.
1409 * cgraphbuild.c (mark_address): Likewise.
1410 (mark_load): Likewise.
1411 (mark_store): Likewise.
1412 * cgraphunit.c (decide_is_symbol_needed): Likewise.
1414 (enqueue_node): Likewise.
1415 (referred_to_p): Likewise.
1416 (cgraph_process_same_body_aliases): Likewise.
1417 (analyze_functions): Likewise.
1418 (handle_alias_pairs): Likewise.
1419 (output_weakrefs): Likewise.
1420 (compile): Likewise.
1421 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1422 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1423 * ipa-prop.c (remove_described_reference): Likewise.
1424 (try_decrement_rdesc_refcount): Likewise.
1425 (ipa_edge_duplication_hook): Likewise.
1426 * ipa-ref.c (ipa_record_reference): Likewise.
1427 (ipa_maybe_record_reference): Likewise.
1428 (ipa_clone_ref): Likewise.
1429 (ipa_clone_references): Likewise.
1430 (ipa_clone_referring): Likewise.
1431 (ipa_find_reference): Likewise.
1432 (ipa_remove_stmt_references): Likewise.
1433 (ipa_clear_stmts_in_references): Likewise.
1434 * ipa-ref.h (symtab_node_base): Likewise.
1435 (ipa_ref): Likewise.
1436 (ipa_record_reference): Likewise.
1437 (ipa_maybe_record_reference): Likewise.
1438 (ipa_clone_references): Likewise.
1439 (ipa_clone_referring): Likewise.
1440 (ipa_clone_ref): Likewise.
1441 (ipa_find_reference): Likewise.
1442 (ipa_remove_stmt_references): Likewise.
1443 (ipa_clear_stmts_in_references): Likewise.
1444 * ipa-reference.c (ipa_reference_write_optimization_summary):
1446 * ipa.c (enqueue_node): Likewise.
1447 (process_references): Likewise.
1448 (walk_polymorphic_call_targets): Likewise.
1449 (symtab_remove_unreachable_nodes): Likewise.
1450 (address_taken_from_non_vtable_p): Likewise.
1451 (comdat_can_be_unshared_p_1): Likewise.
1452 (comdat_can_be_unshared_p): Likewise.
1453 (can_replace_by_local_alias): Likewise.
1454 (function_and_variable_visibility): Likewise.
1455 * is-a.h: Likewise (within example in comment).
1456 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
1457 (lto_symtab_encoder_encode): Likewise.
1458 (lto_symtab_encoder_delete_node): Likewise.
1459 (lto_symtab_encoder_in_partition_p): Likewise.
1460 (lto_set_symtab_encoder_in_partition): Likewise.
1461 (output_refs): Likewise.
1462 (compute_ltrans_boundary): Likewise.
1463 (output_symtab): Likewise.
1464 (input_node): Likewise.
1465 (input_ref): Likewise.
1466 (input_edge): Likewise.
1467 (input_cgraph_1): Likewise.
1468 (input_refs): Likewise.
1469 (output_cgraph_opt_summary): Likewise.
1470 (input_node_opt_summary): Likewise.
1471 (input_cgraph_opt_section): Likewise.
1472 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1474 * lto-streamer-out.c (lto_output): Likewise.
1475 (output_symbol_p): Likewise.
1476 (produce_symtab): Likewise.
1477 * lto-streamer.h (lto_encoder_entry): Likewise.
1478 (lto_free_function_in_decl_state_for_node): Likewise.
1479 (lto_symtab_encoder_encode): Likewise.
1480 (lto_symtab_encoder_delete_node): Likewise.
1481 (lto_symtab_encoder_in_partition_p): Likewise.
1482 (lto_set_symtab_encoder_in_partition): Likewise.
1483 (lto_symtab_encoder_lookup): Likewise.
1484 (lsei_node): Likewise.
1485 (lto_symtab_encoder_deref): Likewise.
1486 * symtab.c (symtab_hash): Likewise.
1487 (assembler_name_hash): Likewise.
1488 (symtab_nodes): Likewise.
1489 (hash_node): Likewise.
1490 (eq_node): Likewise.
1491 (hash_node_by_assembler_name): Likewise.
1492 (eq_assembler_name): Likewise.
1493 (insert_to_assembler_name_hash): Likewise.
1494 (unlink_from_assembler_name_hash): Likewise.
1495 (symtab_prevail_in_asm_name_hash): Likewise.
1496 (symtab_register_node): Likewise.
1497 (symtab_insert_node_to_hashtable): Likewise.
1498 (symtab_unregister_node): Likewise.
1499 (symtab_get_node): Likewise.
1500 (symtab_remove_node): Likewise.
1501 (symtab_initialize_asm_name_hash): Likewise.
1502 (symtab_node_for_asm): Likewise.
1503 (symtab_add_to_same_comdat_group): Likewise.
1504 (symtab_dissolve_same_comdat_group_list): Likewise.
1505 (symtab_node_asm_name): Likewise.
1506 (symtab_node_name): Likewise.
1507 (dump_symtab_base): Likewise.
1508 (dump_symtab_node): Likewise.
1509 (dump_symtab): Likewise.
1510 (debug_symtab_node): Likewise.
1511 (verify_symtab_base): Likewise.
1512 (verify_symtab_node): Likewise.
1513 (verify_symtab): Likewise.
1514 (symtab_used_from_object_file_p): Likewise.
1515 (symtab_node_availability): Likewise.
1516 (symtab_alias_ultimate_target): Likewise.
1517 (fixup_same_cpp_alias_visibility): Likewise.
1518 (symtab_resolve_alias): Likewise.
1519 (symtab_for_node_and_aliases): Likewise.
1520 (symtab_for_node_and_aliases): Likewise.
1521 (symtab_nonoverwritable_alias_1): Likewise.
1522 (symtab_nonoverwritable_alias): Likewise.
1523 (symtab_semantically_equivalent_p): Likewise.
1524 * value-prof.c (init_node_map): Likewise.
1525 * varasm.c (find_decl): Likewise.
1526 * varpool.c (varpool_node_for_asm): Likewise.
1527 (varpool_remove_unreferenced_decls): Likewise.
1529 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1531 Manual part of renaming of symtab_node_base to symtab_node.
1533 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
1534 clashes with the preferred name for the base class.
1535 (const_symtab_node): Remove redundant typedef.
1537 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1539 * optabs.c (expand_vec_perm): Avoid vector mode punning
1540 SUBREGs in SET_DEST.
1541 * expmed.c (store_bit_field_1): Likewise.
1542 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
1543 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
1544 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
1545 * config/i386/i386.c (ix86_expand_vector_move_misalign,
1546 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
1547 ix86_expand_sse_unpack, ix86_expand_args_builtin,
1548 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
1549 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
1550 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
1551 expand_vec_perm_vpshufb2_vpermq,
1552 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
1553 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
1554 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
1555 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
1558 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
1560 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
1561 Use gcc_unreachable for unhandled modes. Do not check results of
1562 expand_simple_binop. If not expanded to target, move the result.
1564 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
1565 Shiva Chen <shiva0217@gmail.com>
1567 * config.gcc (nds32*-*-*): Add nds32 target.
1568 * config/nds32/nds32.c: New file.
1569 * config/nds32/nds32.h: New file.
1570 * config/nds32/nds32.md: New file.
1571 * config/nds32/constants.md: New file.
1572 * config/nds32/constraints.md: New file.
1573 * config/nds32/iterators.md: New file.
1574 * config/nds32/nds32-doubleword.md: New file.
1575 * config/nds32/nds32-intrinsic.md: New file.
1576 * config/nds32/nds32_intrinsic.h: New file.
1577 * config/nds32/nds32-modes.def: New file.
1578 * config/nds32/nds32-multiple.md: New file.
1579 * config/nds32/nds32.opt: New file.
1580 * config/nds32/nds32-opts.h: New file.
1581 * config/nds32/nds32-protos.h: New file.
1582 * config/nds32/nds32-peephole2.md: New file.
1583 * config/nds32/pipelines.md: New file.
1584 * config/nds32/predicates.md: New file.
1585 * config/nds32/t-mlibs: New file.
1586 * common/config/nds32: New directory and files.
1588 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
1589 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
1590 * doc/install.texi (Cross-Compiler-Specific Options): Document
1591 --with-nds32-lib for nds32 target.
1592 * doc/extend.texi (Function Attributes, Target Builtins): Document
1593 nds32 specific attributes.
1595 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
1597 * lra-constraints (process_alt_operands): Use the result
1598 elimination register for operand when matching constraints.
1600 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1602 * tree-vrp.c (maybe_set_nonzero_bits): New function.
1603 (remove_range_assertions): Call it.
1605 * tree.c (tree_ctz): New function.
1606 * tree.h (tree_ctz): New prototype.
1607 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
1608 first argument from tree to const_tree.
1609 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
1610 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
1611 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
1613 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
1614 * tree-vect-loop.c (vect_analyze_loop_operations,
1615 vect_transform_loop): Don't force scalar loop for bound just because
1616 number of iterations is unknown, only do it if it is not known to be
1617 a multiple of vectorization_factor.
1618 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
1620 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
1621 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
1622 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
1623 (set_nonzero_bits, get_nonzero_bits): New prototypes.
1624 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
1625 a default def isn't partially constant.
1626 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
1627 is known to be partially zero.
1628 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
1629 to see if a default def isn't partially constant.
1630 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
1631 creation of a range, if VR_RANGE, try to improve nonzero_bits from
1633 (set_nonzero_bits, get_nonzero_bits): New functions.
1635 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
1636 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
1637 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
1638 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
1639 uses but in the condition and ASSERT_EXPR and the other successor of
1640 the predecessor bb is __builtin_unreachable (), set_range_info of the
1641 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
1643 2013-10-31 Martin Jambor <mjambor@suse.cz>
1645 PR rtl-optimization/58934
1647 2013-10-30 Martin Jambor <mjambor@suse.cz>
1648 PR rtl-optimization/10474
1649 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1651 (interesting_dest_for_shprep): New function.
1652 (split_live_ranges_for_shrink_wrap): Likewise.
1653 (ira): Calculate dominance info and df analysis. Call
1654 split_live_ranges_for_shrink_wrap.
1656 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1657 Yury Gribov <y.gribov@samsung.com>
1660 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
1661 shadow pointer to avoid clobbering the main one.
1663 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1665 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
1667 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1670 * ira-color.c (update_costs_from_copies): Add new parameter. Use
1671 it for calling update_costs_from_allocno.
1672 (assign_hard_reg): Call restore_costs_from_copies only for
1673 !retry_p. Pass new argument to update_costs_from_copies.
1674 (color_pass): Pass new argument to update_costs_from_copies.
1675 (ira_mark_allocation_change): Ditto.
1677 2013-10-30 Sharad Singhai <singhai@google.com>
1680 * opts.c (common_handle_option): Remove deprecated option
1681 -ftree-vectorizer-verbose.
1682 * doc/invoke.texi (Debugging Options): Ditto.
1683 * opts-global.c (handle_common_deferred_options): Ditto.
1684 (dump_remap_tree_vectorizer_verbose): Delete.
1685 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
1687 2013-10-30 DJ Delorie <dj@redhat.com>
1689 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
1690 that take no arguments.
1692 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
1695 * reload1.c (update_eliminables_and_spill): New function, broken
1697 (reload): Use it. Check for frame size change after frame size
1698 alignment, and call update_eliminables_and_spill first if continue-ing.
1700 2013-10-30 Cong Hou <congh@google.com>
1703 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
1704 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
1705 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
1707 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
1709 PR rtl-optimization/58369
1710 * reload1.c (compute_reload_subreg_offset): New function.
1711 (choose_reload_regs): Use it to pass endian-correct
1712 offset to subreg_regno_offset.
1714 2013-10-30 Tobias Burnus <burnus@net-b.de>
1717 * tree-cfg.c (replace_loop_annotate): Replace warning by
1720 2013-10-30 Jason Merrill <jason@redhat.com>
1722 * configure.ac (loose_warn): Add -Wno-format if
1723 --disable-build-format-warnings.
1725 2013-10-30 David Malcolm <dmalcolm@redhat.com>
1727 * cgraphunit.c (analyze_functions): Split symtab_node declarations
1728 onto multiple lines to make things easier for rename_symtab.py.
1730 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
1731 (symtab_semantically_equivalent_p): Likewise.
1733 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1736 * lra.c (check_rtl): Remove address check before LRA work.
1738 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
1740 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
1741 POINTER_PLUS_EXPR in the defining statement.
1743 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1745 * regmove.c: Remove.
1746 * tree-pass.h (make_pass_regmove): Remove.
1747 * timevar.def (TV_REGMOVE): Remove.
1748 * passes.def (pass_regmove): Remove.
1749 * opts.c (default_options_table): Remove entry for regmove.
1750 * doc/passes.texi: Remove regmove pass description.
1751 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
1753 (-fdump-rtl-regmove): Ditto.
1754 * common.opt (foptimize-register-move, fregmove): Ignore.
1755 * Makefile.in (OBJS): Remove regmove.o.
1756 * regmove.c: Remove.
1757 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
1759 (struct ira_allocno) New member allocno_prefs.
1760 (ALLOCNO_PREFS): New macro.
1761 (ira_prefs, ira_prefs_num): New external vars.
1762 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
1763 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
1764 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
1766 (ira_add_allocno_copy_to_list): Remove prototype.
1767 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
1768 (ira_pref_iterator): New type.
1769 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
1770 (FOR_EACH_PREF): New macro.
1771 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
1772 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
1774 (ira_setup_alts): New function.
1775 (decrease_live_ranges_number): New function.
1776 (ira): Call the above function.
1777 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
1778 (ira_create_allocno): Initialize allocno prefs.
1779 (pref_pool, pref_vec): New static vars.
1780 (initiate_prefs, find_allocno_pref, ira_create_pref): New
1782 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
1783 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
1784 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
1785 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
1786 (ira_add_allocno_copy_to_list): Make static. Rename to
1787 add_allocno_copy_to_list.
1788 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
1789 swap_allocno_copy_ends_if_necessary.
1790 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
1791 ira_remove_allocno_prefs.
1792 (ira_flattening): Ditto.
1793 (ira_build): Call initiate_prefs, print_prefs.
1794 (ira_destroy): Call finish_prefs.
1795 * ira-color.c (struct update_cost_record): New.
1796 (struct allocno_color_data): Add new member update_cost_records.
1797 (update_cost_record_pool): New static var.
1798 (init_update_cost_records, get_update_cost_record): New functions.
1799 (free_update_cost_record_list, finish_update_cost_records): Ditto.
1800 (struct update_cost_queue_elem): Add member from.
1801 (initiate_cost_update): Call init_update_cost_records.
1802 (finish_cost_update): Call finish_update_cost_records.
1803 (queue_update_cost, get_next_update_cost): Add new param from.
1804 (Update_allocno_cost, update_costs_from_allocno): New functions.
1805 (update_costs_from_prefs): Ditto.
1806 (update_copy_costs): Rename to update_costs_from_copies.
1807 (restore_costs_from_copies): New function.
1808 (update_conflict_hard_regno_costs): Don't go back.
1809 (assign_hard_reg): Call restore_costs_from_copies. Add printing
1811 (pop_allocnos): Add priniting more debug info.
1812 (color_allocnos): Remove prefs for conflicting hard regs.
1813 Call update_costs_from_prefs.
1814 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
1815 (get_dup_num): Rename, modify, and move to ira.c
1816 (process_regs_for_copy): Add prefs.
1817 (add_insn_allocno_copies): Put src as first arg of
1818 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
1819 * ira-costs.c (record_reg_classes): Modify and move code into
1820 record_operands_costs.
1821 (find_costs_and_classes): Create prefs for the hard reg of small
1823 (process_bb_node_for_hard_reg_moves): Add prefs.
1825 2013-10-30 Richard Biener <rguenther@suse.de>
1828 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
1829 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
1831 (pre_and_rev_post_order_compute): ... which now wraps it.
1832 * graph.c (draw_cfg_nodes_no_loops): Use
1833 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
1836 2013-10-30 Christian Bruel <christian.bruel@st.com>
1838 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
1839 (sh_expand_cmpstr): Handle known align and schedule improvements.
1840 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
1841 * config/sh/sh.md (cmpstrnsi): New pattern.
1843 2013-10-30 Martin Jambor <mjambor@suse.cz>
1845 PR rtl-optimization/10474
1846 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1848 (interesting_dest_for_shprep): New function.
1849 (split_live_ranges_for_shrink_wrap): Likewise.
1850 (ira): Calculate dominance info and df analysis. Call
1851 split_live_ranges_for_shrink_wrap.
1853 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1856 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1858 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1860 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1861 * tree.h (POINTER_BOUNDS_P): New.
1862 (BOUNDED_TYPE_P): New.
1864 (pointer_bounds_type_node): New.
1865 * tree.c (build_common_tree_nodes): Initialize
1866 pointer_bounds_type_node.
1867 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1868 (gimple_call_num_nobnd_args): New.
1869 (gimple_call_nobnd_arg): New.
1870 (gimple_return_retbnd): New.
1871 (gimple_return_set_retbnd): New
1872 * gimple.c (gimple_build_return): Increase number of ops
1873 for return statement.
1874 (gimple_call_get_nobnd_arg_index): New.
1875 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1877 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1879 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1880 with "chkp ctor" and "bnd_legacy" attributes.
1881 * gimplify.c (gimplify_init_constructor): Avoid infinite
1882 loop during gimplification of bounds initializer.
1884 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1886 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1887 (handle_bnd_legacy): New.
1888 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1889 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1892 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1894 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1895 (BT_FN_PTR_CONST_PTR): New.
1896 (BT_FN_CONST_PTR_CONST_PTR): New.
1897 (BT_FN_PTR_CONST_PTR_SIZE): New.
1898 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1899 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1900 (BT_FN_VOID_CONST_PTR_SIZE): New.
1901 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1902 * chkp-builtins.def: New.
1903 * builtins.def: include chkp-builtins.def.
1904 (DEF_CHKP_BUILTIN): New.
1905 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1906 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1907 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1908 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1909 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1910 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1911 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1912 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1913 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1914 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1915 * common.opt (fcheck-pointer-bounds): New.
1916 * toplev.c (process_options): Check Pointer Bounds Checker is
1918 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1920 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1922 * target.def (builtin_chkp_function): New.
1923 (chkp_bound_type): New.
1924 (chkp_bound_mode): New.
1925 (fn_abi_va_list_bounds_size): New.
1926 (load_bounds_for_arg): New.
1927 (store_bounds_for_arg): New.
1928 * targhooks.h (default_load_bounds_for_arg): New.
1929 (default_store_bounds_for_arg): New.
1930 (default_fn_abi_va_list_bounds_size): New.
1931 (default_chkp_bound_type): New.
1932 (default_chkp_bound_mode): New.
1933 (default_builtin_chkp_function): New.
1934 * targhooks.c (default_load_bounds_for_arg): New.
1935 (default_store_bounds_for_arg): New.
1936 (default_fn_abi_va_list_bounds_size): New.
1937 (default_chkp_bound_type): New.
1938 (default_chkp_bound_mode); New.
1939 (default_builtin_chkp_function): New.
1940 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1941 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1942 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1943 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1944 (TARGET_CHKP_BOUND_TYPE): New.
1945 (TARGET_CHKP_BOUND_MODE): New.
1946 * doc/tm.texi: Regenerated.
1947 * langhooks.h (lang_hooks): Add chkp_supported field.
1948 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1949 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1951 2013-10-29 Andrew Pinski <apinski@cavium.com>
1953 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1954 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1955 fails, combining the branches anyways.
1956 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1957 increases the number of combinings.
1958 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1960 2013-10-29 Mike Stump <mikestump@comcast.net>
1962 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1963 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1964 (make_vector_mode): Increase namebuf to 16.
1965 (emit_insn_modes_h): When processing BImode, don't
1966 also match partial int modes.
1967 (emit_class_narrowest_mode): Likewise.
1969 * config/bfin/bfin-modes.def: Add precision to PDI.
1970 * config/m32c/m32c-modes.def: Add precision to PSI.
1971 * config/msp430/msp430-modes.def: Add precision to PSI.
1972 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1973 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1975 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1978 * config/sh/sh.md (*addc): Rename existing variations to ...
1979 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1980 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1981 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1982 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1984 2013-10-29 Teresa Johnson <tejohnson@google.com>
1987 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1988 insanities when updating probabilities.
1990 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1992 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1993 prettyprinter for cgraph_node to reflect the conversion of the
1994 symtable types to a C++ class hierarchy: it now *is* a
1995 symtab_node_base, rather than having one (named "symbol").
1997 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1999 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
2000 __cilkrts_pop_frame. If matched, then return true for built-in
2002 (expand_builtin): Added BUILT_IN_CILK_DETACH and
2003 BUILT_IN_CILK_POP_FRAME case.
2004 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
2005 (lhs_cilk_detect_spawn): Likewise.
2006 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
2007 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
2008 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
2009 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
2010 (LANG_HOOKS_CILKPLUS): Likewise.
2011 * tree.h (CILK_SPAWN_FN): Likewise.
2012 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
2013 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
2014 (OBJS): Added cilk-common.o.
2015 (BUILTINS_DEF): Added cilk-builtins.def.
2016 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
2017 (lhd_cilk_detect_spawn): Likewise.
2018 * langhooks.h (lang_hooks_for_cilkplus): New struct.
2019 (struct lang_hooks): Added new field called "cilkplus."
2020 * cilk-common.c: New file.
2022 * cilk-builtins.def: Likewise.
2023 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
2024 "__cilk" macro and set it to 200.
2025 * function.h (struct function::cilk_frame_decl): New field.
2026 (struct function::is_cilk_function): Likewise.
2027 (struct function::calls_cilk_spawn): Likewise.
2028 * gimplify.c (gimplify_call_expr): Added a check if the function call
2029 being gimplified is a spawn detach point. If so, then add pop_frame
2030 and detach function calls.
2031 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
2032 for gimplifying _Cilk_spawn and _Cilk_sync statements.
2033 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
2034 function. If so, added a _Cilk_sync and gimplified it.
2035 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
2036 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
2037 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
2039 (can_inline_edge_p): Prevent inling of spawnee function.
2040 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
2041 for functions that use Cilk keywords.
2042 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
2043 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
2044 CILK_SYNC_STMT cases.
2045 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
2047 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
2048 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
2049 * passes.texi (Cilk Keywords): New section that describes the compiler
2050 code changes for handling Cilk Keywords.
2052 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2054 Patch autogenerated by refactor_symtab.py from
2055 https://github.com/davidmalcolm/gcc-refactoring-scripts
2056 revision 58bb219cc090b2f4516a9297d868c245495ee622
2058 * asan.c (asan_finish_file): Update for conversion of symtab types to
2059 a true class hierarchy.
2060 * cfgexpand.c (estimated_stack_frame_size): Likewise.
2061 * cgraph.c (cgraph_get_body): Likewise.
2062 (cgraph_get_create_real_symbol_node): Likewise.
2063 (verify_cgraph_node): Likewise.
2064 (verify_edge_corresponds_to_fndecl): Likewise.
2065 (verify_edge_count_and_frequency): Likewise.
2066 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
2067 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
2068 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
2069 (cgraph_node_cannot_return): Likewise.
2070 (cgraph_set_pure_flag_1): Likewise.
2071 (cgraph_set_const_flag_1): Likewise.
2072 (cgraph_set_nothrow_flag_1): Likewise.
2073 (cgraph_make_node_local_1): Likewise.
2074 (cgraph_for_node_and_aliases): Likewise.
2075 (cgraph_for_node_thunks_and_aliases): Likewise.
2076 (cgraph_node_can_be_local_p): Likewise.
2077 (cgraph_node_cannot_be_local_p_1): Likewise.
2078 (cgraph_function_body_availability): Likewise.
2079 (dump_cgraph_node): Likewise.
2080 (cgraph_rtl_info): Likewise.
2081 (cgraph_mark_address_taken_node): Likewise.
2082 (cgraph_remove_node): Likewise.
2083 (cgraph_release_function_body): Likewise.
2084 (cgraph_update_edges_for_call_stmt_node): Likewise.
2085 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
2086 (cgraph_make_edge_direct): Likewise.
2087 (cgraph_resolve_speculation): Likewise.
2088 (cgraph_speculative_call_info): Likewise.
2089 (cgraph_turn_edge_to_speculative): Likewise.
2090 (cgraph_create_edge_1): Likewise.
2091 (cgraph_set_call_stmt): Likewise.
2092 (cgraph_node_for_asm): Likewise.
2093 (cgraph_add_thunk): Likewise.
2094 (cgraph_same_body_alias): Likewise.
2095 (cgraph_create_function_alias): Likewise.
2096 (cgraph_create_node): Likewise.
2097 (cgraph_create_empty_node): Likewise.
2098 (record_function_versions): Likewise.
2099 (used_from_object_file_p): Likewise.
2100 * cgraph.h (symtab_can_be_discarded): Likewise.
2101 (symtab_real_symbol_p): Likewise.
2102 (cgraph_mark_force_output_node): Likewise.
2103 (cgraph_edge_recursive_p): Likewise.
2104 (symtab_alias_target): Likewise.
2105 (varpool_all_refs_explicit_p): Likewise.
2106 (varpool_can_remove_if_no_refs): Likewise.
2107 (cgraph_only_called_directly_or_aliased_p): Likewise.
2108 (cgraph_next_function_with_gimple_body): Likewise.
2109 (cgraph_first_function_with_gimple_body): Likewise.
2110 (cgraph_function_with_gimple_body_p): Likewise.
2111 (cgraph_next_function): Likewise.
2112 (cgraph_first_function): Likewise.
2113 (cgraph_next_defined_function): Likewise.
2114 (cgraph_first_defined_function): Likewise.
2115 (varpool_next_defined_variable): Likewise.
2116 (varpool_first_defined_variable): Likewise.
2117 (varpool_next_static_initializer): Likewise.
2118 (varpool_first_static_initializer): Likewise.
2119 (varpool_next_variable): Likewise.
2120 (varpool_first_variable): Likewise.
2121 (varpool_node_name): Likewise.
2122 (varpool): Likewise.
2124 (is_a_helper <varpool_node>::test): Likewise.
2125 (is_a_helper <cgraph_node>::test): Likewise.
2126 (varpool_variable_node): Likewise.
2127 (cgraph_function_or_thunk_node): Likewise.
2128 (varpool_alias_target): Likewise.
2129 (cgraph_alias_target): Likewise.
2130 (cgraph_node_name): Likewise.
2131 (varpool_node_asm_name): Likewise.
2132 (cgraph_node_asm_name): Likewise.
2133 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
2134 (cgraph_rebuild_references): Likewise.
2135 (rebuild_cgraph_edges): Likewise.
2136 (record_eh_tables): Likewise.
2137 (build_cgraph_edges): Likewise.
2138 (mark_store): Likewise.
2139 (mark_load): Likewise.
2140 (mark_address): Likewise.
2141 (record_type_list): Likewise.
2142 (record_reference): Likewise.
2143 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
2144 (cgraph_materialize_clone): Likewise.
2145 (cgraph_function_versioning): Likewise.
2146 (cgraph_copy_node_for_versioning): Likewise.
2147 (update_call_expr): Likewise.
2148 (cgraph_find_replacement_node): Likewise.
2149 (cgraph_create_virtual_clone): Likewise.
2150 (cgraph_clone_node): Likewise.
2151 * cgraphunit.c (compile): Likewise.
2152 (output_weakrefs): Likewise.
2153 (output_in_order): Likewise.
2154 (expand_function): Likewise.
2155 (assemble_thunks_and_aliases): Likewise.
2156 (expand_thunk): Likewise.
2157 (mark_functions_to_output): Likewise.
2158 (handle_alias_pairs): Likewise.
2159 (analyze_functions): Likewise.
2160 (walk_polymorphic_call_targets): Likewise.
2161 (varpool_finalize_decl): Likewise.
2162 (process_function_and_variable_attributes): Likewise.
2163 (cgraph_process_same_body_aliases): Likewise.
2164 (analyze_function): Likewise.
2165 (cgraph_add_new_function): Likewise.
2166 (cgraph_finalize_function): Likewise.
2167 (referred_to_p): Likewise.
2168 (cgraph_reset_node): Likewise.
2169 (cgraph_process_new_functions): Likewise.
2170 (enqueue_node): Likewise.
2171 (decide_is_symbol_needed): Likewise.
2172 * coverage.c (coverage_compute_profile_id): Likewise.
2173 * dbxout.c (dbxout_expand_expr): Likewise.
2174 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
2175 (reference_to_unused): Likewise.
2176 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2177 * gimplify.c (unvisit_body): Likewise.
2178 (unshare_body): Likewise.
2179 * ipa-cp.c (ipcp_generate_summary): Likewise.
2180 (ipcp_decision_stage): Likewise.
2181 (identify_dead_nodes): Likewise.
2182 (decide_whether_version_node): Likewise.
2183 (decide_about_value): Likewise.
2184 (perhaps_add_new_callers): Likewise.
2185 (create_specialized_node): Likewise.
2186 (update_profiling_info): Likewise.
2187 (ipcp_propagate_stage): Likewise.
2188 (estimate_local_effects): Likewise.
2189 (good_cloning_opportunity_p): Likewise.
2190 (devirtualization_time_bonus): Likewise.
2191 (propagate_constants_accross_call): Likewise.
2192 (initialize_node_lattices): Likewise.
2193 (ipcp_cloning_candidate_p): Likewise.
2194 (determine_versionability): Likewise.
2195 (print_all_lattices): Likewise.
2196 (print_lattice): Likewise.
2197 (ipcp_discover_new_direct_edges): Likewise.
2198 * ipa-devirt.c (ipa_devirt): Likewise.
2199 (likely_target_p): Likewise.
2200 (update_type_inheritance_graph): Likewise.
2201 (possible_polymorphic_call_target_p): Likewise.
2202 (dump_possible_polymorphic_call_targets): Likewise.
2203 (devirt_variable_node_removal_hook): Likewise.
2204 (record_binfo): Likewise.
2205 (maybe_record_node): Likewise.
2206 (build_type_inheritance_graph): Likewise.
2207 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2208 (inline_generate_summary): Likewise.
2209 (inline_analyze_function): Likewise.
2210 (do_estimate_growth): Likewise.
2211 (simple_edge_hints): Likewise.
2212 (estimate_node_size_and_time): Likewise.
2213 (estimate_edge_devirt_benefit): Likewise.
2214 (compute_inline_parameters): Likewise.
2215 (estimate_function_body_sizes): Likewise.
2216 (compute_bb_predicates): Likewise.
2217 (initialize_inline_failed): Likewise.
2218 (dump_inline_summary): Likewise.
2219 (dump_inline_edge_summary): Likewise.
2220 * ipa-inline-transform.c (inline_transform): Likewise.
2221 (preserve_function_body_p): Likewise.
2222 (save_inline_function_body): Likewise.
2223 (inline_call): Likewise.
2224 (clone_inlined_nodes): Likewise.
2225 (can_remove_node_now_p): Likewise.
2226 (can_remove_node_now_p_1): Likewise.
2227 * ipa-inline.c (early_inliner): Likewise.
2228 (early_inline_small_functions): Likewise.
2229 (inline_always_inline_functions): Likewise.
2230 (ipa_inline): Likewise.
2231 (flatten_function): Likewise.
2232 (inline_small_functions): Likewise.
2233 (speculation_useful_p): Likewise.
2234 (recursive_inlining): Likewise.
2235 (update_caller_keys): Likewise.
2236 (reset_edge_caches): Likewise.
2237 (update_edge_key): Likewise.
2238 (edge_badness): Likewise.
2239 (relative_time_benefit): Likewise.
2240 (want_inline_self_recursive_call_p): Likewise.
2241 (want_inline_small_function_p): Likewise.
2242 (want_early_inline_function_p): Likewise.
2243 (num_calls): Likewise.
2244 (can_early_inline_edge_p): Likewise.
2245 (can_inline_edge_p): Likewise.
2246 (report_inline_failed_reason): Likewise.
2247 * ipa-profile.c (ipa_profile): Likewise.
2248 (ipa_propagate_frequency): Likewise.
2249 (ipa_propagate_frequency_1): Likewise.
2250 (ipa_profile_generate_summary): Likewise.
2251 * ipa-prop.c (ipcp_transform_function): Likewise.
2252 (read_replacements_section): Likewise.
2253 (ipa_prop_read_section): Likewise.
2254 (ipa_modify_call_arguments): Likewise.
2255 (ipa_print_node_params): Likewise.
2256 (propagate_controlled_uses): Likewise.
2257 (update_indirect_edges_after_inlining): Likewise.
2258 (remove_described_reference): Likewise.
2259 (ipa_make_edge_direct_to_target): Likewise.
2260 (ipa_analyze_node): Likewise.
2261 (ipa_analyze_params_uses): Likewise.
2262 (ipa_compute_jump_functions): Likewise.
2263 (ipa_get_callee_param_type): Likewise.
2264 (ipa_print_node_jump_functions): Likewise.
2265 (ipa_initialize_node_params): Likewise.
2266 (ipa_populate_param_decls): Likewise.
2267 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
2268 (write_agg_replacement_chain): Likewise.
2269 (ipa_write_node_info): Likewise.
2270 (ipa_edge_duplication_hook): Likewise.
2271 (try_decrement_rdesc_refcount): Likewise.
2272 * ipa-pure-const.c (propagate_nothrow): Likewise.
2273 (propagate_pure_const): Likewise.
2274 (pure_const_read_summary): Likewise.
2275 (pure_const_write_summary): Likewise.
2276 (analyze_function): Likewise.
2277 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
2278 (ipa_ref_referring_ref_list): Likewise.
2279 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
2280 (ipa_remove_stmt_references): Likewise.
2281 (ipa_find_reference): Likewise.
2282 (ipa_dump_referring): Likewise.
2283 (ipa_dump_references): Likewise.
2284 (ipa_record_reference): Likewise.
2285 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
2286 (ipa_reference_write_optimization_summary): Likewise.
2287 (write_node_summary_p): Likewise.
2288 (propagate): Likewise.
2289 (read_write_all_from_decl): Likewise.
2290 (generate_summary): Likewise.
2291 (analyze_function): Likewise.
2292 (propagate_bits): Likewise.
2293 (ipa_reference_get_not_written_global): Likewise.
2294 (ipa_reference_get_not_read_global): Likewise.
2295 * ipa-split.c (execute_split_functions): Likewise.
2296 (split_function): Likewise.
2297 * ipa-utils.c (ipa_merge_profiles): Likewise.
2298 (dump_cgraph_node_set): Likewise.
2299 (ipa_reverse_postorder): Likewise.
2300 (ipa_edge_within_scc): Likewise.
2301 (ipa_get_nodes_in_cycle): Likewise.
2302 (ipa_free_postorder_info): Likewise.
2303 (ipa_reduced_postorder): Likewise.
2304 (searchc): Likewise.
2305 (recursive_call_p): Likewise.
2306 * ipa.c (ipa_cdtor_merge): Likewise.
2307 (record_cdtor_fn): Likewise.
2308 (function_and_variable_visibility): Likewise.
2309 (varpool_externally_visible_p): Likewise.
2310 (cgraph_externally_visible_p): Likewise.
2311 (comdat_can_be_unshared_p): Likewise.
2312 (comdat_can_be_unshared_p_1): Likewise.
2313 (address_taken_from_non_vtable_p): Likewise.
2314 (ipa_discover_readonly_nonaddressable_vars): Likewise.
2315 (symtab_remove_unreachable_nodes): Likewise.
2316 (walk_polymorphic_call_targets): Likewise.
2317 (process_references): Likewise.
2318 (enqueue_node): Likewise.
2319 (has_addr_references_p): Likewise.
2320 (cgraph_non_local_node_p_1): Likewise.
2321 * is-a.h (varpool_analyze_node): Likewise.
2322 * lto-cgraph.c (input_symtab): Likewise.
2323 (merge_profile_summaries): Likewise.
2324 (input_cgraph_1): Likewise.
2325 (input_edge): Likewise.
2326 (input_varpool_node): Likewise.
2327 (input_node): Likewise.
2328 (input_overwrite_node): Likewise.
2329 (compute_ltrans_boundary): Likewise.
2330 (output_refs): Likewise.
2331 (lto_output_varpool_node): Likewise.
2332 (lto_output_node): Likewise.
2333 (reachable_from_other_partition_p): Likewise.
2334 (referenced_from_other_partition_p): Likewise.
2335 (lto_output_edge): Likewise.
2336 (output_node_opt_summary): Likewise.
2337 (add_node_to): Likewise.
2338 (reachable_from_this_partition_p): Likewise.
2339 (lto_set_symtab_encoder_in_partition): Likewise.
2340 (lto_symtab_encoder_in_partition_p): Likewise.
2341 (lto_set_symtab_encoder_encode_initializer): Likewise.
2342 (lto_symtab_encoder_encode_initializer_p): Likewise.
2343 (lto_set_symtab_encoder_encode_body): Likewise.
2344 (lto_symtab_encoder_encode_body_p): Likewise.
2345 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
2347 * lto-streamer-in.c (lto_read_body): Likewise.
2348 (fixup_call_stmt_edges): Likewise.
2349 (fixup_call_stmt_edges_1): Likewise.
2350 * lto-streamer-out.c (produce_symtab): Likewise.
2351 (output_symbol_p): Likewise.
2352 (write_symbol): Likewise.
2353 (lto_output): Likewise.
2354 (copy_function): Likewise.
2355 (output_function): Likewise.
2356 * passes.c (function_called_by_processed_nodes_p): Likewise.
2357 (ipa_write_optimization_summaries): Likewise.
2358 (ipa_write_summaries): Likewise.
2359 (do_per_function_toporder): Likewise.
2360 (do_per_function): Likewise.
2361 (dump_passes): Likewise.
2362 * symtab.c (symtab_semantically_equivalent_p): Likewise.
2363 (symtab_nonoverwritable_alias): Likewise.
2364 (symtab_nonoverwritable_alias_1): Likewise.
2365 (symtab_for_node_and_aliases): Likewise.
2366 (symtab_resolve_alias): Likewise.
2367 (fixup_same_cpp_alias_visibility): Likewise.
2368 (symtab_alias_ultimate_target): Likewise.
2369 (symtab_used_from_object_file_p): Likewise.
2370 (verify_symtab_base): Likewise.
2371 (dump_symtab_base): Likewise.
2372 (symtab_node_name): Likewise.
2373 (symtab_node_asm_name): Likewise.
2374 (symtab_dissolve_same_comdat_group_list): Likewise.
2375 (symtab_add_to_same_comdat_group): Likewise.
2376 (symtab_unregister_node): Likewise.
2377 (symtab_insert_node_to_hashtable): Likewise.
2378 (symtab_register_node): Likewise.
2379 (unlink_from_assembler_name_hash): Likewise.
2380 (insert_to_assembler_name_hash): Likewise.
2381 (eq_assembler_name): Likewise.
2382 (hash_node_by_assembler_name): Likewise.
2383 (eq_node): Likewise.
2384 (hash_node): Likewise.
2385 * toplev.c (wrapup_global_declaration_2): Likewise.
2386 * trans-mem.c (ipa_tm_execute): Likewise.
2387 (ipa_tm_transform_clone): Likewise.
2388 (ipa_tm_transform_transaction): Likewise.
2389 (ipa_tm_transform_calls_redirect): Likewise.
2390 (ipa_tm_insert_gettmclone_call): Likewise.
2391 (ipa_tm_insert_irr_call): Likewise.
2392 (ipa_tm_create_version): Likewise.
2393 (ipa_tm_create_version_alias): Likewise.
2394 (ipa_tm_mark_forced_by_abi_node): Likewise.
2395 (ipa_tm_mark_force_output_node): Likewise.
2396 (ipa_tm_diagnose_tm_safe): Likewise.
2397 (ipa_tm_mayenterirr_function): Likewise.
2398 (ipa_tm_scan_irr_function): Likewise.
2399 (ipa_tm_note_irrevocable): Likewise.
2400 (ipa_tm_scan_calls_clone): Likewise.
2401 (get_cg_data): Likewise.
2402 * tree-eh.c (tree_could_trap_p): Likewise.
2403 * tree-emutls.c (ipa_lower_emutls): Likewise.
2404 (create_emultls_var): Likewise.
2405 (lower_emutls_function_body): Likewise.
2406 (gen_emutls_addr): Likewise.
2407 (emutls_decl): Likewise.
2408 (new_emutls_decl): Likewise.
2409 * tree-inline.c (tree_function_versioning): Likewise.
2410 (optimize_inline_calls): Likewise.
2411 (expand_call_inline): Likewise.
2412 (estimate_num_insns): Likewise.
2413 (copy_bb): Likewise.
2414 (delete_unreachable_blocks_update_callgraph): Likewise.
2415 * tree-nested.c (gimplify_all_functions): Likewise.
2416 (create_nesting_tree): Likewise.
2417 (check_for_nested_with_variably_modified): Likewise.
2418 * tree-pretty-print.c (dump_function_header): Likewise.
2419 * tree-profile.c (tree_profiling): Likewise.
2420 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
2421 (modify_function): Likewise.
2422 (convert_callers): Likewise.
2423 (convert_callers_for_node): Likewise.
2424 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
2425 (associate_varinfo_to_alias): Likewise.
2426 (create_variable_info_for): Likewise.
2427 (get_constraint_for_ssa_var): Likewise.
2428 * tree-vectorizer.c (increase_alignment): Likewise.
2429 * tree.c (find_decls_types_in_var): Likewise.
2430 (find_decls_types_in_node): Likewise.
2431 (free_lang_data_in_decl): Likewise.
2432 * value-prof.c (gimple_ic_transform): Likewise.
2433 (gimple_ic): Likewise.
2434 (check_ic_target): Likewise.
2435 (init_node_map): Likewise.
2436 * varasm.c (decl_binds_to_current_def_p): Likewise.
2437 (default_binds_local_p_1): Likewise.
2438 (dump_tm_clone_pairs): Likewise.
2439 (assemble_alias): Likewise.
2440 (find_decl): Likewise.
2441 (mark_decl_referenced): Likewise.
2442 * varpool.c (varpool_for_node_and_aliases): Likewise.
2443 (varpool_extra_name_alias): Likewise.
2444 (varpool_create_variable_alias): Likewise.
2445 (add_new_static_var): Likewise.
2446 (varpool_finalize_named_section_flags): Likewise.
2447 (varpool_remove_unreferenced_decls): Likewise.
2448 (enqueue_node): Likewise.
2449 (varpool_assemble_decl): Likewise.
2450 (assemble_aliases): Likewise.
2451 (varpool_analyze_node): Likewise.
2452 (cgraph_variable_initializer_availability): Likewise.
2453 (varpool_add_new_variable): Likewise.
2454 (ctor_for_folding): Likewise.
2455 (dump_varpool_node): Likewise.
2456 (varpool_remove_initializer): Likewise.
2457 (varpool_remove_node): Likewise.
2458 (varpool_node_for_decl): Likewise.
2459 (varpool_create_empty_node): Likewise.
2460 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
2461 (ix86_get_function_versions_dispatcher): Likewise.
2463 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2465 * cgraph.h (symtab_node_base): Convert to a class;
2466 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
2467 chain_next/prev from symtab_node_def.
2468 (cgraph_node): Inherit from symtab_node; add GTY option
2469 tag ("SYMTAB_FUNCTION").
2470 (varpool_node): Inherit from symtab_node; add GTY option
2471 tag ("SYMTAB_VARIABLE").
2472 (symtab_node_def): Remove.
2473 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
2474 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
2475 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
2476 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
2478 * ipa-ref.h (symtab_node_def): Drop.
2479 (symtab_node): Change underlying type from symtab_node_def to
2481 (const_symtab_node): Likwise.
2483 * is-a.h: Update examples in comment.
2485 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
2486 (assembler_name_hash): Likewise.
2488 2013-10-29 Martin Liska <marxin.liska@gmail.com>
2490 * doc/tree-ssa.texi (gimple_phi_result): Document.
2491 (gimple_phi_num_args, gimple_phi_arg): Likewise.
2492 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
2493 (PHI_RESULT, PHI_NUM_ARGS): Remove.
2494 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
2496 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2498 * expr.h: Revert change and include tree-core.h.
2499 * rtl.h: Revert change and don't include tree-core.h.
2501 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2503 * config/darwin.c: Include gimple.h.
2504 * config/i386/winnt.c: Likewise.
2506 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
2508 PR tree-optimization/19831
2509 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
2511 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2513 * tree-outof-ssa.h: Remove include files.
2514 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
2516 * tree-ssa-coalesce.c: Likewise.
2517 * cfgexpand.c: Likewise.
2518 * tree-ssa-ter.c: Likewise.
2519 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
2520 * lto-streamer.h: Likewise.
2521 * cgraphbuild.c: Add gimple.h to include list.
2522 * data-streamer-in.c: Likewise.
2523 * ipa-cp.c: Likewise.
2524 * tree-streamer.c: Likewise.
2525 * lto-compress.c: Likewise.
2526 * ipa-reference.c: Likewise.
2527 * data-streamer-out.c: Likewise.
2528 * lto-cgraph.c: Likewise.
2529 * cgraphclones.c: Likewise.
2530 * ipa-utils.c: Likewise.
2531 * data-streamer.c: Likewise.
2532 * ipa-split.c: Likewise.
2533 * lto-section-in.c: Likewise.
2534 * tree-streamer-out.c: Likewise.
2535 * ipa-prop.c: Likewise.
2536 * tree-streamer-in.c: Likewise.
2537 * symtab.c: Likewise.
2538 * opts-global.c: Likewise.
2539 * lto-opts.c: Likewise.
2540 * lto-section-out.c: Likewise.
2541 * lto-streamer.c: Likewise.
2542 * rtl.h: Add tree-core.h to include list.
2543 * expr.h: Remove tree-core.h from include list.
2544 * gimple.h: Likewise.
2545 * ipa-utils.h: Likewise.
2546 * streamer-hooks.h: Likewise.
2547 * streamer-hooks.c: Include input.h.
2549 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2551 * config/arm/arm.c (cortexa7_extra_costs): New table.
2552 (arm_cortex_a7_tune): New.
2553 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
2555 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
2557 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
2559 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2561 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
2562 to use autogenerated markers for a class-hierarchy, every class
2563 must have a GTY marker.
2564 * gengtype.h (struct type): Add linked list of subclasses to
2565 the "s" member of the union.
2566 (add_subclass): New decl.
2567 * gengtype-state.c (read_state_struct_type): Set up subclass
2569 * gengtype.c (get_ultimate_base_class): New.
2570 (add_subclass): New.
2571 (new_structure): Set up subclass linked list.
2572 (set_gc_used_type): Propagate usage information to subclasses.
2573 (output_mangled_typename): Use get_ultimate_base_class.
2574 (walk_subclasses): Use the subclass linked list, avoiding an
2575 O(N^2) when writing out all types.
2576 (walk_type): Issue an error if the base class is missing a tag,
2577 rather than generating bogus C code. Add a gcc_unreachable
2578 default case, in case people omit tags from concrete subclasses,
2579 or get the values wrong.
2580 (write_func_for_structure): Issue an error for subclasses for
2581 which the base doesn't have a "desc", since otherwise the
2582 autogenerated routines for the base would silently fail to visit
2583 any subclass fields.
2584 (write_root): Use get_ultimate_base_class, tweaking constness of
2585 tp to match that function's signature.
2587 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2589 * doc/gty.texi (GTY Options): Add note about inheritance to
2590 description of desc and tag.
2591 (Inheritance and GTY): New.
2593 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2595 * gengtype-parse.c (opts_have): Drop "static" so that
2596 we can use this from gengtype.c.
2597 * gengtype.c (set_gc_used_type): Mark any base class as used;
2598 update field traversal to visit inherited fields.
2599 (output_mangled_typename): Convert references to classes within
2600 an inheritance hierarchy to reference the ultimate base class,
2601 since only it will have gt_ functions.
2602 (get_string_option): New.
2603 (walk_subclasses): New.
2604 (walk_type): Treat GTY structs that have a "desc" as being the
2605 root of an inheritance hierarchy. Generate a switch on it
2606 within the marking function which walks all subclasses, adding
2607 cases for them via walk_subclasses. For subclasses, visit all
2608 fields of the type (including inherited ones).
2609 (write_func_for_structure): Don't write fns for subclasses, only
2610 for the ultimate base class within an inheritance hierarchy.
2611 Subclasses-marking will be handled by the base class marking functions.
2612 (write_types): Likewise.
2613 (write_local_func_for_structure): Likewise.
2614 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
2615 a "tag" option (and are thus concrete subclasses).
2616 (write_root): Use the marker function for the ultimate base class.
2617 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
2618 (opts_have): Add declaration.
2620 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
2622 * lra-spills.c (lra_final_code_change): Remove useless move insns
2623 originated from moves of pseudos.
2625 2013-10-28 Jeff Law <law@redhat.com>
2627 * lower-subreg.c (resolve_simple_move): Fix comment typo.
2629 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
2631 * df-scan.c (df_collection_rec): Adjust.
2632 (copy_defs): New constant.
2633 (copy_uses): Likewise.
2634 (copy_eq_uses): Likewise.
2635 (copy_mw): Likewise.
2636 (copy_all): Likewise.
2637 (df_insn_rescan): Adjust.
2638 (df_notes_rescan): Likewise.
2639 (df_swap_refs): Likewise.
2640 (df_sort_and_compress_refs): Likewise.
2641 (df_sort_and_compress_mws): Likewise.
2642 (df_install_refs): Likewise.
2643 (df_install_mws): Likewise.
2644 (df_refs_add_to_chains): Add flags parameter controlling which vectors
2646 (df_bb_refs_record): Adjust.
2647 (df_record_entry_block_defs): Likewise.
2648 (df_record_exit_block_defs): Likewise.
2649 (df_refs_verify): Likewise.
2650 (df_mws_verify): Likewise.
2651 (df_insn_refs_verify): Likewise.
2652 (df_bb_verify): Likewise.
2653 * ipa-pure-const.c (finish_state): Remove.
2654 (propagate): Adjust.
2655 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
2656 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
2657 var-tracking.c: Adjust.
2658 * vec.c (stack_vecs): Remove.
2659 (register_stack_vec): Likewise.
2660 (stack_vec_register_index): Likewise.
2661 (unregister_stack_vec): Likewise.
2662 * vec.h (struct va_stack): Remove.
2663 (struct vec<T, A, vl_ptr>): Specialize as
2664 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
2665 allocation strategy compatable with the vl_ptr layout.
2666 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
2667 specialization anyway.
2668 (class stack_vec): New class.
2669 (vec_stack_alloc): Remove.
2670 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
2672 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2673 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2674 Sergey Lega <sergey.s.lega@intel.com>
2675 Anna Tikhonova <anna.tikhonova@intel.com>
2676 Ilya Tocar <ilya.tocar@intel.com>
2677 Andrey Turetskiy <andrey.turetskiy@intel.com>
2678 Ilya Verbin <ilya.verbin@intel.com>
2679 Kirill Yukhin <kirill.yukhin@intel.com>
2680 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2682 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
2683 (*prefetch_avx512pf_<mode>): New.
2684 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
2685 (avx512f_maskcmp<mode>3): Ditto.
2686 (vashrv16si3): Ditto.
2688 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2689 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2690 Sergey Lega <sergey.s.lega@intel.com>
2691 Anna Tikhonova <anna.tikhonova@intel.com>
2692 Ilya Tocar <ilya.tocar@intel.com>
2693 Andrey Turetskiy <andrey.turetskiy@intel.com>
2694 Ilya Verbin <ilya.verbin@intel.com>
2695 Kirill Yukhin <kirill.yukhin@intel.com>
2696 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2698 * config/i386/i386.md (any_truncate): New.
2699 (trunsuffix): Ditto.
2700 * config/i386/predicates.md (const_8_to_9_operand): New.
2701 (const_10_to_11_operand): Ditto.
2702 (const_12_to_13_operand): Ditto.
2703 (const_14_to_15_operand): Ditto.
2704 (const_16_to_19_operand): Ditto.
2705 (const_20_to_23_operand): Ditto.
2706 (const_24_to_27_operand): Ditto.
2707 (const_28_to_31_operand): Ditto.
2708 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
2709 (cvtusi2<ssescalarmodesuffix>32): New.
2710 (cvtusi2<ssescalarmodesuffix>64): Ditto.
2711 (ufloatv16siv16sf2): Ditto.
2712 (avx512f_fix_notruncv16sfv16si): Ditto.
2713 (avx512f_ufix_notruncv16sfv16si): Ditto.
2714 (avx512f_vcvtss2usi): Ditto.
2715 (avx512f_vcvtss2usiq): Ditto.
2716 (avx512f_vcvttss2usi): Ditto.
2717 (avx512f_vcvttss2usiq): Ditto.
2718 (avx512f_vcvtsd2usi): Ditto.
2719 (avx512f_vcvtsd2usiq): Ditto.
2720 (avx512f_vcvttsd2usi): Ditto.
2721 (avx512f_vcvttsd2usiq): Ditto.
2722 (ufloatv8siv8df): Ditto.
2723 (avx512f_cvtdq2pd512_2): Ditto.
2724 (avx512f_cvtpd2dq512): Ditto.
2725 (avx512f_ufix_notruncv8dfv8si): Ditto.
2726 (avx512f_cvtpd2ps512): Ditto.
2727 (vec_unpacks_lo_v16sf): Ditto.
2728 (vec_unpacks_hi_v16sf): Ditto.
2729 (vec_unpacks_float_hi_v16si): Ditto.
2730 (vec_unpacks_float_lo_v16si): Ditto.
2731 (avx512f_unpckhps512): Ditto.
2732 (avx512f_unpcklps512): Ditto.
2733 (avx512f_movshdup512): Ditto.
2734 (avx512f_movsldup512): Ditto.
2735 (vec_extract_lo_v32hi): Ditto.
2736 (vec_extract_hi_v32hi): Ditto.
2737 (vec_extract_lo_v64qi): Ditto.
2738 (vec_extract_hi_v64qi): Ditto.
2739 (avx512f_unpckhpd512): Ditto.
2740 (avx512f_movddup512): Ditto.
2741 (avx512f_unpcklpd512): Ditto.
2742 (*avx512f_unpcklpd512): Ditto.
2743 (avx512f_shufps512_1): Ditto.
2744 (avx512f_shufpd512_1): Ditto.
2745 (avx512f_interleave_highv8di): Ditto.
2746 (avx512f_interleave_lowv8di): Ditto.
2747 (PMOV_DST_MODE): Ditto.
2748 (pmov_src_mode): Ditto.
2749 (pmov_src_lower): Ditto.
2751 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
2752 (*avx512f_<code>v8div16qi2): Ditto.
2753 (*avx512f_<code>v8div16qi2_store): Ditto.
2754 (vec_widen_umult_even_v16si): Ditto.
2755 (*vec_widen_umult_even_v16si): Ditto.
2756 (vec_widen_smult_even_v16si): Ditto.
2757 (*vec_widen_smult_even_v16si): Ditto.
2758 (avx512f_interleave_highv16si): Ditto.
2759 (avx512f_interleave_lowv16si): Ditto.
2760 (avx512f_<code>v16qiv16si2): Ditto.
2761 (avx512f_<code>v16hiv16si2): Ditto.
2762 (avx512f_<code>v8qiv8di2): Ditto.
2763 (avx512f_<code>v8hiv8di2): Ditto.
2764 (avx512f_<code>v8siv8di2): Ditto.
2765 (avx512cd_maskb_vec_dupv8di): Ditto.
2766 (avx512cd_maskw_vec_dupv16si): Ditto.
2767 (avx512f_vcvtph2ps512): Ditto.
2768 (avx512f_vcvtps2ph512): Ditto.
2769 (VEC_EXTRACT_MODE): Extened with wider modes.
2770 (VEC_PERM_AVX2): Ditto.
2771 (VEC_PERM_CONST): Ditto.
2773 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
2775 * config/arc/arc.c (arc_ccfsm_post_advance):
2776 Add comment about TYPE_RETURN.
2778 2013-10-28 Bin Cheng <bin.cheng@arm.com>
2780 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
2781 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
2782 (strip_offset): Convert offset to unsigned number.
2784 2013-10-27 Tom de Vries <tom@codesourcery.com>
2786 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
2787 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
2789 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
2791 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
2792 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
2793 (TARGET_BIG_ENDIAN): New macro.
2794 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
2795 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
2796 * config/sh/sh.c: Likewise.
2797 * config/sh/sh.h: Likewise.
2799 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
2801 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
2802 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
2804 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2806 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
2808 (cmpstr_t, cmpstrsi): Fix formatting.
2810 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2813 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
2814 addressing, do not use general_operand for memory operands.
2816 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
2819 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2820 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2822 2013-10-26 Jeff Law <law@redhat.com>
2824 * predict.c (PRED_MUDFLAP): Remove.
2825 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
2827 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
2828 (OBJS): Remove tree-nomudflap.o
2829 (GTFILES): Remove tree-mudflap.c
2830 * builtins.c (expand_builtin_alloc): Remove mudflap support.
2831 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2832 (mfwrap_spec, mflib_spec): Likewise.
2833 (cpp_unique_options, cc1_options, static_specs): Likewise.
2834 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
2835 * passes.def: Likewise.
2836 * toplev.c (compile_file, process_options): Likewise.
2837 * tree-inline.c (copy_tree_r): Likewise.
2838 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
2839 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
2840 (build_constant_desc, output_constant_def_contents): Likewise.
2841 (categorize_decl_for_section): Likewise.
2842 * tree-mudflap.c: Removed.
2843 * tree-mudflap.h: Removed.
2844 * tree-nomudflap.c: Removed.
2845 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
2846 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
2847 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2848 * config/sol2.h (MFLIB_SPEC): Likewise.
2849 * doc/install.texi: Remove mudflap references.
2850 * doc/passes.texi: Similarly.
2851 * doc/sourcebuild.texi: Similarly.
2852 * doc/invoke.texi: Remove mudlfap related options.
2854 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2856 PR rtl-optimization/58759
2857 * lra-constraints.c (lra_constraints): Remove wrong condition to
2858 remove insn setting up an equivalent pseudo.
2860 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2862 * config/rs6000/rs6000-protos.h
2863 (rs6000_secondary_memory_needed_mode): New prototype.
2864 * config/rs6000/rs6000.c: Include ira.h.
2865 (TARGET_LRA_P): Redefine.
2866 (rs6000_legitimate_offset_address_p): Call
2867 legitimate_constant_pool_address_p in strict mode for LRA.
2868 (rs6000_legitimate_address_p): Ditto.
2869 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2870 (rs6000_emit_move): Add LRA version of code to generate load/store
2872 (rs6000_secondary_memory_needed_mode): New.
2873 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2874 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2875 constants, memory, and FP registers.
2876 (rs6000_lra_p): New.
2877 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2878 * config/rs6000/rs6000.opt (mlra): New option.
2879 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2881 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2883 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2884 has_single_use () and not do the conversion if has_single_use ()
2885 returns false for the multiplication result.
2887 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2889 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2890 to "NODE", since this works on a "tree", not an
2892 (CONSTANT_CLASS_P): Likewise.
2895 (INDIRECT_REF_P): Likewise.
2896 (REFERENCE_CLASS_P): Likewise.
2897 (COMPARISON_CLASS_P): Likewise.
2898 (UNARY_CLASS_P): Likewise.
2899 (BINARY_CLASS_P): Likewise.
2900 (STATEMENT_CLASS_P): Likewise.
2901 (VL_EXP_CLASS_P): Likewise.
2902 (EXPRESSION_CLASS_P): Likewise.
2903 (IS_TYPE_OR_DECL_P): Likewise.
2905 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2907 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2908 ADDR_EXPR in the defining statement.
2910 2013-10-25 Richard Biener <rguenther@suse.de>
2912 PR tree-optimization/58626
2913 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2914 anti_dd, output_dd and input_dd.
2915 (struct rdg_edge): Remove level and relation members.
2916 (RDGE_LEVEL, RDGE_RELATION): Remove.
2917 (dot_rdg_1): Adjust.
2918 (create_rdg_edge_for_ddr): Remove.
2919 (create_rdg_edges_for_scalar): Adjust.
2920 (create_edge_for_control_dependence): Likewise.
2921 (create_rdg_edges): Split into ...
2922 (create_rdg_flow_edges): ... this
2923 (create_rdg_cd_edges): ... and this.
2925 (build_rdg): Likewise, do not compute data dependences or
2927 (pg_add_dependence_edges): New function.
2929 (distribute_loop): First apply all non-dependence based
2930 partition mergings. Then compute dependences between partitions
2931 and merge and order partitions according to them.
2933 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2935 PR rtl-optimization/58831
2936 * alias.c (init_alias_analysis): At the beginning of each iteration,
2937 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2939 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2941 * recog.c (search_ofs): New static variable moved from...
2942 (peep2_find_free_register): ...here.
2943 (peephole2_optimize): Initialize it.
2945 2013-10-25 Tobias Burnus <burnus@net-b.de>
2947 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2949 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2951 * config/i386/i386.h (TARGET_MPX): New define.
2952 (TARGET_MPX_P): Ditto.
2954 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2956 * config/i386/constraints.md (B): New.
2959 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2960 * config/i386/i386-modes.def (BND32): New.
2962 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2963 * config/i386/i386.c (isa_opts): Add mmpx.
2964 (regclass_map): Add bound registers.
2965 (dbx_register_map): Likewise.
2966 (dbx64_register_map): Likewise.
2967 (svr4_dbx_register_map): Likewise.
2969 (ix86_option_override_internal): Support MPX ISA.
2970 (ix86_conditional_register_usage): Support bound registers.
2971 (print_reg): Likewise.
2972 (ix86_code_end): Add MPX bnd prefix.
2973 (output_set_got): Likewise.
2974 (ix86_output_call_insn): Likewise.
2975 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2976 (ix86_print_operand_punct_valid_p): Likewise.
2977 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2979 (ix86_class_likely_spilled_p): Add bound regs support.
2980 (ix86_hard_regno_mode_ok): Likewise.
2981 (x86_order_regs_for_local_alloc): Likewise.
2982 (ix86_bnd_prefixed_insn_p): New.
2983 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2984 (FIXED_REGISTERS): Add bound registers.
2985 (CALL_USED_REGISTERS): Likewise.
2986 (REG_ALLOC_ORDER): Likewise.
2987 (HARD_REGNO_NREGS): Likewise.
2989 (VALID_BND_REG_MODE): New.
2990 (FIRST_BND_REG): New.
2991 (LAST_BND_REG): New.
2992 (reg_class): Add BND_REGS.
2993 (REG_CLASS_NAMES): Likewise.
2994 (REG_CLASS_CONTENTS): Likewise.
2996 (ANY_BND_REG_P): New.
2998 (HI_REGISTER_NAMES): Add bound registers.
2999 * config/i386/i386.md (UNSPEC_BNDMK): New.
3000 (UNSPEC_BNDMK_ADDR): New.
3001 (UNSPEC_BNDSTX): New.
3002 (UNSPEC_BNDLDX): New.
3003 (UNSPEC_BNDLDX_ADDR): New.
3004 (UNSPEC_BNDCL): New.
3005 (UNSPEC_BNDCU): New.
3006 (UNSPEC_BNDCN): New.
3007 (UNSPEC_MPX_FENCE): New.
3010 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
3011 (length_immediate): Likewise.
3012 (prefix_0f): Likewise.
3014 (prefix_rep): Check for bnd prefix.
3015 (length_nobnd): New.
3016 (length): Use length_nobnd if specified.
3021 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
3024 (simple_return_internal): Likewise.
3025 (simple_return_pop_internal): Likewise.
3026 (*indirect_jump): Add MPX bnd prefix.
3027 (*tablejump_1): Likewise.
3028 (simple_return_internal_long): Likewise.
3029 (simple_return_indirect_internal): Likewise.
3033 (*mov<mode>_internal_mpx): New.
3034 (<mode>_<bndcheck>): New.
3035 (*<mode>_<bndcheck>): New.
3040 * config/i386/predicates.md (lea_address_operand): Rename to...
3041 (address_no_seg_operand): ... this.
3042 (address_mpx_no_base_operand): New.
3043 (address_mpx_no_index_operand): New.
3044 (bnd_mem_operator): New.
3045 * config/i386/i386.opt (mmpx): New.
3046 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
3047 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
3049 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
3051 * mode-classes.def (MODE_POINTER_BOUNDS): New.
3052 * tree.def (POINTER_BOUNDS_TYPE): New.
3053 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
3054 (POINTER_BOUNDS_MODE): New.
3055 (make_pointer_bounds_mode): New.
3056 * machmode.h (POINTER_BOUNDS_MODE_P): New.
3057 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
3058 (layout_type): Support POINTER_BOUNDS_TYPE.
3059 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
3060 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
3061 (type_contains_placeholder_1): Likewise.
3062 * tree.h (POINTER_BOUNDS_TYPE_P): New.
3063 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
3064 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
3066 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
3068 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
3069 mode of address computation when calling memory_address_addr_space.
3071 2013-08-24 Richard Henderson <rth@twiddle.net>
3074 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
3075 instead of create_convert_operand_to.
3076 (maybe_emit_sync_lock_test_and_set): Likewise.
3077 (expand_atomic_compare_and_swap): Likewise.
3078 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
3080 2013-08-24 Sriraman Tallam <tmsriram@google.com>
3082 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
3083 Change param_is to use the struct and not the pointer to the struct.
3085 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
3087 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
3088 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
3089 Move to targhooks.c.
3090 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
3091 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
3092 Relocate from gimplify.c.
3093 * targhooks.h: Add 2 prototypes.
3094 * tree.h: Delete 2 prototypes.
3096 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
3098 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
3099 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
3101 2013-10-24 Cong Hou <congh@google.com>
3103 * convert.c (convert_to_real): Guard those unsafe math function
3104 convertions with flag_unsafe_math_optimizations. Handle sqrt()
3107 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
3110 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
3112 (cgraph_create_edge): Use the new argument.
3113 (cgraph_create_indirect_edge): Likewise.
3115 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
3117 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
3119 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
3120 changes statep->state.
3122 2013-10-24 Tobias Burnus <burnus@net-b.de>
3125 * tree-cfg.c (replace_loop_annotate): New function.
3126 (execute_build_cfg): Call it.
3127 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
3128 * internal-fn.c (expand_ANNOTATE): New function.
3129 * internal-fn.def (ANNOTATE): Define as new internal function.
3130 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
3131 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
3132 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
3133 * doc/extend.texi (Pragmas): Document #pragma ivdep.
3134 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
3136 2013-10-17 Ian Bolton <ian.bolton@arm.com>
3137 Marcus Shawcroft <marcus.shawcroft@arm.com>
3139 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
3140 Special case reload SP+C into none GENERAL_REGS.
3142 2013-10-24 Michael Matz <matz@suse.de>
3144 * gengtype.c (is_file_equal): Check that files will be same length.
3146 2013-10-25 Christian Bruel <christian.bruel@st.com>
3148 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
3149 * config/sh/t-sh (sh-mem.o): New rule.
3150 * config/sh/sh-mem.cc (expand_block_move): Moved here.
3151 (sh_expand_cmpstr): New function.
3152 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
3153 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
3154 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
3155 (rotlhi3_8): Rename.
3157 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3159 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
3160 * Makefile.in (CXXDEPMODE): Assign and change users.
3161 (CCDEPMODE): Delete.
3162 * configure: Regenerate.
3164 2013-10-23 David Malcolm <dmalcolm@redhat.com>
3166 * gengtype-parse.c (require_without_advance): New.
3167 (type): For GTY-marked types that are not GTY((user)), parse any
3168 base classes, requiring them to be single-inheritance, and not
3169 be templates. For non-GTY-marked types and GTY((user)),
3170 continue to skip over any C++ inheritance specification.
3171 * gengtype-state.c (state_writer::write_state_struct_type):
3172 Write base class of type (if any).
3173 (read_state_struct_type): Read base class of type (if any).
3174 * gengtype.c (new_structure): Add a "base_class" parameter.
3175 (create_optional_field_): Update for new parameter to new_structure.
3176 (adjust_field_rtx_def): Likewise.
3177 (adjust_field_tree_exp): Likewise.
3178 * gengtype.h (struct type): Add "base_class" field to the s
3180 (new_structure): Add "base" parameter.
3182 2013-10-23 Sriraman Tallam <tmsriram@google.com>
3185 * config/i386/i386.c (ix86_option_override_internal):
3186 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
3187 (ix86_valid_target_attribute_tree):
3188 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
3189 Change TARGET_SSE to TARGET_SSE_P (opts->...)
3191 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
3193 * tree-ssa-loop.h: Remove include files.
3194 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
3195 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
3196 * graphite-clast-to-gimple.c: Likewise.
3197 * graphite-scop-detection.c: Likewise.
3198 * graphite-sese-to-poly.c: Likewise.
3199 * ipa-inline-analysis.c: Likewise.
3200 * ipa-pure-const.c: Likewise.
3201 * loop-init.c: Likewise.
3202 * passes.c: Likewise.
3203 * predict.c: Likewise.
3204 * tree-cfg.c: Likewise.
3205 * tree-cfgcleanup.c: Likewise.
3206 * tree-chrec.c: Likewise.
3207 * tree-data-ref.c: Likewise.
3208 * tree-loop-distribution.c: Likewise.
3209 * tree-parloops.c: Likewise.
3210 * tree-predcom.c: Likewise.
3211 * tree-scalar-evolution.c: Likewise.
3212 * tree-ssa-address.c: Likewise.
3213 * tree-ssa.c: Likewise.
3214 * tree-ssa-dce.c: Likewise.
3215 * tree-ssa-loop.c: Likewise.
3216 * tree-ssa-loop-im.c: Likewise.
3217 * tree-ssa-loop-ivcanon.c: Likewise.
3218 * tree-ssa-loop-ivopts.c: Likewise.
3219 * tree-ssa-loop-manip.c: Likewise.
3220 * tree-ssa-loop-niter.c: Likewise.
3221 * tree-ssa-loop-prefetch.c: Likewise.
3222 * tree-ssa-loop-unswitch.c: Likewise.
3223 * tree-ssa-reassoc.c: Likewise.
3224 * tree-vect-data-refs.c: Likewise.
3225 * tree-vect-loop.c: Likewise.
3226 * tree-vect-loop-manip.c: Likewise.
3227 * tree-vectorizer.c: Likewise.
3228 * tree-vect-stmts.c: Likewise.
3229 * tree-vrp.c: Likewise.
3231 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3233 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
3235 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3237 PR tree-optimization/58775
3238 PR tree-optimization/58791
3239 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
3240 (insert_stmt_after): Rewritten, don't move the stmt, but really
3242 (get_stmt_uid_with_default): Remove.
3243 (build_and_add_sum): Use insert_stmt_after and
3244 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
3245 (update_range_test): Set uid on stmts added by
3246 force_gimple_operand_gsi. Don't immediately modify statements
3247 in inter-bb optimization, just update oe->op values.
3248 (optimize_range_tests): Return bool whether any changed have been made.
3249 (update_ops): New function.
3250 (struct inter_bb_range_test_entry): New type.
3251 (maybe_optimize_range_tests): Perform statement changes here.
3252 (not_dominated_by, appears_later_in_bb, get_def_stmt,
3253 ensure_ops_are_available): Remove.
3254 (find_insert_point): Rewritten.
3255 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
3256 return LHS of the (new resp. old) stmt. Don't call
3257 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
3258 instead of last, move new stmt at the right place.
3259 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
3260 (negate_value): Likewise. Set uids.
3261 (break_up_subtract_bb): Initialize uids.
3262 (reassociate_bb): Adjust rewrite_expr_tree caller.
3263 (do_reassoc): Don't call renumber_gimple_stmt_uids.
3265 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
3268 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
3269 TARGET_32BIT final condition.
3270 (mulsi3_internal2 and splitter): Same.
3272 2013-10-23 Jeff Law <law@redhat.com>
3274 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
3275 through joiner blocks with abnormal outgoing edges.
3277 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
3278 Add parameter JOINERS, to allow/disallow threading through joiner
3280 (thread_block): New. Call thread_block_1.
3281 (mark_threaded_blocks): Remove code to filter out certain cases
3282 of threading through joiner blocks.
3283 (thread_through_all_blocks): Document how we can have a dangling
3284 edge AUX field and clear it.
3286 2013-10-23 Ian Lance Taylor <iant@google.com>
3288 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
3289 (C++ Dialect Options): Likewise.
3290 (Optimize Options): Likewise.
3292 2013-10-23 Tom de Vries <tom@codesourcery.com>
3294 PR tree-optimization/58805
3295 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
3297 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3299 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
3300 sequence based on get_range_info returned range.
3302 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
3304 * tree-ssa.h: Remove all #include's
3305 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
3306 * alias.c: Move required includes from tree-ssa.h.
3308 * builtins.c: Likewise.
3309 * calls.c: Likewise.
3310 * cfgexpand.c: Likewise.
3311 * cfghooks.c: Likewise.
3312 * cfgloop.c: Likewise.
3313 * cfgloopmanip.c: Likewise.
3314 * cgraph.c: Likewise.
3315 * cgraphbuild.c: Likewise.
3316 * cgraphclones.c: Likewise.
3317 * cgraphunit.c: Likewise.
3319 * except.c: Likewise.
3321 * final.c: Likewise.
3322 * fold-const.c: Likewise.
3323 * ggc-page.c: Likewise.
3324 * gimple-builder.c: Likewise.
3325 * gimple-fold.c: Likewise.
3326 * gimple-iterator.c: Likewise.
3327 * gimple-low.c: Likewise.
3328 * gimple-pretty-print.c: Likewise.
3329 * gimple-ssa-strength-reduction.c: Likewise.
3330 * gimple-streamer-in.c: Likewise.
3331 * gimple-streamer-out.c: Likewise.
3332 * gimplify.c: Likewise.
3333 * graphite-blocking.c: Likewise.
3334 * graphite-clast-to-gimple.c: Likewise.
3335 * graphite-dependences.c: Likewise.
3336 * graphite-interchange.c: Likewise.
3337 * graphite-optimize-isl.c: Likewise.
3338 * graphite-poly.c: Likewise.
3339 * graphite-scop-detection.c: Likewise.
3340 * graphite-sese-to-poly.c: Likewise.
3341 * graphite.c: Likewise.
3342 * ipa-cp.c: Likewise.
3343 * ipa-inline-analysis.c: Likewise.
3344 * ipa-inline-transform.c: Likewise.
3345 * ipa-inline.c: Likewise.
3346 * ipa-prop.c: Likewise.
3347 * ipa-pure-const.c: Likewise.
3348 * ipa-reference.c: Likewise.
3349 * ipa-split.c: Likewise.
3350 * ipa-utils.c: Likewise.
3351 * loop-init.c: Likewise.
3352 * lto-cgraph.c: Likewise.
3353 * lto-section-in.c: Likewise.
3354 * lto-section-out.c: Likewise.
3355 * lto-streamer-in.c: Likewise.
3356 * lto-streamer-out.c: Likewise.
3357 * lto-streamer.c: Likewise.
3358 * omp-low.c: Likewise.
3359 * passes.c: Likewise.
3360 * predict.c: Likewise.
3361 * print-tree.c: Likewise.
3362 * profile.c: Likewise.
3364 * targhooks.c: Likewise.
3365 * tracer.c: Likewise.
3366 * trans-mem.c: Likewise.
3367 * tree-call-cdce.c: Likewise.
3368 * tree-cfg.c: Likewise.
3369 * tree-cfgcleanup.c: Likewise.
3370 * tree-chrec.c: Likewise.
3371 * tree-complex.c: Likewise.
3372 * tree-data-ref.c: Likewise.
3373 * tree-dfa.c: Likewise.
3374 * tree-eh.c: Likewise.
3375 * tree-emutls.c: Likewise.
3376 * tree-if-conv.c: Likewise.
3377 * tree-inline.c: Likewise.
3378 * tree-into-ssa.c: Likewise.
3379 * tree-loop-distribution.c: Likewise.
3380 * tree-mudflap.c: Likewise.
3381 * tree-nested.c: Likewise.
3382 * tree-nrv.c: Likewise.
3383 * tree-object-size.c: Likewise.
3384 * tree-outof-ssa.c: Likewise.
3385 * tree-parloops.c: Likewise.
3386 * tree-phinodes.c: Likewise.
3387 * tree-predcom.c: Likewise.
3388 * tree-pretty-print.c: Likewise.
3389 * tree-profile.c: Likewise.
3390 * tree-scalar-evolution.c: Likewise.
3391 * tree-sra.c: Likewise.
3392 * tree-ssa-address.c: Likewise.
3393 * tree-ssa-alias.c: Likewise.
3394 * tree-ssa-ccp.c: Likewise.
3395 * tree-ssa-coalesce.c: Likewise.
3396 * tree-ssa-copy.c: Likewise.
3397 * tree-ssa-copyrename.c: Likewise.
3398 * tree-ssa-dce.c: Likewise.
3399 * tree-ssa-dom.c: Likewise.
3400 * tree-ssa-dse.c: Likewise.
3401 * tree-ssa-forwprop.c: Likewise.
3402 * tree-ssa-ifcombine.c: Likewise.
3403 * tree-ssa-live.c: Likewise.
3404 * tree-ssa-loop-ch.c: Likewise.
3405 * tree-ssa-loop-im.c: Likewise.
3406 * tree-ssa-loop-ivcanon.c: Likewise.
3407 * tree-ssa-loop-ivopts.c: Likewise.
3408 * tree-ssa-loop-manip.c: Likewise.
3409 * tree-ssa-loop-niter.c: Likewise.
3410 * tree-ssa-loop-prefetch.c: Likewise.
3411 * tree-ssa-loop-unswitch.c: Likewise.
3412 * tree-ssa-loop.c: Likewise.
3413 * tree-ssa-math-opts.c: Likewise.
3414 * tree-ssa-operands.c: Likewise.
3415 * tree-ssa-phiopt.c: Likewise.
3416 * tree-ssa-phiprop.c: Likewise.
3417 * tree-ssa-pre.c: Likewise.
3418 * tree-ssa-propagate.c: Likewise.
3419 * tree-ssa-reassoc.c: Likewise.
3420 * tree-ssa-sccvn.c: Likewise.
3421 * tree-ssa-sink.c: Likewise.
3422 * tree-ssa-strlen.c: Likewise.
3423 * tree-ssa-structalias.c: Likewise.
3424 * tree-ssa-tail-merge.c: Likewise.
3425 * tree-ssa-ter.c: Likewise.
3426 * tree-ssa-threadedge.c: Likewise.
3427 * tree-ssa-threadupdate.c: Likewise.
3428 * tree-ssa-uncprop.c: Likewise.
3429 * tree-ssa-uninit.c: Likewise.
3430 * tree-ssa.c: Likewise.
3431 * tree-ssanames.c: Likewise.
3432 * tree-stdarg.c: Likewise.
3433 * tree-streamer-in.c: Likewise.
3434 * tree-switch-conversion.c: Likewise.
3435 * tree-tailcall.c: Likewise.
3436 * tree-vect-data-refs.c: Likewise.
3437 * tree-vect-generic.c: Likewise.
3438 * tree-vect-loop-manip.c: Likewise.
3439 * tree-vect-loop.c: Likewise.
3440 * tree-vect-patterns.c: Likewise.
3441 * tree-vect-slp.c: Likewise.
3442 * tree-vect-stmts.c: Likewise.
3443 * tree-vectorizer.c: Likewise.
3444 * tree-vrp.c: Likewise.
3447 * value-prof.c: Likewise.
3448 * var-tracking.c: Likewise.
3449 * varpool.c: Likewise.
3450 * vtable-verify.c: Likewise.
3452 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3454 * config/tilegx/tilegx.c: Include "tree.h".
3456 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3458 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
3459 the info, not after it.
3460 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
3461 dump_ssaname_info call.
3462 (pp_gimple_stmt_1): Adjust caller.
3463 (dump_phi_nodes): Likewise. Don't print "# " here.
3465 2013-10-22 Jan Hubicka <jh@suse.cz>
3467 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
3469 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
3470 * i386.c (expand_small_movmem_or_setmem): New function.
3471 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
3473 (alg_usable_p): Add support for value ranges; cleanup.
3474 (ix86_expand_set_or_movmem): Add support for misaligned moves.
3476 2013-10-22 Sterling Augustine <saugustine@google.com>
3478 * doc/invoke.texi: Document -ggnu-pubnames.
3479 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
3481 * dwarf2out.c: Include gdb/gdb-index.h.
3482 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
3483 debug_generate_pub_sections.
3484 (is_java, output_pubtables, output_pubname): New functions.
3485 (include_pubname_in_output): Handle debug_generate_pub_sections at
3487 (size_of_pubnames): Use new local space_for_flags based on
3488 debug_generate_pub_sections.
3489 (output_pubnames): Unify pubnames and pubtypes output logic.
3490 Genericize comments. Call output_pubname.
3491 (dwarf2out_finish): Move logic to output_pubnames and call it.
3493 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
3496 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
3497 Remove CCCmode handling.
3498 <case LTU>: Return 'c' suffix for CCCmode.
3499 <case GEU>: Return 'nc' suffix for CCCmode.
3500 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
3501 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
3502 (*sub<mode>3_cc_overflow): Ditto.
3503 (*subsi3_zext_cc_overflow): Ditto.
3505 2013-10-22 Steve Ellcey <sellcey@mips.com>
3507 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
3508 (AND (NOT OP1) (NOT OP2)).
3510 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
3512 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3513 meaning of merge-high and merge-low masks for little endian; avoid
3514 use of vector-pack masks for little endian for mismatched modes.
3516 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3518 * config/tilepro/tilepro.c: Include "tree.h".
3520 2013-10-22 Andreas Schwab <schwab@suse.de>
3522 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
3525 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
3527 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
3528 [Ed Smith-Rowland]): New entries.
3529 ([Stephen M. Webb]): Update.
3531 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
3533 * tree-ssa-ter.h: Remove duplicate copy of file contents.
3535 2013-10-21 Marek Polacek <polacek@redhat.com>
3538 * fold-const.c (fold_range_test): Return 0 if the type is not
3541 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
3543 * system.h: Move hwint.h include further down.
3544 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
3545 gcc_checking_asserts.
3546 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
3548 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3550 Fix volatile issues in optimize_bit_field_compare.
3551 * fold-const.c (optimize_bit_field_compare): Bail out if
3552 lvolatilep or rvolatilep.
3554 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3556 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
3557 and get_inner_reference returning different pmode for non-volatile
3558 bit-field members dependent on flag_strict_volatile_bitfields.
3559 * stor-layout.c (layout_decl): Remove special handling of
3560 flag_strict_volatile_bitfields.
3561 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
3562 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
3564 2013-10-21 Paulo Matos <pmatos@broadcom.com>
3566 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
3569 2013-10-21 Jeff Law <law@redhat.com>
3571 * tree-ssa-threadedge.c (thread_through_normal_block): New
3572 argument VISITED. Remove VISISTED as a local variable. When we
3573 have a threadable jump, verify the destination of the jump has not
3575 (thread_across_edge): Allocate VISITED bitmap once at function
3576 scope and use it throughout. Make sure to set appropriate bits in
3577 VISITED for E (start of jump thread path).
3578 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
3579 through a joiner if any edge on the path has a recorded jump thread.
3581 2013-10-21 Ian Lance Taylor <iant@google.com>
3583 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
3584 don't imply that attributes can solve all problems.
3585 (Directory Options): Fix typo.
3587 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3589 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
3590 extend and extend_add.
3592 2013-10-21 Richard Biener <rguenther@suse.de>
3594 PR tree-optimization/58794
3595 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
3596 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
3598 2013-10-21 Richard Biener <rguenther@suse.de>
3601 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
3602 to (T) X for sign-changing conversions (or no conversion).
3604 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
3606 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
3608 2013-10-20 Jan Hubicka <jh@suse.cz>
3610 * config/i386/i386-tune.def: Add comment; organize into categories
3612 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3614 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
3615 argument. Update function comment.
3616 (expand_set_or_movmem_via_rep): New function combining
3617 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
3618 (expand_movmem_via_rep_mov): Remove.
3619 expand_setmem_via_rep_stos): Remove.
3620 (expand_movmem_epilogue): Update calls correspondingly.
3621 (expand_setmem_epilogue_via_loop): Likewise.
3623 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
3624 (expand_set_or_movmem_prologue): New function combining
3625 expand_movmem_prologue and expand_setmem_prologue.
3626 (expand_movmem_prologue): Remove.
3627 (expand_setmem_prologue): Remove.
3628 (expand_set_or_movmem_constant_prologue): New function combining
3629 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
3630 (expand_constant_movmem_prologue): Remove.
3631 (expand_constant_setmem_prologue): Remove.
3632 (promote_duplicated_reg): Allow vector-const0 value.
3633 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
3634 and ix86_expand_setmem.
3635 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
3636 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
3638 2013-10-21 Diego Novillo <dnovillo@google.com>
3640 * asan.c: Include tree.h
3641 * bb-reorder.c: Likewise.
3642 * cfgcleanup.c: Likewise.
3643 * cfgloopmanip.c: Likewise.
3644 * data-streamer-in.c: Likewise.
3645 * data-streamer-out.c: Likewise.
3646 * data-streamer.c: Likewise.
3647 * dwarf2cfi.c: Likewise.
3648 * graphite-blocking.c: Likewise.
3649 * graphite-clast-to-gimple.c: Likewise.
3650 * graphite-dependences.c: Likewise.
3651 * graphite-interchange.c: Likewise.
3652 * graphite-optimize-isl.c: Likewise.
3653 * graphite-poly.c: Likewise.
3654 * graphite-scop-detection.c: Likewise.
3655 * graphite-sese-to-poly.c: Likewise.
3656 * graphite.c: Likewise.
3657 * ipa-devirt.c: Likewise.
3658 * ipa-profile.c: Likewise.
3661 * loop-init.c: Likewise.
3662 * loop-unroll.c: Likewise.
3663 * lower-subreg.c: Likewise.
3664 * lto/lto-object.c: Likewise.
3665 * recog.c: Likewise.
3666 * reginfo.c: Likewise.
3667 * tree-loop-distribution.c: Likewise.
3668 * tree-parloops.c: Likewise.
3669 * tree-ssa-strlen.c: Likewise.
3670 * tree-streamer.c: Likewise.
3671 * value-prof.c: Likewise.
3672 * target-globals.c: Likewise.
3673 * expr.h: Include tree-core.h instead of tree.h.
3674 * gimple.h: Likewise.
3675 * ipa-prop.h: Likewise.
3676 * ipa-utils.h: Likewise.
3677 * lto-streamer.h: Likewise.
3678 * streamer-hooks.h: Likewise.
3679 * ipa-reference.h: Include cgraph.h instead of tree.h.
3680 * cgraph.h: Include basic-block.h instead of tree.h.
3681 * tree-streamer.h: Do not include tree.h.
3682 * genattrtab.c (write_header): Generate inclusion of
3684 * genautomata.c (main): Likewise.
3685 * genemit.c: Likewise.
3686 * genopinit.c: Likewise.
3687 * genoutput.c (output_prologue): Likewise.
3688 * genpeep.c: Likewise.
3690 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3692 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
3694 (vec_unpacku_hi_v8hi): Likewise.
3695 (vec_unpacku_lo_v16qi): Likewise.
3696 (vec_unpacku_lo_v8hi): Likewise.
3698 2013-10-20 Jan Hubicka <jh@suse.cz>
3700 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
3701 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
3702 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
3704 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
3706 * config/mips/mips.h (ISA_HAS_WSBH): Define.
3707 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
3709 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
3711 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
3714 * system.h: Undef m_slot.
3716 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3718 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
3719 all elements for both endian flavors.
3721 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3724 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
3725 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
3726 and SI_REG for 64bit SYSV ABI targets.
3728 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3730 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
3731 to multi_reg_return. Clarify that we are skipping USEs of multiple
3732 return registers. Use bool type where appropriate.
3734 2013-10-18 Jan Hubicka <jh@suse.cz>
3736 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
3738 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
3739 (X86_TUNE_PUSH_MEMORY): Likewise.
3740 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
3741 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
3742 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
3744 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
3745 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
3747 (ix86_option_override_internal): Update to use tune features instead
3750 2013-10-18 Cong Hou <congh@google.com>
3752 PR tree-optimization/58508
3753 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
3754 statement that contains data refs with zero-step.
3756 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3758 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
3759 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
3760 * gimple-low.c (gimple_check_call_arg,
3761 gimple_check_call_matching_types): Move to cgraph.c.
3762 * gimple-low.h: Remove prototype.
3763 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
3764 Relocate from gimple-low.c.
3765 * cgraph.h: Add prototype. Don't include basic-block.h.
3766 * gimplify.c: Add gimple-low to include list.
3767 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
3768 * tree-eh.c: Add gimple-low to include list.
3769 * tree-nested.c: Likewise.
3770 * cfgexpand.c: Add tree-ssa-address.h to include list.
3772 * gimple-fold.c: Likewise.
3773 * gimple-ssa-strength-reduction.c: Likewise.
3774 * trans-mem.c: Likewise.
3775 * tree-mudflap.c: Likewise.
3776 * tree-ssa-loop-ivopts.c: Likewise.
3777 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
3778 (degenerate_phi_result): Move to tree-phinodes.c.
3779 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
3780 * tree-ssa-threadedge.c: Likewise.
3781 * tree-vrp.c: Likewise.
3782 * tree-phinodes.c (degenerate_phi_result): Relocate here.
3783 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
3784 * tree-phinodes.h (degenerate_phi_result): Add prototype.
3785 * tree-ssa-copy.c: Include tree-ssa-dom.h.
3786 * tree-ssa-forwprop.c: Likewise.
3787 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
3788 pass_data_cleanup_cfg_post_optimizing,
3789 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
3790 * tree-optimize.c: Delete File.
3791 * graphite.c: Include tree-cfgcleanup.h.
3792 * passes.c: Likewise.
3793 * tree-cfg.c: Likewise.
3794 * tree-profile.c: Likewise.
3795 * tree-ssa-dse.c: Likewise.
3796 * tree-ssa-loop-ivcanon.c: Likewise.
3797 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
3798 * tree-outof-ssa.c: Include sbitmap.h.
3799 * tree-ssa-live.c: Likewise.
3800 * tree-ssa-propagate.c: Likewise.
3801 * tree-ssa-structalias.c: Likewise.
3802 * tree-stdarg.c: Likewise.
3803 * Makefile.in (OBJS): Delete tree-optimize.o.
3804 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
3805 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
3806 * varasm.c: Include basic-block.h.
3807 * cfgloop.h: Include function.h instead of basic-block.h
3808 (bb_loop_depth): Move to cfgloop.c.
3809 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
3811 2013-10-18 Teresa Johnson <tejohnson@google.com>
3813 * predict.c (probably_never_executed): Compare frequency-based
3814 count to number of training runs.
3815 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
3817 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3819 * config/arm/arm.c (cortexa9_extra_costs): New table.
3820 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
3822 2013-10-18 Jeff Law <law@redhat.com>
3824 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
3826 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
3827 (register_jump_thread): Add "registered_jump_thread" debug
3829 * dbgcnt.def (registered_jump_thread): New debug counter.
3831 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3833 * config/rs6000/rs6000.c: Include cgraph.h.
3835 2013-10-18 Teresa Johnson <tejohnson@google.com>
3837 * tree-ssa-tail-merge.c (replace_block_by): Update edge
3838 weights during merging.
3840 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3842 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
3843 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
3844 * tree-ssa-operands.h: Remove prototype.
3845 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
3846 * gimple.c (virtual_operand_p): Relocate from gimple.c.
3847 * gimple.h: Add prototype.
3848 * gimple-ssa.h: Include tree-ssa-operands.h.
3849 * tree-dump.c: Add tree-cfg.h to include list.
3850 * tree-ssa-alias.c: Add ipa-reference.h to include list.
3851 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
3852 * config/i386/i386.c: Don't include tree-flow.h.
3853 * config/rs6000/rs6000.c: Likewise.
3855 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3857 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3859 2013-10-18 Richard Biener <rguenther@suse.de>
3861 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3862 or TYPE_UNSIGNED of integral types.
3863 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3864 NULL_TREE for zero-precision integral types.
3866 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3868 * config/aarch64/arm_neon.h
3869 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3871 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3874 * opth-gen.awk: Define target_flags_explicit.
3876 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3878 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3879 fields to the reg_addr array that describes the valid addressing
3880 mode for any register, general purpose registers, floating point
3881 registers, and Altivec registers.
3882 (FIRST_RELOAD_REG_CLASS): Likewise.
3883 (LAST_RELOAD_REG_CLASS): Likewise.
3884 (struct reload_reg_map_type): Likewise.
3885 (reload_reg_map_type): Likewise.
3886 (RELOAD_REG_VALID): Likewise.
3887 (RELOAD_REG_MULTIPLE): Likewise.
3888 (RELOAD_REG_INDEXED): Likewise.
3889 (RELOAD_REG_OFFSET): Likewise.
3890 (RELOAD_REG_PRE_INCDEC): Likewise.
3891 (RELOAD_REG_PRE_MODIFY): Likewise.
3892 (reg_addr): Likewise.
3893 (mode_supports_pre_incdec_p): New helper functions to say whether
3894 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3895 (mode_supports_pre_modify_p): Likewise.
3896 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3897 print the valid address mode bits for each mode.
3898 (rs6000_debug_print_mode): Likewise.
3899 (rs6000_debug_reg_global): Likewise.
3900 (rs6000_setup_reg_addr_masks): New function to set up the address
3901 mask bits for each type.
3902 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3903 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3904 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3905 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3906 PRE_MODIFY are supported.
3907 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3908 registers, instead of {src,dest}_av_p.
3909 (rs6000_print_options_internal): Tweak the debug output slightly.
3911 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3913 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3916 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3918 * tree-flow.h (struct omp_region): Move to omp-low.c.
3919 Remove omp_ prototypes and variables.
3920 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3921 (copy_var_decl): Relocate prototype from tree-flow.h.
3922 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3923 * tree.h: Move prototype to omp-low.h.
3924 * omp-low.h: New File. Relocate prototypes here.
3925 * omp-low.c (struct omp_region): Make local here.
3926 (root_omp_region): Make static.
3927 (copy_var_decl) Move to gimple.c.
3928 (new_omp_region): Make static.
3929 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3930 * tree-cfg.c: Include omp-low.h.
3931 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3932 * gimplify.c: Include omp-low.h.
3933 * tree-parloops.c: Likewise.
3935 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3937 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3938 operands are in memory, prefer to force non-matched operand 1 to
3941 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3944 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3945 restrict TImode addresses to single indirect registers if both
3946 -mquad-memory and -mvsx-timode are used.
3947 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3948 we should emit load/store quad. Remove using %y for quad memory
3951 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3952 constraints to allow load/store quad on machines where TImode is
3953 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3954 for TImode to load integer constants.
3956 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3958 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3960 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3962 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3963 handling of STACK_REG.
3965 2013-10-17 Richard Biener <rguenther@suse.de>
3967 PR tree-optimization/58143
3968 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3970 (rewrite_to_defined_overflow): Likewise.
3971 (move_computations_dom_walker::before_dom): Rewrite stmts
3972 with undefined signed overflow that are not always executed
3973 into unsigned arithmetic.
3975 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3978 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3979 explicit isa flag to be an options variable, instead of using
3980 global_options_set. Remove define from rs6000.h.
3981 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3983 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3984 Initialize rs6000_isa_flags_explicit.
3985 (rs6000_function_specific_save): Add gcc_options* parameter, so
3986 that the powerpc builds after the 2013-10-15 changes.
3987 (rs6000_function_specific_restore): Likewise.
3989 2013-10-16 DJ Delorie <dj@redhat.com>
3991 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3992 op is a REG before checking REGNO.
3993 (rl78_alloc_physical_registers): Verify pattern is a SET before
3996 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3998 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
4000 (vec_unpacks_lo_v4sf): Likewise.
4001 (vec_unpacks_float_hi_v4si): Likewise.
4002 (vec_unpacks_float_lo_v4si): Likewise.
4003 (vec_unpacku_float_hi_v4si): Likewise.
4004 (vec_unpacku_float_lo_v4si): Likewise.
4006 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4008 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
4009 (vsx_concat_v2sf): Likewise.
4011 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
4013 * config/aarch64/aarch64.md
4014 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
4016 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
4018 PR tree-optimization/58697
4019 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
4020 estimated_loop_iterations_int.
4021 (max_stmt_executions_int): Call get_max_loop_iterations_int.
4022 (get_max_loop_iterations_int): New. HWINT version of
4023 get_max_loop_iterations.
4024 * cfgloop.h: Add prototypes.
4025 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
4026 * loop-unroll.c (decide_peel_once_rolling): Call
4027 get_estimated_loop_iterations_int.
4028 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
4029 * tree-ssa-loop-niter.h: Tweak prototypes.
4031 2013-10-16 David Malcolm <dmalcolm@redhat.com>
4033 * gengtype-parse.c (struct_field_seq): Ignore access-control
4034 keywords ("public:" etc).
4036 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
4038 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
4039 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
4041 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
4043 * config/arm/arm.opt (mlra): New option.
4044 * config/arm/arm.c (arm_lra_p): New function.
4045 (TARGET_LRA_P): Define.
4047 2013-10-16 Paulo Matos <pmatos@broadcom.com>
4049 * tree-core.h (tree_code_name): Remove.
4050 * tree.h (get_tree_code_name): New prototype.
4051 * tree.c (tree_code_name): Make static.
4052 (get_tree_code_name): New function.
4053 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
4054 tree_class_check_failed, tree_range_check_failed,
4055 tree_not_class_check_failed, omp_clause_check_failed,
4056 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
4057 wrapper get_tree_code_name instead of calling tree_code_name directly.
4058 * tree-vrp.c (dump_asserts_for): Likewise.
4059 * tree-dump.c (dequeue_and_dump): Likewise.
4060 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
4061 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
4062 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
4063 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
4064 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
4065 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
4066 dump_gimple_omp_for): Likewise.
4067 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
4068 * tree-ssa-pre.c (print_pre_expr): Likewise.
4069 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
4070 * print-tree.c (print_node_brief, print_node): Likewise.
4071 * gimple.c (gimple_check_failed): Likewise.
4072 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
4073 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
4074 * config/mep/mep.c (mep_validate_vliw): Likewise.
4075 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
4076 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
4078 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
4080 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
4083 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
4085 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
4086 (MULTILIB_MATCHES): Add multilib for -march=v8.
4088 2013-10-15 Sriraman Tallam <tmsriram@google.com>
4091 * optc-save-gen.awk: Add extra parameter to the save and restore
4093 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
4094 * tree.c (build_optimization_node): New parameter. Add extra parameter
4095 to call to cl_optimization_save.
4096 (build_target_option_node): New parameter. Add extra parameter
4097 to call to cl_target_option_save.
4098 * tree.h (build_optimization_node): New parameter.
4099 (build_target_option_node): New parameter.
4100 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
4101 build_optimization_node and build_target_option_node.
4102 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
4103 (handle_pragma_push_options): Ditto.
4104 * toplev.c (process_options): Ditto.
4105 * opts.c (init_options_struct): Check for opts_set non-null.
4106 * target.def (target_option.save): New parameter.
4107 (target_option.restore): New parameter.
4108 * tm.texi: Generate.
4109 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
4110 (ix86_pragma_target_parse): Ditto.
4111 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
4113 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
4114 to build_optimization_node and build_target_option_node.
4115 (rs6000_valid_attribute_p): Ditto.
4116 (rs6000_pragma_target_parse): Ditto.
4117 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
4119 * config/i386/i386.h:
4120 TARGET_64BIT_P: New Macro
4121 TARGET_MMX_P: New Macro.
4122 TARGET_3DNOW_P: New Macro.
4123 TARGET_3DNOW_A_P: New Macro.
4124 TARGET_SSE_P: New Macro.
4125 TARGET_SSE2_P: New Macro.
4126 TARGET_SSE3_P: New Macro.
4127 TARGET_SSSE3_P: New Macro.
4128 TARGET_SSE4_1_P: New Macro.
4129 TARGET_SSE4_2_P: New Macro.
4130 TARGET_AVX_P: New Macro.
4131 TARGET_AVX2_P: New Macro.
4132 TARGET_AVX512F_P: New Macro.
4133 TARGET_AVX512PF_P: New Macro.
4134 TARGET_AVX512ER_P: New Macro.
4135 TARGET_AVX512CD_P: New Macro.
4136 TARGET_FMA_P: New Macro.
4137 TARGET_SSE4A_P: New Macro.
4138 TARGET_FMA4_P: New Macro.
4139 TARGET_XOP_P: New Macro.
4140 TARGET_LWP_P: New Macro.
4141 TARGET_ABM_P: New Macro.
4142 TARGET_BMI_P: New Macro.
4143 TARGET_BMI2_P: New Macro.
4144 TARGET_LZCNT_P: New Macro.
4145 TARGET_TBM_P: New Macro.
4146 TARGET_POPCNT_P: New Macro.
4147 TARGET_SAHF_P: New Macro.
4148 TARGET_MOVBE_P: New Macro.
4149 TARGET_CRC32_P: New Macro.
4150 TARGET_AES_P: New Macro.
4151 TARGET_PCLMUL_P: New Macro.
4152 TARGET_CMPXCHG16B_P: New Macro.
4153 TARGET_FSGSBASE_P: New Macro.
4154 TARGET_RDRND_P: New Macro.
4155 TARGET_F16C_P: New Macro.
4156 TARGET_RTM_P: New Macro.
4157 TARGET_HLE_P: New Macro.
4158 TARGET_RDSEED_P: New Macro.
4159 TARGET_PRFCHW_P: New Macro.
4160 TARGET_ADX_P: New Macro.
4161 TARGET_FXSR_P: New Macro.
4162 TARGET_XSAVE_P: New Macro.
4163 TARGET_XSAVEOPT_P: New Macro.
4164 TARGET_LP64_P: New Macro.
4165 TARGET_X32_P: New Macro.
4166 TARGET_FPMATH_DEFAULT_P: New Macro.
4167 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
4168 * config/i386/i386.c (ix86_option_override_internal): New parameters.
4170 Change ix86_tune_string to access opts->x_ix86_tune_string.
4171 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
4172 Change ix86_arch_string to access opts->x_ix86_arch_string.
4173 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
4174 Change ix86_pmode to access opts->x_ix86_pmode.
4175 Change ix86_abi to access opts->x_ix86_abi.
4176 Change ix86_cmodel to access opts->x_ix86_cmodel.
4177 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
4178 Change ix86_isa_flags_explicit to access
4179 opts->x_ix86_isa_flags_explicit.
4180 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
4181 Change ix86_regparm to access opts->x_ix86_regparm.
4182 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
4183 Change ix86_preferred_stack_boundary_arg to access
4184 opts->x_ix86_preferred_stack_boundary_arg.
4185 Change ix86_force_align_arg_pointer to access
4186 opts->x_ix86_force_align_arg_pointer.
4187 Change ix86_incoming_stack_boundar_arg to access
4188 opts->x_ix86_incoming_stack_boundar_arg.
4189 Change ix86_fpmath to access opts->x_ix86_fpmath.
4190 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
4191 Change ix86_recip_name to access opts->x_ix86_recip_name.
4192 Change ix86_stack_protector_guard to access
4193 opts->x_ix86_stack_protector_guard.
4194 Change ix86_tune_memcpy_strategy to access
4195 opts->x_ix86_tune_memcpy_strategy.
4196 Change ix86_tune_memset_strategy to access
4197 opts->x_ix86_tune_memset_strategy.
4198 Change global_options to access opts.
4199 Change global_options_set to access opts_set.
4200 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
4201 Change TARGET_MMX to TARGET_MMX_P (opts->...).
4202 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
4203 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
4204 Change TARGET_SSE to TARGET_SSE_P (opts->...).
4205 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
4206 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
4207 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
4208 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
4209 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
4210 Change TARGET_AVX to TARGET_AVX_P (opts->...).
4211 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
4212 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
4213 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
4214 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
4215 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
4216 Change TARGET_FMA to TARGET_FMA_P (opts->...).
4217 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
4218 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
4219 Change TARGET_XOP to TARGET_XOP_P (opts->...).
4220 Change TARGET_LWP to TARGET_LWP_P (opts->...).
4221 Change TARGET_ABM to TARGET_ABM_P (opts->...).
4222 Change TARGET_BMI to TARGET_BMI_P (opts->...).
4223 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
4224 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
4225 Change TARGET_TBM to TARGET_TBM_P (opts->...).
4226 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
4227 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
4228 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
4229 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
4230 Change TARGET_AES to TARGET_AES_P (opts->...).
4231 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
4232 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
4233 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
4234 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
4235 Change TARGET_F16C to TARGET_F16C_P (opts->...).
4236 Change TARGET_RTM to TARGET_RTM_P (opts->...).
4237 Change TARGET_HLE to TARGET_HLE_P (opts->...).
4238 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
4239 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
4240 Change TARGET_ADX to TARGET_ADX_P (opts->...).
4241 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
4242 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
4243 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
4244 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
4245 Change TARGET_X32 to TARGET_X32_P (opts->...).
4246 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
4247 Change TARGET_FLOAT_RETURNS_IN_80387 to
4248 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
4249 (ix86_function_specific_save): New parameter. Use opts-> fields
4250 to replace global fields.
4251 (ix86_function_specific_restore): Ditto.
4252 (ix86_valid_target_attribute_inner_p): New parameters.
4254 Fix call to ix86_handle_option and set_option.
4255 (ix86_valid_target_attribute_tree): New parameters.
4256 Change global_options to access opts.
4257 Change global_options_set to access opts_set.
4258 Fix call to ix86_valid_target_attribute_inner_p.
4259 Change ix86_tune_string to access opts->x_ix86_tune_string.
4260 Change ix86_arch_string to access opts->x_ix86_arch_string.
4261 Change ix86_fpmath to access opts->x_ix86_fpmath
4262 Fix call to ix86_option_override_internal.
4263 Fix call to ix86_add_new_builtins.
4264 Fix calls to build_optimization_node and build_target_option_node.
4265 (ix86_valid_target_attribute_p): Remove access to global_options.
4266 Use new gcc_options structure func_options.
4267 Fix call to ix86_valid_target_attribute_tree.
4268 Fix call to build_optimization_node.
4269 (get_builtin_code_for_version): Fix call to
4270 ix86_valid_target_attribute_tree.
4272 2013-10-15 David Malcolm <dmalcolm@redhat.com>
4274 * Makefile.in (PICFLAG): New.
4275 (enable_host_shared): New.
4276 (INTERNAL_CFLAGS): Use PICFLAG.
4277 (LIBIBERTY): Use pic build of libiberty.a if configured with
4278 --enable-host-shared.
4279 * configure.ac: Add --enable-host-shared, setting up new
4281 * configure: Regenerate.
4282 * doc/install.texi (--enable-shared): Add note contrasting it with ...
4283 (--enable-host-shared): New option.
4285 2013-10-15 Richard Biener <rguenther@suse.de>
4287 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
4288 for built-in functions.
4290 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
4292 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
4293 (optimize_range_tests_1): New function,
4294 extracted from optimize_range_tests.
4295 (optimize_range_tests_xor): Similarly.
4296 (optimize_range_tests_diff): New function.
4297 (optimize_range_tests): Use optimize_range_tests_1.
4299 2013-10-15 Cong Hou <congh@google.com>
4301 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
4302 requirement of the reduction pattern so that one operand of the
4303 reduction operation can come from outside of the loop.
4305 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4307 * config/arm/neon-schedgen.ml: Remove.
4308 * config/arm/cortex-a9-neon.md: Remove comment regarding
4311 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4313 * config/arm/types: Remove old neon types.
4315 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4317 * config/arm/cortex-a7.md
4318 (cortex_a7_neon_type): New.
4319 (cortex_a7_neon_mul): Update for new types.
4320 (cortex_a7_neon_mla): Likewise.
4321 (cortex_a7_neon): Likewise.
4323 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4325 * config/arm/cortex-a15-neon.md
4326 (cortex_a15_neon_type): New,
4328 (cortex_a15_neon_int_1): Remove.
4329 (cortex_a15_neon_int_2): Likewise.
4330 (cortex_a15_neon_int_3): Likewise.
4331 (cortex_a15_neon_int_4): Likewise.
4332 (cortex_a15_neon_int_5): Likewise.
4333 (cortex_a15_neon_vqneg_vqabs): Likewise.
4334 (cortex_a15_neon_vmov): Likewise.
4335 (cortex_a15_neon_vaba): Likewise.
4336 (cortex_a15_neon_vaba_qqq): Likewise.
4337 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4338 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4339 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4341 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4342 (cortex_a15_neon_mla_qqq_8_16): Likewise.
4343 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
4344 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4345 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4346 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
4347 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4348 (cortex_a15_neon_shift_1): Likewise.
4349 (cortex_a15_neon_shift_2): Likewise.
4350 (cortex_a15_neon_shift_3): Likewise.
4351 (cortex_a15_neon_vshl_ddd): Likewise.
4352 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4353 (cortex_a15_neon_vsra_vrsra): Likewise.
4354 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4355 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4356 (cortex_a15_neon_bp_3cycle): Likewise.
4357 (cortex_a15_neon_ldm_2): Likewise.
4358 (cortex_a15_neon_stm_2): Likewise.
4359 (cortex_a15_neon_mcr): Likewise.
4360 (cortex_a15_neon_mrc): Likewise.
4361 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
4362 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
4363 (cortex_a15_neon_fp_vmul_ddd): Likewise.
4364 (cortex_a15_neon_fp_vmul_qqd): Likewise.
4365 (cortex_a15_neon_fp_vmla_ddd): Likewise.
4366 (cortex_a15_neon_fp_vmla_qqq): Likewise.
4367 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4368 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4369 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4370 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4371 (cortex_a15_neon_bp_simple): Likewise.
4372 (cortex_a15_neon_bp_2cycle): Likewise.
4373 (cortex_a15_neon_bp_3cycle): Likewise.
4374 (cortex_a15_neon_vld1_1_2_regs): Likewise.
4375 (cortex_a15_neon_vld1_3_4_regs): Likewise.
4376 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4377 (cortex_a15_neon_vld2_4_regs): Likewise.
4378 (cortex_a15_neon_vld3_vld4): Likewise.
4379 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4380 (cortex_a15_neon_vst1_3_4_regs): Likewise.
4381 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
4382 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
4383 (cortex_a15_neon_vst3_vst4): Rename to...
4384 (cortex_a15_neon_vst4): This, update for new attributes.
4385 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
4386 (cortex_a15_neon_vld3_vld4_lane): Likewise.
4387 (cortex_a15_neon_vst1_vst2_lane): Likewise.
4388 (cortex_a15_neon_vst3_vst4_lane): Likewise.
4389 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
4390 (cortex_a15_neon_ldm_2): Likewise.
4391 (cortex_a15_neon_stm_2): Likewise.
4392 (cortex_a15_neon_mcr): Likewise.
4393 (cortex_a15_neon_mcr_2_mcrr): Likewise.
4394 (cortex_a15_neon_mrc): Likewise.
4395 (cortex_a15_neon_mrrc): Likewise.
4397 (cortex_a15_neon_abd): New.
4398 (cortex_a15_neon_abd_q): Likewise.
4399 (cortex_a15_neon_aba): Likewise.
4400 (cortex_a15_neon_aba_q): Likewise.
4401 (cortex_a15_neon_acc): Likewise.
4402 (cortex_a15_neon_acc_q): Likewise.
4403 (cortex_a15_neon_arith_basic): Likewise.
4404 (cortex_a15_neon_arith_complex): Likewise.
4405 (cortex_a15_neon_multiply): Likewise.
4406 (cortex_a15_neon_multiply_q): Likewise.
4407 (cortex_a15_neon_mla): Likewise.
4408 (cortex_a15_neon_mla_q): Likewise.
4409 (cortex_a15_neon_sat_mla_long): Likewise.
4410 (cortex_a15_neon_shift_acc): Likewise.
4411 (cortex_a15_neon_shift_imm_basic): Likewise.
4412 (cortex_a15_neon_shift_imm_complex): Likewise.
4413 (cortex_a15_neon_shift_reg_basic): Likewise.
4414 (cortex_a15_neon_shift_reg_basic_q): Likewise.
4415 (cortex_a15_neon_shift_reg_complex): Likewise.
4416 (cortex_a15_neon_shift_reg_complex_q): Likewise.
4417 (cortex_a15_neon_fp_negabs): Likewise
4418 (cortex_a15_neon_fp_arith): Likewise
4419 (cortex_a15_neon_fp_arith_q): Likewise
4420 (cortex_a15_neon_fp_cvt_int): Likewise
4421 (cortex_a15_neon_fp_cvt_int_q): Likewise
4422 (cortex_a15_neon_fp_cvt_16): Likewise
4423 (cortex_a15_neon_fp_mul): Likewise
4424 (cortex_a15_neon_fp_mul_q): Likewise
4425 (cortex_a15_neon_fp_mla): Likewise
4426 (cortex_a15_neon_fp_mla_q): Likewise
4427 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
4428 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
4429 (cortex_a15_neon_bitops): Likewise.
4430 (cortex_a15_neon_bitops_q): Likewise.
4431 (cortex_a15_neon_from_gp): Likewise.
4432 (cortex_a15_neon_from_gp_q): Likewise.
4433 (cortex_a15_neon_tbl3_tbl4): Likewise.
4434 (cortex_a15_neon_zip_q): Likewise.
4435 (cortex_a15_neon_to_gp): Likewise.
4436 (cortex_a15_neon_load_a): Likewise.
4437 (cortex_a15_neon_load_b): Likewise.
4438 (cortex_a15_neon_load_c): Likewise.
4439 (cortex_a15_neon_load_d): Likewise.
4440 (cortex_a15_neon_load_e): Likewise.
4441 (cortex_a15_neon_load_f): Likewise.
4442 (cortex_a15_neon_store_a): Likewise.
4443 (cortex_a15_neon_store_b): Likewise.
4444 (cortex_a15_neon_store_c): Likewise.
4445 (cortex_a15_neon_store_d): Likewise.
4446 (cortex_a15_neon_store_e): Likewise.
4447 (cortex_a15_neon_store_f): Likewise.
4448 (cortex_a15_neon_store_g): Likewise.
4449 (cortex_a15_neon_store_h): Likewise.
4450 (cortex_a15_vfp_to_from_gp): Likewise.
4452 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4454 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
4456 (cortex_a9_neon_vshl_ddd): Remove.
4457 (cortex_a9_neon_vst3_vst4): Likewise.
4458 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4460 (cortex_a9_neon_bit_ops_q): New.
4462 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
4463 (cortex_a9_neon_int_2): Likewise.
4464 (cortex_a9_neon_int_3): Likewise.
4465 (cortex_a9_neon_int_4): Likewise.
4466 (cortex_a9_neon_int_5): Likewise.
4467 (cortex_a9_neon_vqneg_vqabs): Likewise.
4468 (cortex_a9_neon_vmov): Likewise.
4469 (cortex_a9_neon_vaba): Likewise.
4470 (cortex_a9_neon_vaba_qqq): Likewise.
4471 (cortex_a9_neon_shift_1): Likewise.
4472 (cortex_a9_neon_shift_2): Likewise.
4473 (cortex_a9_neon_shift_3): Likewise.
4474 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4475 (cortex_a9_neon_vsra_vrsra): Likewise.
4476 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4477 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4478 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4480 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4481 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4482 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4484 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4485 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4486 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4487 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4488 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4489 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4490 (cortex_a9_neon_fp_vsum): Likewise.
4491 (cortex_a9_neon_fp_vmul_ddd): Likewise.
4492 (cortex_a9_neon_fp_vmul_qqd): Likewise.
4493 (cortex_a9_neon_fp_vmla_ddd): Likewise.
4494 (cortex_a9_neon_fp_vmla_qqq): Likewise.
4495 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
4496 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
4497 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4498 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4499 (cortex_a9_neon_bp_simple): Likewise.
4500 (cortex_a9_neon_bp_2cycle): Likewise.
4501 (cortex_a9_neon_bp_3cycle): Likewise.
4502 (cortex_a9_neon_ldr): Likewise.
4503 (cortex_a9_neon_str): Likewise.
4504 (cortex_a9_neon_vld1_1_2_regs): Likewise.
4505 (cortex_a9_neon_vld1_3_4_regs): Likewise.
4506 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4507 (cortex_a9_neon_vld2_4_regs): Likewise.
4508 (cortex_a9_neon_vld3_vld4): Likewise.
4509 (cortex_a9_neon_vld1_vld2_lane): Likewise.
4510 (cortex_a9_neon_vld3_vld4_lane): Likewise.
4511 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4512 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4513 (cortex_a9_neon_vst1_3_4_regs): Likewise.
4514 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
4515 (cortex_a9_neon_vst1_vst2_lane): Likewise.
4516 (cortex_a9_neon_vst3_vst4_lane): Likewise.
4517 (cortex_a9_neon_mcr): Likewise.
4518 (cortex_a9_neon_mcr_2_mcrr): Likewise.
4519 (cortex_a9_neon_mrc): Likewise.
4520 (cortex_a9_neon_mrrc): Likewise.
4522 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4524 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
4526 (cortex_a8_neon_vshl_ddd): Remove.
4527 (cortex_a8_neon_vst3_vst4): Likewise.
4528 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4530 (cortex_a8_neon_bit_ops_q): New.
4532 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
4533 (cortex_a8_neon_int_2): Likewise..
4534 (cortex_a8_neon_int_3): Likewise.
4535 (cortex_a8_neon_int_5): Likewise.
4536 (cortex_a8_neon_vqneg_vqabs): Likewise.
4537 (cortex_a8_neon_int_4): Likewise.
4538 (cortex_a8_neon_vaba): Likewise.
4539 (cortex_a8_neon_vaba_qqq): Likewise.
4540 (cortex_a8_neon_shift_1): Likewise.
4541 (cortex_a8_neon_shift_2): Likewise.
4542 (cortex_a8_neon_shift_3): Likewise.
4543 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4544 (cortex_a8_neon_vsra_vrsra): Likewise.
4545 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4546 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4547 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4549 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4550 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4551 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4553 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4554 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4555 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4556 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4557 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4558 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4559 (cortex_a8_neon_fp_vsum): Likewise.
4560 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4561 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4562 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4563 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4564 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4565 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4566 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4567 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4568 (cortex_a8_neon_bp_simple): Likewise.
4569 (cortex_a8_neon_bp_2cycle): Likewise.
4570 (cortex_a8_neon_bp_3cycle): Likewise.
4571 (cortex_a8_neon_ldr): Likewise.
4572 (cortex_a8_neon_str): Likewise.
4573 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4574 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4575 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4576 (cortex_a8_neon_vld2_4_regs): Likewise.
4577 (cortex_a8_neon_vld3_vld4): Likewise.
4578 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4579 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4580 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4581 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4582 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4583 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4584 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4585 (cortex_a8_neon_mcr): Likewise.
4586 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4587 (cortex_a8_neon_mrc): Likewise.
4588 (cortex_a8_neon_mrrc): Likewise.
4590 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4592 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
4594 * config/aarch64/aarch64-simd.md (neon_type): Remove.
4595 (aarch64_simd_dup<mode>): Add "type" attribute.
4596 (aarch64_dup_lane<mode>): Likewise.
4597 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
4598 (*aarch64_simd_mov<mode>): Likewise.
4599 (aarch64_simd_mov_from_<mode>low): Likewise.
4600 (aarch64_simd_mov_from_<mode>high): Likewise.
4601 (orn<mode>3): Likewise.
4602 (bic<mode>3): Likewise.
4603 (add<mode>3): Likewise.
4604 (sub<mode>3): Likewise.
4605 (mul<mode>3): Likewise.
4606 (*aarch64_mul3_elt<mode>): Likewise.
4607 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4608 (*aarch64_mul3_elt_to_128df): Likewise.
4609 (*aarch64_mul3_elt_to_64v2df): Likewise.
4610 (neg<mode>2): Likewise.
4611 (abs<mode>2): Likewise.
4612 (abd<mode>_3): Likewise.
4613 (aba<mode>_3): Likewise.
4614 (fabd<mode>_3): Likewise.
4615 (*fabd_scalar<mode>3): Likewise.
4616 (and<mode>3): Likewise.
4617 (ior<mode>3): Likewise.
4618 (xor<mode>3): Likewise.
4619 (one_cmpl<mode>2): Likewise.
4620 (aarch64_simd_vec_set<mode>): Likewise.
4621 (aarch64_simd_lshr<mode>): Likewise.
4622 (aarch64_simd_ashr<mode>): Likewise.
4623 (aarch64_simd_imm_shl<mode>): Likewise.
4624 (aarch64_simd_reg_sshl<mode): Likewise.
4625 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
4626 (aarch64_simd_reg_shl<mode>_signed): Likewise.
4627 (aarch64_simd_vec_setv2di): Likewise.
4628 (aarch64_simd_vec_set<mode>): Likewise.
4629 (aarch64_mla<mode>): Likewise.
4630 (*aarch64_mla_elt<mode>): Likewise.
4631 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4632 (aarch64_mls<mode>): Likewise.
4633 (*aarch64_mls_elt<mode>): Likewise.
4634 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4635 (<su><maxmin><mode>3): Likewise.
4636 (move_lo_quad_<mode>): Likewise.
4637 (aarch64_simd_move_hi_quad_<mode>): Likewise.
4638 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
4639 (vec_pack_trunc_<mode>): Likewise.
4640 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
4641 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
4642 (*aarch64_<su>mlal_lo<mode>): Likewise.
4643 (*aarch64_<su>mlal_hi<mode>): Likewise.
4644 (*aarch64_<su>mlsl_lo<mode>): Likewise.
4645 (*aarch64_<su>mlsl_hi<mode>): Likewise.
4646 (*aarch64_<su>mlal<mode>): Likewise.
4647 (*aarch64_<su>mlsl<mode>): Likewise.
4648 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
4649 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
4650 (add<mode>3): Likewise.
4651 (sub<mode>3): Likewise.
4652 (mul<mode>3): Likewise.
4653 (div<mode>3): Likewise.
4654 (neg<mode>2): Likewise.
4655 (abs<mode>2): Likewise.
4656 (fma<mode>4): Likewise.
4657 (*aarch64_fma4_elt<mode>): Likewise.
4658 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4659 (*aarch64_fma4_elt_to_128df): Likewise.
4660 (*aarch64_fma4_elt_to_64v2df): Likewise.
4661 (fnma<mode>4): Likewise.
4662 (*aarch64_fnma4_elt<mode>): Likewise.
4663 (*aarch64_fnma4_elt_<vswap_width_name><mode>
4664 (*aarch64_fnma4_elt_to_128df): Likewise.
4665 (*aarch64_fnma4_elt_to_64v2df): Likewise.
4666 (<frint_pattern><mode>2): Likewise.
4667 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
4668 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
4669 (vec_unpacks_lo_v4sf): Likewise.
4670 (aarch64_float_extend_lo_v2df): Likewise.
4671 (vec_unpacks_hi_v4sf): Likewise.
4672 (aarch64_float_truncate_lo_v2sf): Likewise.
4673 (aarch64_float_truncate_hi_v4sf): Likewise.
4674 (aarch64_vmls<mode>): Likewise.
4675 (<su><maxmin><mode>3): Likewise.
4676 (<maxmin_uns><mode>3): Likewise.
4677 (reduc_<sur>plus_<mode>): Likewise.
4678 (reduc_<sur>plus_v2di): Likewise.
4679 (reduc_<sur>plus_v2si): Likewise.
4680 (reduc_<sur>plus_<mode>): Likewise.
4681 (aarch64_addpv4sf): Likewise.
4682 (clz<mode>2): Likewise.
4683 (reduc_<maxmin_uns>_<mode>): Likewise.
4684 (reduc_<maxmin_uns>_v2di): Likewise.
4685 (reduc_<maxmin_uns>_v2si): Likewise.
4686 (reduc_<maxmin_uns>_<mode>): Likewise.
4687 (reduc_<maxmin_uns>_v4sf): Likewise.
4688 (aarch64_simd_bsl<mode>_internal): Likewise.
4689 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
4690 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
4691 (aarch64_get_lane<mode>): Likewise.
4692 (*aarch64_combinez<mode>): Likewise.
4693 (aarch64_combine<mode>): Likewise.
4694 (aarch64_simd_combine<mode>): Likewise.
4695 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
4696 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
4697 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
4698 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
4699 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
4700 (aarch64_<sur>h<addsub><mode>): Likewise.
4701 (aarch64_<sur><addsub>hn<mode>): Likewise.
4702 (aarch64_<sur><addsub>hn2<mode>): Likewise.
4703 (aarch64_pmul<mode>): Likewise.
4704 (aarch64_<su_optab><optab><mode>): Likewise.
4705 (aarch64_<sur>qadd<mode>): Likewise.
4706 (aarch64_sqmovun<mode>): Likewise.
4707 (aarch64_<sur>qmovn<mode>): Likewise.
4708 (aarch64_s<optab><mode>): Likewise.
4709 (aarch64_sq<r>dmulh<mode>): Likewise.
4710 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4711 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
4712 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4713 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
4714 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4715 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4716 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
4717 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
4718 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
4719 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
4720 (aarch64_sqdmull<mode>): Likewise.
4721 (aarch64_sqdmull_lane<mode>_internal): Likewise.
4722 (aarch64_sqdmull_n<mode>): Likewise.
4723 (aarch64_sqdmull2<mode>_internal): Likewise.
4724 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
4725 (aarch64_sqdmull2_n<mode>_internal): Likewise.
4726 (aarch64_<sur>shl<mode>): Likewise.
4727 (aarch64_<sur>q<r>shl<mode>
4728 (aarch64_<sur>shll_n<mode>): Likewise.
4729 (aarch64_<sur>shll2_n<mode>): Likewise.
4730 (aarch64_<sur>shr_n<mode>): Likewise.
4731 (aarch64_<sur>sra_n<mode>): Likewise.
4732 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
4733 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
4734 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
4735 (aarch64_cm<optab><mode>): Likewise.
4736 (aarch64_cm<optab>di): Likewise.
4737 (aarch64_cm<optab><mode>): Likewise.
4738 (aarch64_cm<optab>di): Likewise.
4739 (aarch64_cmtst<mode>): Likewise.
4740 (aarch64_cmtstdi): Likewise.
4741 (aarch64_cm<optab><mode>): Likewise.
4742 (*aarch64_fac<optab><mode>): Likewise.
4743 (aarch64_addp<mode>): Likewise.
4744 (aarch64_addpdi): Likewise.
4745 (sqrt<mode>2): Likewise.
4746 (vec_load_lanesoi<mode>): Likewise.
4747 (vec_store_lanesoi<mode>): Likewise.
4748 (vec_load_lanesci<mode>): Likewise.
4749 (vec_store_lanesci<mode>): Likewise.
4750 (vec_load_lanesxi<mode>): Likewise.
4751 (vec_store_lanesxi<mode>): Likewise.
4752 (*aarch64_mov<mode>): Likewise.
4753 (aarch64_ld2<mode>_dreg): Likewise.
4754 (aarch64_ld2<mode>_dreg): Likewise.
4755 (aarch64_ld3<mode>_dreg): Likewise.
4756 (aarch64_ld3<mode>_dreg): Likewise.
4757 (aarch64_ld4<mode>_dreg): Likewise.
4758 (aarch64_ld4<mode>_dreg): Likewise.
4759 (aarch64_tbl1<mode>): Likewise.
4760 (aarch64_tbl2v16qi): Likewise.
4761 (aarch64_combinev16qi): Likewise.
4762 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
4763 (aarch64_st2<mode>_dreg): Likewise.
4764 (aarch64_st2<mode>_dreg): Likewise.
4765 (aarch64_st3<mode>_dreg): Likewise.
4766 (aarch64_st3<mode>_dreg): Likewise.
4767 (aarch64_st4<mode>_dreg): Likewise.
4768 (aarch64_st4<mode>_dreg): Likewise.
4769 (*aarch64_simd_ld1r<mode>): Likewise.
4770 (aarch64_frecpe<mode>): Likewise.
4771 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
4772 (aarch64_frecps<mode>): Likewise.
4774 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4776 * config/arm/iterators.md (V_elem_ch): New.
4778 (VQH_type): Likewise.
4779 * config/arm/arm.md (is_neon_type): New.
4780 (conds): Use is_neon_type.
4781 (anddi3_insn): Update type attribute.
4782 (xordi3_insn): Likewise.
4783 (one_cmpldi2): Likewise.
4784 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
4785 * config/arm/neon.md (neon_mov): Update type attribute.
4786 (*movmisalign<mode>_neon_store): Likewise.
4787 (*movmisalign<mode>_neon_load): Likewise.
4788 (vec_set<mode>_internal): Likewise.
4789 (vec_set<mode>_internal): Likewise.
4790 (vec_setv2di_internal): Likewise.
4791 (vec_extract<mode>): Likewise.
4792 (vec_extract<mode>): Likewise.
4793 (vec_extractv2di): Likewise.
4794 (*add<mode>3_neon): Likewise.
4795 (adddi3_neon): Likewise.
4796 (*sub<mode>3_neon): Likewise.
4797 (subdi3_neon): Likewise.
4798 (fma<VCVTF:mode>4): Likewise.
4799 (fma<VCVTF:mode>4_intrinsic): Likewise.
4800 (*fmsub<VCVTF:mode>4): Likewise.
4801 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4802 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4803 (ior<mode>3): Likewise.
4804 (and<mode>3): Likewise.
4805 (orn<mode>3_neon): Likewise.
4806 (orndi3_neon): Likewise.
4807 (bic<mode>3_neon): Likewise.
4808 (bicdi3_neon): Likewise.
4809 (xor<mode>3): Likewise.
4810 (one_cmpl<mode>2): Likewise.
4811 (abs<mode>2): Likewise.
4812 (neg<mode>2): Likewise.
4813 (negdi2_neon): Likewise.
4814 (*umin<mode>3_neon): Likewise.
4815 (*umax<mode>3_neon): Likewise.
4816 (*smin<mode>3_neon): Likewise.
4817 (*smax<mode>3_neon): Likewise.
4818 (vashl<mode>3): Likewise.
4819 (vashr<mode>3_imm): Likewise.
4820 (vlshr<mode>3_imm): Likewise.
4821 (ashl<mode>3_signed): Likewise.
4822 (ashl<mode>3_unsigned): Likewise.
4823 (neon_load_count): Likewise.
4824 (ashldi3_neon_noclobber): Likewise.
4825 (ashldi3_neon): Likewise.
4826 (signed_shift_di3_neon): Likewise.
4827 (unsigned_shift_di3_neon): Likewise.
4828 (ashrdi3_neon_imm_noclobber): Likewise.
4829 (lshrdi3_neon_imm_noclobber): Likewise.
4830 (<shift>di3_neon): Likewise.
4831 (widen_ssum<mode>3): Likewise.
4832 (widen_usum<mode>3): Likewise.
4833 (quad_halves_<code>v4si): Likewise.
4834 (quad_halves_<code>v4sf): Likewise.
4835 (quad_halves_<code>v8hi): Likewise.
4836 (quad_halves_<code>v16qi): Likewise.
4837 (reduc_splus_v2di): Likewise.
4838 (neon_vpadd_internal<mode>): Likewise.
4839 (neon_vpsmin<mode>): Likewise.
4840 (neon_vpsmax<mode>): Likewise.
4841 (neon_vpumin<mode>): Likewise.
4842 (neon_vpumax<mode>): Likewise.
4843 (*ss_add<mode>_neon): Likewise.
4844 (*us_add<mode>_neon): Likewise.
4845 (*ss_sub<mode>_neon): Likewise.
4846 (*us_sub<mode>_neon): Likewise.
4847 (neon_vadd<mode>_unspec): Likewise.
4848 (neon_vaddl<mode>): Likewise.
4849 (neon_vaddw<mode>): Likewise.
4850 (neon_vhadd<mode>): Likewise.
4851 (neon_vqadd<mode>): Likewise.
4852 (neon_vaddhn<mode>): Likewise.
4853 (neon_vmul<mode>): Likewise.
4854 (neon_vfms<VCVTF:mode>): Likewise.
4855 (neon_vmlal<mode>): Likewise.
4856 (neon_vmls<mode>): Likewise.
4857 (neon_vmlsl<mode>): Likewise.
4858 (neon_vqdmulh<mode>): Likewise.
4859 (neon_vqdmlal<mode>): Likewise.
4860 (neon_vqdmlsl<mode>): Likewise.
4861 (neon_vmull<mode>): Likewise.
4862 (neon_vqdmull<mode>): Likewise.
4863 (neon_vsub<mode>_unspec): Likewise.
4864 (neon_vsubl<mode>): Likewise.
4865 (neon_vsubw<mode>): Likewise.
4866 (neon_vqsub<mode>): Likewise.
4867 (neon_vhsub<mode>): Likewise.
4868 (neon_vsubhn<mode>): Likewise.
4869 (neon_vceq<mode>): Likewise.
4870 (neon_vcge<mode>): Likewise.
4871 (neon_vcgeu<mode>): Likewise.
4872 (neon_vcgt<mode>): Likewise.
4873 (neon_vcgtu<mode>): Likewise.
4874 (neon_vcle<mode>): Likewise.
4875 (neon_vclt<mode>): Likewise.
4876 (neon_vcage<mode>): Likewise.
4877 (neon_vcagt<mode>): Likewise.
4878 (neon_vtst<mode>): Likewise.
4879 (neon_vabd<mode>): Likewise.
4880 (neon_vabdl<mode>): Likewise.
4881 (neon_vaba<mode>): Likewise.
4882 (neon_vabal<mode>): Likewise.
4883 (neon_vmax<mode>): Likewise.
4884 (neon_vmin<mode>): Likewise.
4885 (neon_vpaddl<mode>): Likewise.
4886 (neon_vpadal<mode>): Likewise.
4887 (neon_vpmax<mode>): Likewise.
4888 (neon_vpmin<mode>): Likewise.
4889 (neon_vrecps<mode>): Likewise.
4890 (neon_vrsqrts<mode>): Likewise.
4891 (neon_vqabs<mode>): Likewise.
4892 (neon_vqneg<mode>): Likewise.
4893 (neon_vcls<mode>): Likewise.
4894 (clz<mode>2): Likewise.
4895 (popcount<mode>2): Likewise.
4896 (neon_vrecpe<mode>): Likewise.
4897 (neon_vrsqrte<mode>): Likewise.
4898 (neon_vget_lane<mode>_sext_internal): Likewise.
4899 (neon_vget_lane<mode>_zext_internal): Likewise.
4900 (neon_vdup_n<mode>): Likewise.
4901 (neon_vdup_n<mode>): Likewise.
4902 (neon_vdup_nv2di): Likewise.
4903 (neon_vdup_lane<mode>_interal): Likewise.
4904 (*neon_vswp<mode>): Likewise.
4905 (neon_vcombine<mode>): Likewise.
4906 (float<mode><V_cvtto>2): Likewise.
4907 (floatuns<mode><V_cvtto>2): Likewise.
4908 (fix_trunc<mode><V_cvtto>2): Likewise.
4909 (fixuns_trunc<mode><V_cvtto>2
4910 (neon_vcvt<mode>): Likewise.
4911 (neon_vcvt<mode>): Likewise.
4912 (neon_vcvtv4sfv4hf): Likewise.
4913 (neon_vcvtv4hfv4sf): Likewise.
4914 (neon_vcvt_n<mode>): Likewise.
4915 (neon_vcvt_n<mode>): Likewise.
4916 (neon_vmovn<mode>): Likewise.
4917 (neon_vqmovn<mode>): Likewise.
4918 (neon_vqmovun<mode>): Likewise.
4919 (neon_vmovl<mode>): Likewise.
4920 (neon_vmul_lane<mode>): Likewise.
4921 (neon_vmul_lane<mode>): Likewise.
4922 (neon_vmull_lane<mode>): Likewise.
4923 (neon_vqdmull_lane<mode>): Likewise.
4924 (neon_vqdmulh_lane<mode>): Likewise.
4925 (neon_vqdmulh_lane<mode>): Likewise.
4926 (neon_vmla_lane<mode>): Likewise.
4927 (neon_vmla_lane<mode>): Likewise.
4928 (neon_vmlal_lane<mode>): Likewise.
4929 (neon_vqdmlal_lane<mode>): Likewise.
4930 (neon_vmls_lane<mode>): Likewise.
4931 (neon_vmls_lane<mode>): Likewise.
4932 (neon_vmlsl_lane<mode>): Likewise.
4933 (neon_vqdmlsl_lane<mode>): Likewise.
4934 (neon_vext<mode>): Likewise.
4935 (neon_vrev64<mode>): Likewise.
4936 (neon_vrev32<mode>): Likewise.
4937 (neon_vrev16<mode>): Likewise.
4938 (neon_vbsl<mode>_internal): Likewise.
4939 (neon_vshl<mode>): Likewise.
4940 (neon_vqshl<mode>): Likewise.
4941 (neon_vshr_n<mode>): Likewise.
4942 (neon_vshrn_n<mode>): Likewise.
4943 (neon_vqshrn_n<mode>): Likewise.
4944 (neon_vqshrun_n<mode>): Likewise.
4945 (neon_vshl_n<mode>): Likewise.
4946 (neon_vqshl_n<mode>): Likewise.
4947 (neon_vqshlu_n<mode>): Likewise.
4948 (neon_vshll_n<mode>): Likewise.
4949 (neon_vsra_n<mode>): Likewise.
4950 (neon_vsri_n<mode>): Likewise.
4951 (neon_vsli_n<mode>): Likewise.
4952 (neon_vtbl1v8qi): Likewise.
4953 (neon_vtbl2v8qi): Likewise.
4954 (neon_vtbl3v8qi): Likewise.
4955 (neon_vtbl4v8qi): Likewise.
4956 (neon_vtbl1v16qi): Likewise.
4957 (neon_vtbl2v16qi): Likewise.
4958 (neon_vcombinev16qi): Likewise.
4959 (neon_vtbx1v8qi): Likewise.
4960 (neon_vtbx2v8qi): Likewise.
4961 (neon_vtbx3v8qi): Likewise.
4962 (neon_vtbx4v8qi): Likewise.
4963 (*neon_vtrn<mode>_insn): Likewise.
4964 (*neon_vzip<mode>_insn): Likewise.
4965 (*neon_vuzp<mode>_insn): Likewise.
4966 (neon_vld1<mode>): Likewise.
4967 (neon_vld1_lane<mode>): Likewise.
4968 (neon_vld1_lane<mode>): Likewise.
4969 (neon_vld1_dup<mode>): Likewise.
4970 (neon_vld1_dup<mode>): Likewise.
4971 (neon_vld1_dupv2di): Likewise.
4972 (neon_vst1<mode>): Likewise.
4973 (neon_vst1_lane<mode>): Likewise.
4974 (neon_vst1_lane<mode>): Likewise.
4975 (neon_vld2<mode>): Likewise.
4976 (neon_vld2<mode>): Likewise.
4977 (neon_vld2_lane<mode>): Likewise.
4978 (neon_vld2_lane<mode>): Likewise.
4979 (neon_vld2_dup<mode>): Likewise.
4980 (neon_vst2<mode>): Likewise.
4981 (neon_vst2<mode>): Likewise.
4982 (neon_vst2_lane<mode>): Likewise.
4983 (neon_vst2_lane<mode>): Likewise.
4984 (neon_vld3<mode>): Likewise.
4985 (neon_vld3qa<mode>): Likewise.
4986 (neon_vld3qb<mode>): Likewise.
4987 (neon_vld3_lane<mode>): Likewise.
4988 (neon_vld3_lane<mode>): Likewise.
4989 (neon_vld3_dup<mode>): Likewise.
4990 (neon_vst3<mode>): Likewise.
4991 (neon_vst3qa<mode>): Likewise.
4992 (neon_vst3qb<mode>): Likewise.
4993 (neon_vst3_lane<mode>): Likewise.
4994 (neon_vst3_lane<mode>): Likewise.
4995 (neon_vld4<mode>): Likewise.
4996 (neon_vld4qa<mode>): Likewise.
4997 (neon_vld4qb<mode>): Likewise.
4998 (neon_vld4_lane<mode>): Likewise.
4999 (neon_vld4_lane<mode>): Likewise.
5000 (neon_vld4_dup<mode>): Likewise.
5001 (neon_vst4<mode>): Likewise.
5002 (neon_vst4qa<mode>): Likewise.
5003 (neon_vst4qb<mode>): Likewise.
5004 (neon_vst4_lane<mode>): Likewise.
5005 (neon_vst4_lane<mode>): Likewise.
5006 (neon_vec_unpack<US>_lo_<mode>): Likewise.
5007 (neon_vec_unpack<US>_hi_<mode>): Likewise.
5008 (neon_vec_<US>mult_lo_<mode>): Likewise.
5009 (neon_vec_<US>mult_hi_<mode>): Likewise.
5010 (neon_vec_<US>shiftl_<mode>): Likewise.
5011 (neon_unpack<US>_<mode>): Likewise.
5012 (neon_vec_<US>mult_<mode>): Likewise.
5013 (vec_pack_trunc_<mode>): Likewise.
5014 (neon_vec_pack_trunc_<mode>): Likewise.
5015 (neon_vabd<mode>_2): Likewise.
5016 (neon_vabd<mode>_3): Likewise.
5018 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5020 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
5021 (load_pair): Update type attribute.
5022 (store_pair): Update type attribute.
5023 * config/aarch64/iterators.md (q): New.
5025 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5027 * config/arm/types.md: Add new types for Neon insns.
5029 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5030 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5031 Sergey Lega <sergey.s.lega@intel.com>
5032 Anna Tikhonova <anna.tikhonova@intel.com>
5033 Ilya Tocar <ilya.tocar@intel.com>
5034 Andrey Turetskiy <andrey.turetskiy@intel.com>
5035 Ilya Verbin <ilya.verbin@intel.com>
5036 Kirill Yukhin <kirill.yukhin@intel.com>
5037 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5039 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
5040 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
5041 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
5043 (srcp14<mode>): Ditto.
5044 (rsqrt14<mode>): Ditto.
5045 (rsqrt14<mode>): Ditto.
5046 (avx512f_vmscalef<mode>): Ditto.
5047 (avx512f_scalef<mode>): Ditto.
5048 (avx512f_getexp<mode>): Ditto.
5049 (avx512f_sgetexp<mode>): Ditto.
5050 (avx512f_fixupimm<mode>): Ditto.
5051 (avx512f_sfixupimm<mode>): Ditto.
5052 (avx512f_rndscale<mode>): Ditto.
5053 (*avx512er_exp2<mode>): Ditto.
5054 (*avx512er_rcp28<mode>): Ditto.
5055 (avx512er_rsqrt28<mode>): Ditto.
5056 (avx512f_getmant<mode>): Ditto.
5057 (avx512f_getmant<mode>): Ditto.
5058 (avx512f_rndscale<mode>): Fix formatting.
5060 2013-10-15 Martin Jambor <mjambor@suse.cz>
5062 * ipa-utils.h (ipa_edge_within_scc): Declare.
5063 * ipa-cp.c (edge_within_scc): Moved...
5064 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
5066 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5067 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5068 Sergey Lega <sergey.s.lega@intel.com>
5069 Anna Tikhonova <anna.tikhonova@intel.com>
5070 Ilya Tocar <ilya.tocar@intel.com>
5071 Andrey Turetskiy <andrey.turetskiy@intel.com>
5072 Ilya Verbin <ilya.verbin@intel.com>
5073 Kirill Yukhin <kirill.yukhin@intel.com>
5074 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5076 * config/i386/predicates.md (const_8_to_15_operand): New.
5077 (const_16_to_31_operand): Ditto.
5078 * config/i386/sse.md (V8FI): New.
5080 (reduc_splus_v8df): Ditto.
5081 (reduc_splus_v16sf): Ditto.
5082 (avx512f_vextract<shuffletype>32x4_1): Ditto.
5083 (vec_extract_hi_<mode>): Ditto.
5084 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
5085 (vec_set_lo_<mode>): Ditto.
5086 (vec_set_hi_<mode>): Ditto.
5087 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
5088 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
5089 (avx512f_pshufd_1): Ditto.
5090 (avx512f_broadcast<mode>): Ditto.
5091 (avx512f_broadcast<mode>): Ditto.
5092 (define_split): Split vec_extract_lo into move.
5093 (ssequartermode): Ditto.
5094 (ssedoublemode): Extened with wider modes.
5095 (vec_extract_lo_<mode>): Ditto.
5097 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5098 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5099 Sergey Lega <sergey.s.lega@intel.com>
5100 Anna Tikhonova <anna.tikhonova@intel.com>
5101 Ilya Tocar <ilya.tocar@intel.com>
5102 Andrey Turetskiy <andrey.turetskiy@intel.com>
5103 Ilya Verbin <ilya.verbin@intel.com>
5104 Kirill Yukhin <kirill.yukhin@intel.com>
5105 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5107 * config/i386/predicates.md (register_or_constm1_operand): New.
5108 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
5109 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
5110 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
5111 UNSPEC_SCATTER_PREFETCH
5113 (avx512f_ucmp<mode>3): Ditto.
5114 (avx512f_vternlog<mode>): Ditto.
5115 (avx512f_align<mode>): Ditto.
5116 (<shift_insn><mode>3): Ditto.
5117 (avx512f_<rotate>v<mode>): Ditto.
5118 (avx512f_<rotate><mode>): Ditto.
5119 (avx512f_eq<mode>3): Ditto.
5120 (avx512f_eq<mode>3_1): Ditto.
5121 (avx512f_gt<mode>3): Ditto.
5122 (avx512f_testm<mode>3): Ditto.
5123 (avx512f_testnm<mode>3): Ditto.
5124 (avx512pf_gatherpf<mode>): Ditto.
5125 (*avx512pf_gatherpf<mode>_mask): Ditto.
5126 (*avx512pf_gatherpf<mode>): Ditto.
5127 (avx512pf_scatterpf<mode>): Ditto.
5128 (*avx512pf_scatterpf<mode>_mask): Ditto.
5129 (*avx512pf_scatterpf<mode>): Ditto.
5130 (avx512f_vec_dup_gpr<mode>): Ditto.
5131 (clz<mode>2): Ditto.
5132 (conflict<mode>): Ditto.
5133 (REDUC_SMINMAX_MODE): Extened with wider modes.
5134 (reduc_<code>_<mode>): Ditto.
5135 (vlshr<mode>3): Ditto.
5136 (vashl<mode>3): Ditto.
5138 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5139 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5140 Sergey Lega <sergey.s.lega@intel.com>
5141 Anna Tikhonova <anna.tikhonova@intel.com>
5142 Ilya Tocar <ilya.tocar@intel.com>
5143 Andrey Turetskiy <andrey.turetskiy@intel.com>
5144 Ilya Verbin <ilya.verbin@intel.com>
5145 Kirill Yukhin <kirill.yukhin@intel.com>
5146 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5148 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
5151 (avx512fmaskmode): Ditto.
5152 (bcstscalarsuff): Ditto.
5153 (avx512f_blendm<mode>): Ditto.
5154 (cmp_imm_predicate): Ditto.
5155 (avx512f_cmp<mode>3): Ditto.
5156 (avx512f_vec_dup<mode>): Ditto.
5157 (avx512f_vec_dup_mem<mode>): Ditto.
5158 (avx512f_vpermi2var<mode>3): Ditto.
5159 (avx512f_vpermt2var<mode>3): Ditto.
5160 (vec_init<mode>): Ditto.
5161 (avx512f_gathersi<mode>): Ditto.
5162 (*avx512f_gathersi<mode>): Ditto.
5163 (*avx512f_gathersi<mode>_2): Ditto.
5164 (avx512f_gatherdi<mode>): Ditto.
5165 (*avx512f_gatherdi<mode>): Ditto.
5166 (*avx512f_gatherdi<mode>_2): Ditto.
5167 (avx512f_scattersi<mode>): Ditto.
5168 (*avx512f_scattersi<mode>): Ditto.
5169 (avx512f_scatterdi<mode>): Ditto.
5170 (*avx512f_scatterdi<mode>): Ditto.
5171 (sseintprefix): Extened with wider modes.
5172 (VEC_GATHER_IDXSI): Ditto.
5173 (VEC_GATHER_IDXDI): Ditto.
5174 (VEC_GATHER_SRCDI): Ditto.
5176 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
5177 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5179 * config/arm/t-aprofile: New file.
5180 * config.gcc: Handle --with-multilib-list option.
5182 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
5184 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
5187 2013-10-15 Richard Biener <rguenther@suse.de>
5189 * tree-loop-distribution.c (build_empty_rdg): Inline into
5191 (rdg_flag_vertex): Inline into single user.
5192 (rdg_flag_vertex_and_dependent): Likewise.
5193 (build_rdg_partition_for_vertex): Remove processed bitmap.
5194 (rdg_build_partitions): Simplify.
5196 2013-10-15 Richard Biener <rguenther@suse.de>
5198 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
5199 Restructure forwarding through conversions and copies to
5200 avoid performing copy-propagation the wrong way. Adjust
5201 recursion invocations.
5202 (forward_propagate_addr_expr): Add argument stating if we
5203 are recursing from a single-use.
5204 (ssa_forward_propagate_and_combine): Adjust.
5206 2013-10-14 David Malcolm <dmalcolm@redhat.com>
5208 * dumpfile.h (gcc::dump_manager): New class, to hold state
5209 relating to dumpfile management.
5210 (get_dump_file_name): Remove in favor of method of dump_manager.
5211 (dump_initialized_p): Likewise.
5212 (dump_start): Likewise.
5213 (dump_finish): Likewise.
5214 (dump_switch_p): Likewise.
5215 (dump_register): Likewise.
5216 (get_dump_file_info): Likewise.
5217 * context.c (gcc::context::context): Construct the dump_manager
5219 * context.h (gcc::context::get_dumps): New.
5220 (gcc::context::m_dumps): New.
5221 * coverage.c (coverage_init): Port to dump_manager API.
5222 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
5223 (extra_dump_files_in_use): Likewise.
5224 (extra_dump_files_alloced): Likewise.
5225 (gcc::dump_manager::dump_manager): New.
5226 (dump_register): Convert to...
5227 (gcc::dump_manager::dump_register): ...method, replacing
5228 function-static next_dump with m_next_dump field.
5229 (get_dump_file_info): Convert to...
5230 (gcc::dump_manager::get_dump_file_info): ...method.
5231 (get_dump_file_name): Convert to...
5232 (gcc::dump_manager::get_dump_file_name): ...method.
5233 (dump_start): Convert to...
5234 (gcc::dump_manager::dump_start): ...method.
5235 (dump_finish): Convert to...
5236 (gcc::dump_manager::dump_finish): ...method.
5237 (dump_begin): Replace body with...
5238 (gcc::dump_manager::dump_begin): ...new method.
5239 (dump_phase_enabled_p): Convert to...
5240 (gcc::dump_manager::dump_phase_enabled_p): ...method.
5241 (dump_phase_enabled_p): Convert to...
5242 (gcc::dump_manager::dump_phase_enabled_p): ...method.
5243 (dump_initialized_p): Convert to...
5244 (gcc::dump_manager::dump_initialized_p): ...method.
5245 (dump_flag_name): Replace body with...
5246 (gcc::dump_manager::dump_flag_name): ...new method.
5247 (dump_enable_all): Convert to...
5248 (gcc::dump_manager::dump_enable_all): ...new method.
5249 (opt_info_enable_passes): Convert to...
5250 (gcc::dump_manager::opt_info_enable_passes): ...new method.
5251 (dump_switch_p_1): Convert to...
5252 (gcc::dump_manager::dump_switch_p_1): ...new method.
5253 (dump_switch_p): Convert to...
5254 (gcc::dump_manager::dump_switch_p): ...new method.
5255 (opt_info_switch_p): Port to dump_manager API.
5256 (enable_rtl_dump_file): Likewise.
5257 * opts-global.c (handle_common_deferred_options): Port to new
5259 * passes.c (pass_manager::finish_optimization_passes): Likewise.
5260 (pass_manager::register_one_dump_file): Likewise.
5261 (pass_manager::register_pass): Likewise.
5262 (pass_init_dump_file): Likewise.
5263 (pass_fini_dump_file): Likewise.
5264 * statistics.c (statistics_early_init): Likewise.
5266 2013-10-14 Richard Biener <rguenther@suse.de>
5268 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
5269 iterative_hash_canonical_type, gimple_canonical_type_hash,
5270 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
5271 gimple_register_canonical_type, print_gimple_types_stats,
5272 free_gimple_type_tables): Move to lto/lto.c
5273 (gt-gimple.h): Do not include.
5274 * gimple.h (gimple_register_canonical_type,
5275 print_gimple_types_stats, free_gimple_type_tables): Remove.
5276 * Makefile.in (GTFILES): Remove gimple.c.
5278 2013-10-14 Travis Snoozy <quandary@remstate.com>
5281 * config/msp430/msp430.c (msp430_option_override): Correct thinko
5282 scanning for msp430x targets.
5284 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
5287 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
5288 (registers_ok_for_ldd_peep): Move around.
5289 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
5290 * config/sparc/sparc.md (widening peepholes): Use it.
5292 2013-10-14 Richard Biener <rguenther@suse.de>
5296 * gimple.c (iterative_hash_canonical_type): Make sure to
5297 record the hash into the correct hashtable slot.
5299 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
5301 PR rtl-optimization/58662
5302 * combine.c (try_combine): Take into account death nodes on I2 when
5303 splitting a PARALLEL of two independent SETs. Fix dump message.
5305 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
5308 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
5309 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
5310 * config/sh/t-sh (sh_treg_combine.o): New entry.
5311 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
5312 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
5313 (register_sh_passes): New function. Register sh_treg_combine pass.
5314 (sh_option_override): Invoke it.
5315 (sh_canonicalize_comparison): Handle op0_preserve_value.
5316 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
5317 opportunities. Canonicalize branch condition.
5318 (nott): Allow only if pseudos can be created for non-SH2A.
5320 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
5323 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
5324 (ix86_expand_movmem): Replace copy_addr_to_reg with
5325 ix86_copy_addr_to_reg.
5326 (ix86_expand_setmem): Likewise.
5328 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
5330 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
5331 provided by ix86_fp_compare_mode instead of CCFPUmode.
5333 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
5335 * config/aarch64/arm_neon.h
5336 (vtbx<1,3>_<psu>8): Fix register constriants.
5338 2013-10-11 Jeff Law <law@redhat.com>
5340 PR tree-optimization/58640
5341 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
5342 threading paths that cross over two loop entry points.
5344 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5346 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
5347 handle vector float as well.
5348 (*vsx_le_perm_load_v4si): Likewise.
5349 (*vsx_le_perm_store_v2di): Likewise.
5350 (*vsx_le_perm_store_v4si): Likewise.
5352 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5354 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
5355 directly to circumvent subtract from splat{31} workaround.
5356 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
5358 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
5359 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
5360 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
5361 (altivec_vperm_<mode>): Convert to define_insn_and_split to
5362 separate big and little endian logic.
5363 (*altivec_vperm_<mode>_internal): New define_insn.
5364 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
5365 separate big and little endian logic.
5366 (*altivec_vperm_<mode>_uns_internal): New define_insn.
5367 (vec_permv16qi): Add little endian logic.
5369 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
5371 * doc/extend.texi (returns_nonnull): Remove arguments.
5373 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5374 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5375 Sergey Lega <sergey.s.lega@intel.com>
5376 Anna Tikhonova <anna.tikhonova@intel.com>
5377 Ilya Tocar <ilya.tocar@intel.com>
5378 Andrey Turetskiy <andrey.turetskiy@intel.com>
5379 Ilya Verbin <ilya.verbin@intel.com>
5380 Kirill Yukhin <kirill.yukhin@intel.com>
5381 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5383 * config/i386/sse.md (VI48F_256_512): New.
5384 (avx2_permvar<mode>): Change to ...
5385 (<avx2_avx512f>_permvar<mode>): This.
5387 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5388 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5389 Sergey Lega <sergey.s.lega@intel.com>
5390 Anna Tikhonova <anna.tikhonova@intel.com>
5391 Ilya Tocar <ilya.tocar@intel.com>
5392 Andrey Turetskiy <andrey.turetskiy@intel.com>
5393 Ilya Verbin <ilya.verbin@intel.com>
5394 Kirill Yukhin <kirill.yukhin@intel.com>
5395 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5397 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
5398 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
5399 * config/i386/sse.md (VI4_AVX): New.
5400 (sf2simodelower): Ditto.
5401 (sse2_cvtps2dq): Change to ...
5402 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
5404 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5405 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5406 Sergey Lega <sergey.s.lega@intel.com>
5407 Anna Tikhonova <anna.tikhonova@intel.com>
5408 Ilya Tocar <ilya.tocar@intel.com>
5409 Andrey Turetskiy <andrey.turetskiy@intel.com>
5410 Ilya Verbin <ilya.verbin@intel.com>
5411 Kirill Yukhin <kirill.yukhin@intel.com>
5412 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5414 * config/i386/sse.md (V_512): New.
5416 (vcond<V_512:mode><VF_512:mode>): Ditto.
5417 (vcond<V_512:mode><VI_512:mode>): Ditto.
5418 (vcondu<V_512:mode><VI_512:mode>): Ditto.
5420 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5421 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5422 Sergey Lega <sergey.s.lega@intel.com>
5423 Anna Tikhonova <anna.tikhonova@intel.com>
5424 Ilya Tocar <ilya.tocar@intel.com>
5425 Andrey Turetskiy <andrey.turetskiy@intel.com>
5426 Ilya Verbin <ilya.verbin@intel.com>
5427 Kirill Yukhin <kirill.yukhin@intel.com>
5428 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5430 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
5431 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
5433 (fma<mode>4): Removed condition.
5434 (fms<mode>4): Ditto.
5435 (fnma<mode>4): Ditto.
5436 (fnms<mode>4): Ditto.
5437 (fma4i_fmadd_<mode>): Ditto.
5438 (*fma_fmadd_<mode>): Ditto.
5439 (*fma_fmsub_<mode>): Ditto.
5440 (*fma_fnmadd_<mode>): Ditto.
5441 (*fma_fnmsub_<mode>): Ditto.
5442 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
5443 (*fma_fmaddsub_<mode>): Ditto.
5444 (*fma_fmsubadd_<mode>): Ditto.
5445 (*fmai_fmadd_<mode>): Ditto.
5446 (*fmai_fmsub_<mode>): Ditto.
5447 (*fmai_fnmadd_<mode>): Ditto.
5448 (*fmai_fnmsub_<mode>): Ditto.
5450 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5451 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5452 Sergey Lega <sergey.s.lega@intel.com>
5453 Anna Tikhonova <anna.tikhonova@intel.com>
5454 Ilya Tocar <ilya.tocar@intel.com>
5455 Andrey Turetskiy <andrey.turetskiy@intel.com>
5456 Ilya Verbin <ilya.verbin@intel.com>
5457 Kirill Yukhin <kirill.yukhin@intel.com>
5458 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5460 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
5461 (VI124_256): Changed to ...
5462 (VI124_256_48_512): This.
5463 (ssepackmode): Extended with wider modes.
5464 (<code><mode>3): Changed iterator.
5465 (*avx2_<code><mode>3): Ditto.
5466 (vec_pack_trunc_<mode>): Ditto.
5468 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5469 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5470 Sergey Lega <sergey.s.lega@intel.com>
5471 Anna Tikhonova <anna.tikhonova@intel.com>
5472 Ilya Tocar <ilya.tocar@intel.com>
5473 Andrey Turetskiy <andrey.turetskiy@intel.com>
5474 Ilya Verbin <ilya.verbin@intel.com>
5475 Kirill Yukhin <kirill.yukhin@intel.com>
5476 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5478 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
5479 (VI8F_256_512): Ditto.
5480 (abs<mode>2): Changed iterator.
5481 (avx2_perm<mode>): Changed to ...
5482 (<avx2_avx512f>_perm<mode>): This.
5483 (avx2_perm<mode>_1): Changed to ...
5484 (<avx2_avx512f>_perm<mode>_1): This.
5486 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5487 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5488 Sergey Lega <sergey.s.lega@intel.com>
5489 Anna Tikhonova <anna.tikhonova@intel.com>
5490 Ilya Tocar <ilya.tocar@intel.com>
5491 Andrey Turetskiy <andrey.turetskiy@intel.com>
5492 Ilya Verbin <ilya.verbin@intel.com>
5493 Kirill Yukhin <kirill.yukhin@intel.com>
5494 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5496 * config/i386/sse.md (VI48_AVX512F): New.
5497 (VI48_AVX2): Changed to ...
5498 (VI48_AVX2_48_AVX512F): This.
5499 (avx2_ashrv<mode>): Changed to ...
5500 (<avx2_avx512f>_ashrv<mode>): This.
5501 (avx2_<shift_insn>v<mode>): Changed to ...
5502 (<avx2_avx512f>_<shift_insn>v<mode>): This.
5504 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5505 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5506 Sergey Lega <sergey.s.lega@intel.com>
5507 Anna Tikhonova <anna.tikhonova@intel.com>
5508 Ilya Tocar <ilya.tocar@intel.com>
5509 Andrey Turetskiy <andrey.turetskiy@intel.com>
5510 Ilya Verbin <ilya.verbin@intel.com>
5511 Kirill Yukhin <kirill.yukhin@intel.com>
5512 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5514 * config/i386/sse.md (VI4_AVX512F): New.
5515 (VI8_AVX2_AVX512F): Ditto.
5516 (mul<mode>3): Extended with wider modes.
5517 (*<sse4_1_avx2>_mul<mode>3): Ditto.
5518 (mul<mode>3): Ditto.
5519 (vec_widen_<s>mult_odd_<mode>): Ditto.
5521 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5522 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5523 Sergey Lega <sergey.s.lega@intel.com>
5524 Anna Tikhonova <anna.tikhonova@intel.com>
5525 Ilya Tocar <ilya.tocar@intel.com>
5526 Andrey Turetskiy <andrey.turetskiy@intel.com>
5527 Ilya Verbin <ilya.verbin@intel.com>
5528 Kirill Yukhin <kirill.yukhin@intel.com>
5529 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5531 * config/i386/sse.md (VI2_AVX512F): New.
5532 (VI124_AVX512F): Ditto.
5533 (sseunpackmode): Extended with wider modes.
5534 (sseunpackfltmode): Ditto.
5535 (vec_unpacks_float_hi_<mode>): Ditto.
5536 (vec_unpacks_float_lo_<mode>): Ditto.
5537 (vec_unpacku_float_hi_<mode>): Ditto.
5538 (vec_unpacku_float_lo_<mode>): Ditto.
5539 (vec_unpacks_lo_<mode>): Ditto.
5540 (vec_unpacks_hi_<mode>): Ditto.
5541 (vec_unpacku_lo_<mode>): Ditto.
5542 (vec_unpacku_hi_<mode>): Ditto.
5544 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5545 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5546 Sergey Lega <sergey.s.lega@intel.com>
5547 Anna Tikhonova <anna.tikhonova@intel.com>
5548 Ilya Tocar <ilya.tocar@intel.com>
5549 Andrey Turetskiy <andrey.turetskiy@intel.com>
5550 Ilya Verbin <ilya.verbin@intel.com>
5551 Kirill Yukhin <kirill.yukhin@intel.com>
5552 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5554 * config/i386/i386.md (multdiv): New.
5555 (multdiv_mnemonic): Ditto.
5556 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
5557 (<sse>_vm<multdiv_mnemonic><mode>3): This.
5558 (<sse>_vmdiv<mode>3): Removed.
5560 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5561 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5562 Sergey Lega <sergey.s.lega@intel.com>
5563 Anna Tikhonova <anna.tikhonova@intel.com>
5564 Ilya Tocar <ilya.tocar@intel.com>
5565 Andrey Turetskiy <andrey.turetskiy@intel.com>
5566 Ilya Verbin <ilya.verbin@intel.com>
5567 Kirill Yukhin <kirill.yukhin@intel.com>
5568 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5570 * config/i386/sse.md (V): Extended with wider modes.
5572 (ssehalfvecmode): Ditto.
5574 (ssepackfltmode): Ditto.
5575 (avx_vec_concat<mode>): Ditto.
5576 (V_256_512): New iterator.
5577 (VF2_512_256): Ditto.
5578 (si2dfmode): New attribute.
5579 (si2dfmodelower): Ditto.
5581 (concat_tg_mode): Ditto.
5582 (floatv4siv4df2): Changed to ...
5583 (float<si2dfmodelower><mode>2): This.
5584 (avx_cvtps2pd256): Changed to ...
5585 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
5586 (vec_pack_trunc_v4df): Changed to ...
5587 (vec_pack_trunc_<mode>): This.
5588 (avx_vpermil<mode>): Changed to ...
5589 (<sse2_avx_avx512f>_vpermil<mode>): This.
5590 (<fixsuffix>fix_truncv8dfv8si2): New.
5591 (vec_pack_sfix_trunc_v8df): Ditto.
5592 (avx512f_rndscale<mode>): Ditto.
5593 (avx512f_roundpd512): Ditto.
5594 (vec_pack_ufix_trunc_<mode>): Updated iterator.
5596 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5597 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5598 Sergey Lega <sergey.s.lega@intel.com>
5599 Anna Tikhonova <anna.tikhonova@intel.com>
5600 Ilya Tocar <ilya.tocar@intel.com>
5601 Andrey Turetskiy <andrey.turetskiy@intel.com>
5602 Ilya Verbin <ilya.verbin@intel.com>
5603 Kirill Yukhin <kirill.yukhin@intel.com>
5604 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5606 * config/i386/i386.md (any_fix): New iterator.
5607 (fixsuffix): New attribute.
5608 * config/i386/sse.md (VF1): Extened with wider modes.
5612 (sseintvecmodelower): Ditto.
5613 (ssescalarmode): Ditto.
5614 (ssescalarnum): Ditto.
5616 (ssexmmmode): Ditto.
5617 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
5618 (<sse>_rcp<mode>2): Change iterator.
5619 (rsqrt<mode>2): Ditto.
5620 (<sse>_rsqrt<mode>2): Ditto.
5621 (avx2_vec_dup<mode>): Ditto.
5622 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
5623 (round<mode>2_sfix): Ditto.
5624 (avx2_pbroadcast<mode>): Ditto.
5625 (*andnot<mode>3): Handle XI mode.
5626 (*<code><mode>3): Ditto.
5627 (AVXTOSSEMODE): Removed.
5628 (avx_vpermil<mode>): Changed to ...
5629 (<sse2_avx_avx512f>_vpermil<mode>): This.
5631 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5632 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5633 Sergey Lega <sergey.s.lega@intel.com>
5634 Anna Tikhonova <anna.tikhonova@intel.com>
5635 Ilya Tocar <ilya.tocar@intel.com>
5636 Andrey Turetskiy <andrey.turetskiy@intel.com>
5637 Ilya Verbin <ilya.verbin@intel.com>
5638 Kirill Yukhin <kirill.yukhin@intel.com>
5639 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5641 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
5642 (<sse>_comi): Ditto.
5643 (<sse>_ucomi): Ditto.
5644 (sse_cvtss2siq_2): Ditto.
5645 (sse2_cvtsd2si): Ditto.
5646 (sse2_cvtsd2siq): Ditto.
5647 (sse2_cvttsd2si): Ditto.
5648 (sse2_cvttsd2siq): Ditto.
5649 (<shift_insn><mode>3): Ditto.
5650 (sse2_cvtsi2sdq): Update constraint and prefix.
5651 (sse_cvtsi2ss): Update prefix.
5652 (sse_cvtsi2ssq): Ditto.
5654 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5656 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
5657 ignore internal calls.
5659 2013-10-11 Richard Biener <rguenther@suse.de>
5661 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
5662 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
5663 TREE_OVERFLOW is set.
5665 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
5667 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
5669 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
5670 * gimple.h: Likewise.
5672 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
5674 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
5677 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
5679 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
5680 * configure: Regenerate.
5682 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5684 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
5685 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
5686 formatting fixes, use pp_colon instead of pp_character (..., ':'),
5687 similarly pp_right_paren.
5688 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
5689 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
5690 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
5691 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
5693 (omp_declare_simd_clauses_equal,
5694 omp_remove_redundant_declare_simd_attrs): New functions.
5695 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
5696 (walk_tree_1): Handle new OpenMP 4.0 clauses.
5697 * tree.h (OMP_LOOP_CHECK): Define.
5698 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
5699 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
5700 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
5701 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
5702 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
5703 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
5704 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
5705 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
5706 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
5707 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
5708 OMP_CLAUSE_SIMDLEN_EXPR): Define.
5709 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
5710 (omp_remove_redundant_declare_simd_attrs): New prototype.
5711 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
5712 GIMPLE_OMP_TEAMS): New codes.
5713 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
5714 * omp-low.c (struct omp_context): Add cancel_label and cancellable
5716 (target_nesting_level): New variable.
5717 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
5718 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
5719 for collapse > 1 static schedule unless ordered.
5720 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
5721 (determine_parallel_type): Adjust get_ws_args_for caller.
5722 (install_var_field): Handle mask & 4 for double indirection.
5723 (scan_sharing_clauses): Ignore shared clause on teams construct.
5724 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
5725 (create_omp_child_function): If inside target or declare target
5726 constructs, set "omp declare target" attribute on the child function.
5727 (find_combined_for): New function.
5728 (scan_omp_parallel): Handle combined loops.
5729 (scan_omp_target, scan_omp_teams): New functions.
5730 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
5731 restrictions and set ctx->cancellable for cancellable constructs.
5732 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
5733 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
5734 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
5735 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
5736 (omp_clause_aligned_alignment): New function.
5737 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
5738 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
5739 on teams constructs. Handle user defined reductions and new
5741 (lower_reduction_clauses): Don't set placeholder to address of ref
5742 if it has already the right type.
5743 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
5744 (expand_parallel_call): Use the new non-_start suffixed builtins,
5745 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
5746 and GOMP_parallel_end after the call.
5747 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
5748 (expand_omp_for_init_counts): Handle combined loops.
5749 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
5751 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
5752 end of cancellable loops.
5753 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
5754 Likewise. Handle collapse > 1 loops.
5755 (expand_omp_simd): Handle combined loops.
5756 (expand_omp_for): Add inner_stmt argument, adjust callers of
5757 expand_omp_for* functions, use expand_omp_for_static*chunk even
5758 for collapse > 1 unless ordered.
5759 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
5760 of cancellable sections.
5761 (expand_omp_single): Remove need_barrier variable, just rely on
5762 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
5763 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
5764 (expand_omp_atomic_load, expand_omp_atomic_store,
5765 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
5766 (expand_omp_target): New function.
5767 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
5768 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
5769 (build_omp_regions_1): Immediately close region for
5770 GF_OMP_TARGET_KIND_UPDATE.
5771 (maybe_add_implicit_barrier_cancel): New function.
5772 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
5774 (lower_omp_single): Likewise. Add clobber after the barrier.
5775 (lower_omp_taskgroup): New function.
5776 (lower_omp_for): Handle combined loops. Adjust
5777 lower_rec_input_clauses caller. Handle cancellation.
5778 (lower_depend_clauses): New function.
5779 (lower_omp_taskreg): Lower depend clauses. Adjust
5780 lower_rec_input_clauses caller. Add clobber after the call. Handle
5782 (lower_omp_target, lower_omp_teams): New functions.
5783 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
5784 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
5785 and GOMP_cancellation_point calls.
5786 (lower_omp): Fold stmts inside of target region.
5787 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
5788 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5789 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
5790 (BT_FN_VOID_OMPFN_PTR_UINT,
5791 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
5792 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
5793 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
5794 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
5795 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
5796 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
5797 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
5798 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
5799 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
5800 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
5801 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
5802 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
5803 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
5804 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
5805 BUILT_IN_GOMP_PARALLEL_END.
5806 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
5807 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5808 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
5809 GF_OMP_FOR_KIND_DISTRIBUTE.
5810 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
5811 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
5812 (dump_gimple_omp_return): Print lhs if it has any.
5813 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
5814 gimple_omp_atomic_seq_cst_p.
5815 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
5816 and GIMPLE_OMP_TEAMS.
5817 * langhooks.c (lhd_omp_mappable_type): New function.
5818 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
5819 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
5821 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
5822 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
5823 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
5824 (struct gimplify_omp_ctx): Add combined_loop field.
5825 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
5826 on stmts inside of target region.
5827 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
5828 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
5829 ORT_TARGET and ORT_TARGET_DATA.
5830 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
5832 (omp_notice_threadprivate_variable): Complain about threadprivate
5833 variables in target region.
5834 (omp_notice_variable): Complain about vars with non-mappable type
5835 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
5836 (omp_check_private): Ignore ORT_TARGET* regions.
5837 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
5838 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
5839 (find_combined_omp_for): New function.
5840 (gimplify_omp_for): Handle gimplification of combined loops.
5841 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
5843 (gimplify_omp_target_update): New function.
5844 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
5845 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
5846 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
5847 (gimplify_body): If fndecl has "omp declare target" attribute, add
5848 implicit ORT_TARGET context around it.
5849 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
5850 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
5851 * tree-nested.c (convert_nonlocal_reference_stmt,
5852 convert_local_reference_stmt, convert_gimple_call): Handle
5853 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5854 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
5855 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5856 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5857 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5858 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5859 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5860 BUILT_IN_GOMP_LOOP_END_CANCEL,
5861 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5862 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5863 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5864 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5865 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5866 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5867 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5868 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5869 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5870 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5871 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5872 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5873 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5874 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5875 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5876 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5877 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5878 gimple_build_omp_teams): New functions.
5879 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5880 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5881 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5882 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5883 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5884 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5885 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5886 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5887 GF_OMP_ATOMIC_SEQ_CST): New.
5888 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5889 gimple_build_omp_teams): New prototypes.
5890 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5891 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5892 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5893 GIMPLE_OMP_SINGLE as end of range.
5894 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5895 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5896 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5897 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5898 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5899 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5900 gimple_omp_target_kind, gimple_omp_target_set_kind,
5901 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5902 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5903 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5904 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5905 gimple_omp_teams_set_clauses): New inlines.
5906 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5907 and GIMPLE_OMP_TASKGROUP.
5908 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5909 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5910 enum omp_clause_proc_bind_kind): New.
5911 (union omp_clause_subcode): Add depend_kind, map_kind and
5912 proc_bind_kind fields.
5913 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5914 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5915 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5916 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5917 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5919 2013-10-10 Teresa Johnson <tejohnson@google.com>
5921 * predict.c (tree_estimate_probability): Add new parameter
5922 for estimate_bb_frequencies.
5923 (estimate_bb_frequencies): Add new parameter to force estimation.
5924 (rebuild_frequencies): When max frequency in function is small,
5925 recompute counts from frequencies.
5926 * predict.h (estimate_bb_frequencies): New parameter.
5928 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5930 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5931 optimizations are disabled.
5933 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5935 * coverage.c (coverage_finish): Fix leak of da_file_name.
5937 2013-10-10 Jan Hubicka <jh@suse.cz>
5939 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5940 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5943 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5946 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5947 if any labels are in FALLTHRU_BB, use a special label emitted
5948 immediately after the asm goto insn rather than label_rtx
5950 (expand_asm_stmt): Adjust caller.
5951 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5952 edge if the last insn in predecessor is a jump with single successor,
5953 but it isn't simplejump_p.
5955 2013-10-10 Richard Biener <rguenther@suse.de>
5957 PR tree-optimization/58656
5958 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5960 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5962 * gimplify.c: Include expr.h and tm_p.h for targets with special
5963 va-arg padding requirements.
5965 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5967 * tree-flow.h: Move some prototypes to gimple.h.
5968 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5969 * gimple.h: Relocate some prototypes from tree-flow.h
5970 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5972 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5973 (build_va_arg_indirect_ref): Relocate and make static.
5974 (std_gimplify_va_arg_expr): Relocate here.
5975 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5976 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5978 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5980 * doc/md.texi: Document the mnemonic attribute.
5982 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5985 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5986 (set_attr_alternative x ...) when searching for user defined
5989 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5991 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5993 2013-10-09 Easwaran Raman <eraman@google.com>
5995 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5996 * cfgexpand.c (defer_stack_allocation): ...use here
5997 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5999 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
6001 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
6002 (operand_equal_for_value_replacement): New function, extracted from
6003 value_replacement and enhanced to catch more cases.
6004 (value_replacement): Use operand_equal_for_value_replacement.
6006 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6008 * loop-doloop.c (doloop_modify, doloop_optimize): Use
6009 get_max_loop_iterations.
6011 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6013 * config/arm/aarch-common.c (arm_early_load_addr_dep):
6014 Place comment above function.
6016 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6018 * tree-flow.h: Remove all remaining prototypes, enums and structs that
6019 are not related to tree-cfg.c.
6020 * tree-ssa-address.h: New file. Relocate prototypes.
6021 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
6022 (addr_for_mem_ref): New. Combine call to get_address_description and
6023 return addr_for_mem_ref.
6024 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
6025 * tree-ssa-live.h: Adjust prototypes.
6026 * passes.c: Include tree-ssa-live.h.
6027 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
6028 * graphite.c (graphite_transform_loops): Make static.
6029 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
6030 make_pass_graphite, pass_data_graphite_transforms,
6031 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
6032 * ipa-pure-const.c (warn_function_noreturn): Make static.
6033 (execute_warn_function_noreturn, gate_warn_function_noreturn,
6034 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
6035 Relocate from tree-cfg.c
6036 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
6038 (execute_warn_function_noreturn, gate_warn_function_noreturn,
6039 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
6040 Move to ipa-pure-const.c.
6041 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
6042 Relocate from tree-optimize.c.
6043 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
6044 make_pass_fixup_cfg): Move to tree-cfg.c.
6045 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
6046 Relocate some prototypes.
6047 * tree-data-ref.h (tree_check_data_deps) Add prototype.
6048 * tree-dump.c (dump_function_to_file): Remove prototype.
6049 Add tree-flow.h to the include file.
6050 * tree-dump.h: Remove prototype.
6051 * tree-parloops.h: New File. Add prototypes.
6052 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
6053 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
6054 from tree-ssa-loop.c.
6055 * tree-predcom.c (run_tree_predictive_commoning,
6056 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
6057 Relocate here from tree-ssa-loop.c.
6058 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
6059 ssa_name_values.release ().
6060 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
6061 (ssa_name_values): Relocate from tree-flow.h.
6062 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
6063 * tree-ssa-loop.c (run_tree_predictive_commoning,
6064 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
6065 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
6066 make_pass_graphite, pass_data_graphite_transforms,
6067 make_pass_graphite_transforms, gate_tree_parallelize_loops,
6068 tree_parallelize_loops, pass_data_parallelize_loops,
6069 make_pass_parallelize_loops): Move to other files.
6070 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
6072 * tree.h: Remove prototypes from tree-address.c.
6074 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6076 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
6077 (struct int_tree_map): Move to tree-hasher.h
6078 (SCALE, LABEL, PERCENT): Move to gimple.h
6079 * tree-flow-inline.h: Delete. Move functions to other files.
6080 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
6081 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
6082 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
6083 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
6084 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
6085 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
6086 * tree-hasher.h: Don't include tree-flow.h.
6087 (struct int_tree_map): Relocate from tree-flow.h.
6088 * tree-sra.c (contains_view_convert_expr_p): Relocate from
6089 tree-flow-inline.h and make static.
6090 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
6092 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
6093 tree-flow-inline.h and make static.
6094 * tree.h (is_global_var, may_be_aliased): Relocate from
6096 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
6097 * value-prof.c: No longer include tree-flow-inline.h.
6098 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
6100 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6102 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
6103 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
6104 (enum move_pos): Move to tree-ssa-loop-im.h
6105 * cfgloop.h: Move some prototypes.
6106 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
6107 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
6108 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
6109 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
6110 (loop_containing_stmt): Relocate from tree-flow-inline.h.
6111 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
6112 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
6113 (enum move_pos): Relocate here.
6114 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
6116 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
6117 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
6118 make_pass_lim): Relocate here from tree-ssa-loop.c.
6119 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
6121 (loop_edge_to_cancel, unloop_loops): Make static.
6122 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
6123 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
6124 (tree_complete_unroll, gate_tree_complete_unroll,
6125 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
6126 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
6127 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
6128 * tree-ssa-loop-ivopts.c: Remove local prototypes.
6129 (stmt_invariant_in_loop_p): Remove unused function.
6130 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
6131 * tree-ssa-loop-manip.h: New file. Add prototypes.
6132 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
6133 (gcov_type_to_double_int): Move to cfgloop.h.
6134 (double_int_cmp, bound_index,
6135 estimate_numbers_of_iterations_loop): Make static.
6136 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
6137 (max_loop_iterations): Factor out get_max_loop_iterations.
6138 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
6140 * tree-ssa-loop-niter.h: New file. Add prototypes.
6141 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
6142 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
6143 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
6144 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
6145 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
6146 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
6147 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
6148 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
6149 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
6150 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
6151 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
6152 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
6153 pass_data_complete_unroll, make_pass_complete_unroll,
6154 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
6155 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
6156 tree-ssa-loop-ivcanon.c.
6157 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
6158 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
6159 tree-ssa-loop-prefetch.c.
6160 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
6162 (get_lsm_tmp_name): Relocate and add suffix parameter.
6163 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
6164 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
6165 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
6166 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
6167 (get_estimated_loop_iterations): Factor out accessor from
6168 estimated_loop_iterations in tree-ssa-loop-niter.c.
6169 (get_max_loop_iterations): Factor out accessor from
6170 _max_loop_iterations in tree-ssa-niter.c.
6171 * loop-unroll.c (decide_unroll_constant_iterations,
6172 decide_unroll_runtime_iterations, decide_peel_simple,
6173 decide_unroll_stupid): Use new get_* accessors.
6175 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
6177 PR tree-optimization/20318
6178 * doc/extend.texi (returns_nonnull): New function attribute.
6179 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
6181 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
6182 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
6184 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
6187 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
6188 false if both components are bitfields.
6190 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6192 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
6193 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
6194 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
6195 * config/aarch64/aarch64.h
6196 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
6197 * config/aarch64/aarch64-simd-builtins.def
6198 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
6200 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6202 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
6204 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6206 * config/aarch64/arm_neon.h (vdiv_f64): Added.
6208 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6210 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
6211 (vneg_f64): New intrinsic.
6212 (vneg_s8): Asm replaced with C.
6213 (vneg_s16): Likewise.
6214 (vneg_s32): Likewise.
6215 (vneg_s64): New intrinsic.
6216 (vnegq_f32): Asm replaced with C.
6217 (vnegq_f64): Likewise.
6218 (vnegq_s8): Likewise.
6219 (vnegq_s16): Likewise.
6220 (vnegq_s32): Likewise.
6221 (vnegq_s64): Likewise.
6223 2013-10-09 Renlin Li <Renlin.Li@arm.com>
6225 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
6227 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6229 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
6230 packed stack special handling.
6231 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
6232 back to fixed stack slots for FPRs saved due to stdarg.
6234 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6236 * config/s390/s390.c (s390_frame_info): Restructure function.
6238 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6240 * config/s390/s390.c (struct s390_frame_layout): New field
6242 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
6243 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
6244 regs_ever_clobbered to char*.
6245 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
6246 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
6247 (s390_register_info): Enable FPR save slots. Move/Copy some
6248 functionality into ...
6249 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
6250 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
6252 (s390_frame_info): Do gpr slot allocation here now. stdarg does
6253 not imply a stack frame.
6254 (s390_init_frame_layout): Remove variable clobbered_regs.
6255 (s390_update_register_info): Remove function.
6256 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
6257 cannot be used for register renaming.
6258 (s390_hard_regno_scratch_ok): New function.
6259 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
6260 (s390_initial_elimination_offset): Change offset calculation of
6261 the return address pointer.
6262 (save_gprs): Deal with only r6 being saved from the call-saved regs.
6263 (restore_gprs): Set frame related flag.
6264 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
6265 (s390_emit_prologue): Call s390_register_info instead of
6266 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
6267 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
6268 (s390_optimize_prologue): Call s390_optimize_register_info.
6269 Try to remove also FPR slot save/restore INSNs. Remove frame
6270 related flags from restore INSNs.
6272 2013-10-08 DJ Delorie <dj@redhat.com>
6274 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
6277 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
6278 avoid conflict with the MI use of %c.
6279 * config/rl78/rl78-real.md: change %c to %C throughout.
6280 * config/rl78/rl78-virt.md: Likewise.
6282 2013-10-08 Jan Hubicka <jh@suse.cz>
6284 * config/i386/i386.c (ix86_option_override_internal): Switch
6285 to SSE math for -ffast-math when target ISA supports SSE2.
6287 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
6289 * tree-flow.h: Remove some prototypes.
6290 * tree.h: Remove some protypes, add a couple.
6291 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
6292 using_eh_for_cleanups_p): Add interface routines for front ends.
6293 * tree-eh.h: New file. Add protoptyes.
6294 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
6295 (add_stmt_to_eh_lp_fn): Make static.
6296 (lower_try_finally): Use new using_eh_for_cleanups_p.
6297 * emit-rtl.c: Include tree-eh.h.
6298 * gimple.h: Include tree-eh.h.
6300 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
6302 PR tree-optimization/58480
6303 * tree-vrp.c (infer_nonnull_range): New function.
6304 (infer_value_range): Call infer_nonnull_range.
6306 2013-10-08 Dehao Chen <dehao@google.com>
6308 PR tree-optimization/58619
6309 * tree-inline.c (copy_phis_for_bb): Combine location data
6312 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6315 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
6316 RTX_FRAME_RELATED_P on INSN.
6318 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6320 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
6321 (altivec_expand_vec_perm_const): Call it.
6323 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6325 * config/rs6000/vector.md (mov<mode>): Emit permuted move
6326 sequences for LE VSX loads and stores at expand time.
6327 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
6329 * config/rs6000/rs6000.c (rs6000_const_vec): New.
6330 (rs6000_gen_le_vsx_permute): New.
6331 (rs6000_gen_le_vsx_load): New.
6332 (rs6000_gen_le_vsx_store): New.
6333 (rs6000_gen_le_vsx_move): New.
6334 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
6335 (*vsx_le_perm_load_v4si): New.
6336 (*vsx_le_perm_load_v8hi): New.
6337 (*vsx_le_perm_load_v16qi): New.
6338 (*vsx_le_perm_store_v2di): New.
6339 (*vsx_le_perm_store_v4si): New.
6340 (*vsx_le_perm_store_v8hi): New.
6341 (*vsx_le_perm_store_v16qi): New.
6342 (*vsx_xxpermdi2_le_<mode>): New.
6343 (*vsx_xxpermdi4_le_<mode>): New.
6344 (*vsx_xxpermdi8_le_V8HI): New.
6345 (*vsx_xxpermdi16_le_V16QI): New.
6346 (*vsx_lxvd2x2_le_<mode>): New.
6347 (*vsx_lxvd2x4_le_<mode>): New.
6348 (*vsx_lxvd2x8_le_V8HI): New.
6349 (*vsx_lxvd2x16_le_V16QI): New.
6350 (*vsx_stxvd2x2_le_<mode>): New.
6351 (*vsx_stxvd2x4_le_<mode>): New.
6352 (*vsx_stxvd2x8_le_V8HI): New.
6353 (*vsx_stxvd2x16_le_V16QI): New.
6355 2013-10-07 Renlin Li <Renlin.Li@arm.com>
6357 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
6359 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6361 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
6362 loop to work also for 31bit ABI.
6363 Save the stack pointer for frame_size > 0.
6365 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6367 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
6368 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
6369 constraint letters from expanders.
6370 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
6371 retry count to general_operand.
6372 ("tabort"): Give operand 0 a mode.
6373 ("tabort_1"): Add mode and constraint letter for operand 0.
6374 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
6376 2013-10-04 Jeff Law <law@redhat.com>
6378 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
6380 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
6382 (thread_across_edge): Here. Call it.
6384 2013-10-04 Cary Coutant <ccoutant@google.com>
6386 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
6387 discarding a location list expression (or a piece of one).
6389 2013-10-03 Jan Hubicka <jh@suse.cz>
6391 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
6392 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
6393 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
6395 2013-10-03 Jan Hubicka <jh@suse.cz>
6397 * config/i386/i386.c (ix86_option_override_internal): Do not enable
6398 accumulate-outgoing-args when producing unwind info.
6400 2013-10-03 Wei Mi <wmi@google.com>
6402 * lra-constraints.c (insert_move_for_subreg): New function
6403 extracted from simplify_operand_subreg.
6404 (simplify_operand_subreg): Add reload for paradoxical subreg.
6406 2013-10-03 Rong Xu <xur@google.com>
6408 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
6409 the candidate of builtin_expect such that we should fix the
6410 size/time estimation.
6411 (estimate_function_body_sizes): Do the acutally size/time fix-up
6414 2013-10-03 Rong Xu <xur@google.com>
6416 * predict.c (tree_predict_by_opcode): Get the probability
6417 for builtin_expect from param builtin_expect_probability.
6418 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
6419 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
6420 * doc/invoke.texi: Add documentation for builtin-expect-probability.
6422 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
6425 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
6426 * calls.c (alloca_call_p): Use get_callee_fndecl.
6427 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
6428 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
6430 (vrp_visit_stmt): Remove duplicated code.
6432 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
6434 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
6435 ceildf2, btruncdf2, instead of vsx_* name.
6437 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
6438 iterators to only do V2DF and V4SF here. Move the DF code to
6439 rs6000.md where it is combined with SF mode. Replace <VSv> with
6440 just 'v' since only vector operations are handled with these insns
6441 after moving the DF support to rs6000.md.
6442 (vsx_sub<mode>3): Likewise.
6443 (vsx_mul<mode>3): Likewise.
6444 (vsx_div<mode>3): Likewise.
6445 (vsx_fre<mode>2): Likewise.
6446 (vsx_neg<mode>2): Likewise.
6447 (vsx_abs<mode>2): Likewise.
6448 (vsx_nabs<mode>2): Likewise.
6449 (vsx_smax<mode>3): Likewise.
6450 (vsx_smin<mode>3): Likewise.
6451 (vsx_sqrt<mode>2): Likewise.
6452 (vsx_rsqrte<mode>2): Likewise.
6453 (vsx_fms<mode>4): Likewise.
6454 (vsx_nfma<mode>4): Likewise.
6455 (vsx_copysign<mode>3): Likewise.
6456 (vsx_btrunc<mode>2): Likewise.
6457 (vsx_floor<mode>2): Likewise.
6458 (vsx_ceil<mode>2): Likewise.
6459 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
6460 (vsx_sminsf3): Likewise.
6461 (vsx_fmadf4): Likewise.
6462 (vsx_fmsdf4): Likewise.
6463 (vsx_nfmadf4): Likewise.
6464 (vsx_nfmsdf4): Likewise.
6465 (vsx_cmpdf_internal1): Likewise.
6467 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
6468 simpler to select whether a target has SPE or traditional floating
6469 point support in iterators.
6470 (TARGET_DF_SPE): Likewise.
6471 (TARGET_SF_FPR): Likewise.
6472 (TARGET_DF_FPR): Likewise.
6473 (TARGET_SF_INSN): Macros to say whether floating point support
6474 exists for a given operation for expanders.
6475 (TARGET_DF_INSN): Likewise.
6477 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
6478 combining of SF/DF mode operations, using both traditional and VSX
6486 (abs<mode>2): Combine SF/DF modes using traditional floating point
6487 instructions. Add support for using the upper DF registers with
6488 VSX support, and SF registers with power8-vector support. Update
6489 expanders for operations supported by both the SPE and traditional
6490 floating point units.
6491 (abs<mode>2_fpr): Likewise.
6492 (nabs<mode>2): Likewise.
6493 (nabs<mode>2_fpr): Likewise.
6494 (neg<mode>2): Likewise.
6495 (neg<mode>2_fpr): Likewise.
6496 (add<mode>3): Likewise.
6497 (add<mode>3_fpr): Likewise.
6498 (sub<mode>3): Likewise.
6499 (sub<mode>3_fpr): Likewise.
6500 (mul<mode>3): Likewise.
6501 (mul<mode>3_fpr): Likewise.
6502 (div<mode>3): Likewise.
6503 (div<mode>3_fpr): Likewise.
6504 (sqrt<mode>3): Likewise.
6505 (sqrt<mode>3_fpr): Likewise.
6506 (fre<Fs>): Likewise.
6507 (rsqrt<mode>2): Likewise.
6508 (cmp<mode>_fpr): Likewise.
6509 (smax<mode>3): Likewise.
6510 (smin<mode>3): Likewise.
6511 (smax<mode>3_vsx): Likewise.
6512 (smin<mode>3_vsx): Likewise.
6513 (negsf2): Delete SF operations that are merged with DF.
6520 (fmasf4_fpr): Likewise.
6521 (fmssf4_fpr): Likewise.
6522 (nfmasf4_fpr): Likewise.
6523 (nfmssf4_fpr): Likewise.
6524 (sqrtsf2): Likewise.
6525 (rsqrtsf_internal1): Likewise.
6526 (smaxsf3): Likewise.
6527 (sminsf3): Likewise.
6528 (cmpsf_internal1): Likewise.
6529 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
6530 (negdf2): Delete DF operations that are merged with SF.
6532 (nabsdf2): Likewise.
6538 (rsqrtdf_internal1): Likewise.
6539 (fmadf4_fpr): Likewise.
6540 (fmsdf4_fpr): Likewise.
6541 (nfmadf4_fpr): Likewise.
6542 (nfmsdf4_fpr): Likewise.
6543 (sqrtdf2): Likewise.
6544 (smaxdf3): Likewise.
6545 (smindf3): Likewise.
6546 (cmpdf_internal1): Likewise.
6547 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
6548 (btrunc<mode>2): Delete separate expander, and combine with the
6549 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
6550 (btrunc<mode>2_fpr): Likewise.
6551 (ceil<mode>2): Likewise.
6552 (ceil<mode>2_fpr): Likewise.
6553 (floor<mode>2): Likewise.
6554 (floor<mode>2_fpr): Likewise.
6555 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
6556 Add support for using the upper registers with VSX and
6557 power8-vector. Move insns to be closer to the define_expands. On
6558 VSX systems, prefer the traditional form of FMA over the VSX
6559 version, since the traditional form allows the target not to
6560 overlap with the inputs.
6561 (fms<mode>4_fpr): Likewise.
6562 (nfma<mode>4_fpr): Likewise.
6563 (nfms<mode>4_fpr): Likewise.
6565 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6566 Richard Earnshaw <richard.earnshaw@arm.com>
6568 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
6569 (struct mult_cost_table): Likewise.
6570 (struct mem_cost_table): Likewise.
6571 (struct fp_cost_table): Likewise.
6572 (struct vector_cost_table): Likewise.
6573 (cpu_cost_table): Likewise.
6574 * config/arm/arm.opt (mold-rts-costs): New option.
6575 (mnew-generic-costs): Likewise.
6576 * config/arm/arm.c (generic_extra_costs): New table.
6577 (cortexa15_extra_costs): Likewise.
6578 (arm_slowmul_tune): Use NULL as new costs.
6579 (arm_fastmul_tune): Likewise.
6580 (arm_strongarm_tune): Likewise.
6581 (arm_xscale_tune): Likewise.
6582 (arm_9e_tune): Likewise.
6583 (arm_v6t2_tune): Likewise.
6584 (arm_cortex_a5_tune): Likewise.
6585 (arm_cortex_a9_tune): Likewise.
6586 (arm_v6m_tune): Likewise.
6587 (arm_fa726te_tune): Likewise.
6588 (arm_cortex_a15_tune): Use cortex15_extra_costs.
6589 (arm_cortex_tune): Use generict_extra_costs.
6590 (shifter_op_p): New function.
6591 (arm_unspec_cost): Likewise.
6592 (LIBCALL_COST): Define.
6593 (arm_new_rtx_costs): New function.
6594 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
6595 table is available. Use old costs otherwise unless mnew-generic-costs
6597 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
6598 (cpu_cost_table): Declare.
6600 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
6603 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
6604 (*subs_mul_imm_<mode>)
6605 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
6606 (*sub_<shift>_<mode>)
6607 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
6608 Remove k constraint.
6610 2013-10-03 Ian Bolton <ian.bolton@arm.com>
6612 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
6614 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
6616 2013-10-02 Teresa Johnson <tejohnson@google.com>
6618 * predict.c (probably_never_executed): New function.
6619 (probably_never_executed_bb_p): Invoke probably_never_executed.
6620 (probably_never_executed_edge_p): Ditto.
6621 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
6622 Treat profile insanities conservatively.
6624 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
6626 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
6628 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
6630 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
6631 use it. Use smaller increase for scratch. Don't increase reject
6632 for early clobber scratch.
6633 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
6634 setting eliminated regs except setting fp from hfp.
6635 (lra_eliminate): Check lra_insn_recog_data on NULL.
6637 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6640 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
6641 setting -mvsx-timode by default until the underlying problem is fixed.
6642 (RS6000_CPU, power7 defaults): Likewise.
6644 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6646 * config/x-linux (host-linux.o): Remove header dependencies.
6647 Use $(COMPILE) and $(POSTCOMPILE).
6648 * config/t-linux-android (linux-android.o): Ditto.
6650 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6652 * Makefile.in (expmed.o-warn): Remove.
6654 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6656 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
6657 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
6659 2013-10-02 Teresa Johnson <tejohnson@google.com>
6661 * dojump.c (do_jump_1): Divide probability between
6662 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
6664 2013-10-02 Tom Tromey <tromey@redhat.com>
6666 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
6668 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6670 * tree-flow.h: Remove some prototypes.
6671 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
6672 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
6673 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
6674 mark_virtual_phi_result_for_renaming): Relocate here.
6675 * tree-into-ssa.h: Add prototypes.
6676 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
6677 single_pred_before_succ_order.
6678 (blocks_in_phiopt_order): Rename and move to cfganal.c.
6679 (nonfreeing_call_p) Move to gimple.c.
6680 * cfganal.c (single_pred_before_succ_order): Move and renamed from
6682 * basic-block.h (single_pred_before_succ_order): Add prototype.
6683 * gimple.c (nonfreeing_call_p): Relocate here.
6684 * gimple.h: Add prototype.
6685 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
6686 * tree-ssa-dom.h: New file. Relocate prototypes here.
6687 * tree-ssa.h: Include tree-ssa-dom.h.
6689 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6691 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
6692 Use $(COMPILE) and $(POSTCOMPILE).
6694 * config/alpha/x-alpha (driver-alpha.o): Ditto.
6696 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6698 * tree-flow.h: Remove some prototypes.
6699 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
6700 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6701 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
6702 propagate_tree_value*): Move from here to...
6703 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
6704 propagate_tree_value*): Relocate here.
6705 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6706 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
6707 * gimple-fold.c: Remove gimple-fold.h from include list.
6708 * tree-vrp.c: Remove gimple-fold.h from include list.
6709 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
6710 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
6711 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
6712 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
6713 * sese.c: Add tree-ssa-propagate.h to include list.
6715 2013-10-02 Richard Biener <rguenther@suse.de>
6717 * tree-loop-distribution.c: Include tree-vectorizer.h for
6719 (enum partition_kind): Remove PKIND_REDUCTION.
6720 (struct partition_s): Remove has_writes member, add reduction_p member.
6721 (partition_alloc): Adjust.
6722 (partition_builtin_p): Likewise.
6723 (partition_has_writes): Remove.
6724 (partition_reduction_p): New function.
6725 (partition_merge_into): Likewise.
6726 (generate_code_for_partition): Commonize builtin partition
6728 (rdg_cannot_recompute_vertex_p): Remove.
6729 (already_processed_vertex_p): Likewise.
6730 (rdg_flag_vertex): Do not set has_writes.
6731 (classify_partition): Adjust.
6732 (rdg_build_partitions): Do not set has_writes, treat all
6733 partitions as useful.
6734 (distribute_loop): Record number of library calls generated. Adjust.
6735 (tree_loop_distribution): Report number of loops and library
6736 calls generated as opt-info.
6738 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6740 * tree-flow.h: Include new .h files. Move prototypes.
6741 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
6742 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
6743 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
6744 * tree-pretty-print.h: Add prototypes from tree-flow.h.
6745 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
6746 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
6747 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
6748 * tree.h (get_ref_base_and_extent): Move prototype out.
6749 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
6751 * gimple-low.h: New File. Add prototypes from tree-flow.h.
6752 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
6753 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
6754 * tree-scalar-evolution.c: Include tree.h.
6755 * sese.c: Include tree.h.
6756 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
6757 * dwarf2out.c: Include tree-dfa.h.
6758 * tree-chrec.c: Include tree.h.
6759 * tree-data-ref.c: Include tree.h.
6761 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
6763 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
6766 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6768 * config/t-sol2 (sol2-c.o): Remove header dependencies.
6769 Use $(COMPILE) and $(POSTCOMPILE).
6770 (sol2-cxx.o): Likewise.
6771 (sol2-stubs.o): Likewise.
6773 * config/x-solaris (host-solaris.o): Likewise.
6775 * config/sparc/t-sparc (sparc.o): Remove.
6776 (sparc-c.o): Remove header dependencies.
6777 Use $(COMPILE) and $(POSTCOMPILE).
6778 * config/sparc/x-sparc: Likewise.
6780 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
6782 * config/arc/arc-opts.h: Add 2013 to Copyright years.
6783 * config/arc/arc700.md: Likewise.
6784 * config/arc/arc-modes.def: Likewise.
6785 * config/arc/arc-simd.h: Likewise.
6786 * config/arc/t-arc-uClibc: Likewise.
6787 * config/arc/t-arc-newlib: Likewise.
6789 2013-10-02 Renlin Li <renlin.li@arm.com>
6791 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6793 (aarch64_expand_epilogue): Likewise.
6795 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6796 Yufeng Zhang <yufeng.zhang@arm.com>
6798 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
6800 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
6801 'base_in' represent a conversion and legal_cast_p_1 holds; set
6802 'base_in' with the returned value from get_unwidened.
6804 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6806 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
6807 plus_constant is not used.
6809 2013-10-01 Wei Mi <wmi@google.com>
6811 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
6812 * config/i386/i386.md: Add define_peephole2 to
6813 break partial reg stall for cvtss2sd/cvtsd2ss.
6815 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6817 * config/arc/arc.c (pass_arc_ifcvt::clone):
6818 Update for ctxt_ -> m_ctxt change.
6820 2013-10-01 Jeff Law <law@redhat.com>
6822 * tree-ssa-threadupdate.c (struct redirection_data): Delete
6823 outgoing_edge and intermediate_edge fields. Instead store the path.
6824 (redirection_data::hash): Hash on the last edge's destination index.
6825 (redirection_data::equal): Check the entire thread path.
6826 (lookup_redirectio_data): Corresponding changes.
6827 (create_edge_and_update_destination_phis): Likewise.
6828 (thread_single_edge): Likewise.
6830 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6831 Diego Novillo <dnovillo@google.com>
6833 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
6834 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
6835 (vld32wh_insn, vld32wl_insn): Delete commented-out old
6836 versions of these patterns.
6838 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
6839 (__builtin_arc_aligned): Likewise.
6841 * config/arc/arc.md: Expand adc_0 comment stating the intended
6842 purpose and why it isn't ready.
6843 Replace commented out call_value_via_label_mixed with a
6844 plain comment about bl_s.
6846 * config/arc/arc.c (stdio.h): Don't include directly.
6847 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
6848 Remove if (1) condition.
6849 (arc_encode_section_info): Fix comment.
6851 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6853 * config/arc/arc.c (arc_conditional_register_usage):
6854 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
6855 Also set reg_alloc_order for DMA config regs.
6857 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6858 Jeremy Bennett <jeremy.bennett@embecosm.com>
6860 * doc/install.texi (--with-cpu): Mention ARC.
6861 (arc-*-elf32): New paragraph.
6862 (arc-linux-uclibc): Likewise.
6863 * doc/md.texi (Machine Constraints): Add ARC part.
6864 * doc/invoke.texi: (menu): Add ARC Options.
6865 (Machine Dependent Options) <ARC Options>: Add synopsis.
6866 (node ARC Options): Add.
6867 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6868 (ARC Built-in Functions): New section defining
6869 generic ARC built-in functions.
6870 (ARC SIMD Built-in Functions): New section defining SIMD specific
6872 (Declaring Attributes of Functions): Extended
6873 description of short_call and long_call attributes for ARC and
6874 added index entries.
6876 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6877 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6878 Joern Rennecke <joern.rennecke@embecosm.com>
6879 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6880 Brendan Kehoe <brendan@zen.org>
6881 Michael Eager <eager@eagercon.com>
6882 Simon Cook <simon.cook@embecosm.com>
6883 Jeremy Bennett <jeremy.bennett@embecosm.com>
6885 * config/arc, common/config/arc: New directories.
6887 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6888 Brendan Kehoe <brendan@zen.org>
6889 Simon Cook <simon.cook@embecosm.com>
6891 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6893 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6895 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6896 * tree-ssa-coalesce.h: New. Move prototype to here.
6897 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6898 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6899 (gimple_can_coalesce_p): Move to...
6900 * gimple.c (gimple_can_coalesce_p): Here.
6902 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6904 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6905 (dump_decl_set): Move to gimple.c.
6906 * gimple.h: Don't include tree-ssa-operands.h.
6907 (dump_decl_set): Add prototype.
6908 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6909 Move to gimple-ssa.h.
6910 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6911 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6912 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6913 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6914 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6916 (dump_decl_set): Relocate here.
6917 * gimple-ssa.h: New file.
6918 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6919 Relocate from gimple.h.
6920 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6921 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6922 swap_tree_operands and remove non-ssa path.
6923 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6924 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6926 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6927 vect_is_simple_reduction_1): Use swap_ssa_operands.
6928 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6929 (enum need_phi_state): Move to tree-into-ssa.c.
6930 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6931 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6932 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6933 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6934 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6935 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6936 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6937 num_imm_uses): Move to ssa-iterators.h.
6938 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6939 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6941 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6942 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6943 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6944 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6945 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6946 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6947 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6948 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6949 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6950 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6951 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6952 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6953 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6954 (set_phi_nodes): Move to tree-phinodes.h.
6955 * tree-ssa-operands.h (enum ssa_op_iter_type,
6956 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6957 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6958 (dump_decl_set): Remove prototype.
6959 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6960 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6961 (set_phi_nodes): Relocate from tree-flow-inline.h.
6962 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6964 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6965 include list. Temporarily add gimple.h to include list.
6966 * ssa-iterators.h: New file.
6967 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6968 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6969 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6970 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6971 Relocate from tree-ssa-operands.h.
6972 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6973 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6974 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6975 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6976 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6977 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6978 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6979 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6980 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6981 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6982 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6983 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6984 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6985 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6986 Relocate from tree-flow-inline.h.
6987 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6989 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6992 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6993 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6994 Insert '\t' to output template.
6995 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6996 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6997 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6998 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
7000 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
7002 * doc/install.texi (Host/target specific installation notes for GCC):
7003 Put @anchor before @heading.
7004 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
7005 Use @email for email addresses.
7007 2013-10-01 Jeff Law <law@redhat.com>
7009 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
7010 a vec. Only delete the path if we create one without successfully
7011 registering a jump thread.
7012 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
7014 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
7015 (paths): New vector of jump threading paths.
7016 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
7017 (THREAD_PATH): New accessor macro for the entire thread path.
7018 (lookup_redirection_data): Get intermediate and final outgoing edge
7019 from the thread path.
7020 (create_edge_and_update_destination_phis): Copy the threading path.
7021 (ssa_fix_duplicate_block_edges): Get edges and block types from the
7022 jump threading path.
7023 (ssa_redirect_edges): Get edges and block types from the jump threading
7024 path. Free the path vector.
7025 (thread_block): Get edges from the jump threading path. Look at the
7026 entire path to see if we thread to a loop exit. If we cancel a jump
7027 thread request, then free the path vector.
7028 (thread_single_edge): Get edges and block types from the jump threading
7029 path. Free the path vector.
7030 (thread_through_loop_header): Get edges and block types from the jump
7031 threading path. Free the path vector.
7032 (mark_threaded_blocks): Iterate over the vector of paths and store
7033 the path on the appropriate edge. Get edges and block types from the
7034 jump threading path.
7035 (mark_threaded_blocks): Get edges and block types from the jump
7036 threading path. Free the path vector.
7037 (thread_through_all_blocks): Use the vector of paths rather than
7038 a vector of 3-edge sets.
7039 (register_jump_thread): Accept pointer to a path vector rather
7040 than the path vector itself. Store the path vector for later use.
7043 2013-10-01 Jakub Jelinek <jakub@redhat.com>
7044 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7047 * config/s390/s390.c (s390_split_branches): Modify check for table
7049 (s390_chunkify_start): Rearrange table jump insn check in order to
7050 deal with compare and branch insns correctly.
7052 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
7056 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
7057 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
7058 define_insn_and_split.
7059 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
7060 (shiftsi3_compare): New pattern.
7062 * config/arm/unspecs.md (UNSPEC_RRX): New.
7064 2013-10-01 Alan Modra <amodra@gmail.com>
7066 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
7067 casing inout operands.
7069 2013-10-01 Richard Biener <rguenther@suse.de>
7071 PR tree-optimization/58553
7072 * tree-loop-distribution.c (struct partition_s): Add niter member.
7073 (classify_partition): Populate niter member for the partition
7074 and properly identify whether the relevant store happens before
7075 or after the loop exit.
7076 (generate_memset_builtin): Use niter member from the partition.
7077 (generate_memcpy_builtin): Likewise.
7079 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
7081 * vec.h (vec_prefix, vec): Prefix member names with "m_".
7082 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
7084 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
7086 * basic-block.h (edge_list): Prefix member names with "m_".
7087 * context.h (context): Likewise.
7088 * domwalk.h (dom_walker): Likewise.
7089 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
7090 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
7091 * hash-table.h (hash_table): Likewise.
7092 * machmode.h (bit_field_mode_iterator): Likewise.
7093 * pass_manager.h (pass_list): Likewise.
7094 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
7095 * tree-pass.h (pass_data): Likewise.
7096 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
7097 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
7098 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
7099 * asan.c (pass_data_asan): Update accordingly.
7100 * cfganal.c (control_dependences::find_control_dependence): Likewise.
7101 (control_dependences::control_dependences): Likewise.
7102 (control_dependences::~control_dependences): Likewise.
7103 (control_dependences::~control_dependences): Likewise.
7104 (control_dependences::get_edges_dependent_on): Likewise.
7105 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
7106 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
7107 * context.c (gcc::context::context): Likewise.
7108 * cprop.c (pass_rtl_cprop::clone): Likewise.
7109 * domwalk.c (dom_walker::walk): Likewise.
7110 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
7111 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
7112 * mode-switching.c (pass_mode_switching::clone): Likewise.
7113 * passes.c (opt_pass::opt_pass): Likewise.
7114 (pass_manager::pass_manager): Likewise.
7115 * predict.c (pass_strip_predict_hints::clone): Likewise.
7116 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
7117 (pass_split_all_insns::clone): Likewise.
7118 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
7120 (bit_field_mode_iterator::next_mode): Likewise.
7121 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
7122 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
7123 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
7124 * tree-complex.c (pass_lower_complex::clone): Likewise.
7125 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
7126 * tree-object-size.c (pass_object_sizes::clone): Likewise.
7127 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
7128 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
7129 (pass_fold_builtins::clone): Likewise.
7130 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
7131 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
7132 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
7133 (pass_cd_dce::clone): Likewise.
7134 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
7135 (pass_phi_only_cprop::clone): Likewise.
7136 * tree-ssa-dse.c (pass_dse::clone): Likewise.
7137 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
7138 * tree-ssa-loop.c (pass_lim::clone): Likewise.
7139 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
7140 * tree-ssa-pre.c (pass_fre::clone): Likewise.
7141 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
7142 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
7143 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
7144 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
7145 * tree-vrp.c (pass_vrp::clone): Likewise.
7146 * tsan.c (pass_tsan::clone): Likewise.
7148 2013-09-30 Jakub Jelinek <jakub@redhat.com>
7151 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
7152 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
7155 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
7156 optimization, punt if sign_bit_p looked through any zero extension.
7158 2013-09-30 Teresa Johnson <tejohnson@google.com>
7160 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
7161 Update redirected out edge count in joiner case.
7162 (ssa_redirect_edges): Common the joiner and non-joiner cases
7163 so that joiner case gets profile updates.
7165 2013-09-30 Richard Biener <rguenther@suse.de>
7167 PR tree-optimization/58554
7168 * tree-loop-distribution.c (classify_partition): Require
7169 unconditionally executed stores for memcpy and memset recognition.
7170 (tree_loop_distribution): Calculate dominance info.
7172 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
7174 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
7175 (NO_PROFILE_COUNTERS): Likewise.
7176 (PROFILE_HOOK): Likewise.
7177 (FUNCTION_PROFILER): Likewise.
7178 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
7180 2013-09-30 Iain Sandoe <iain@codesourcery.com>
7182 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
7183 calls and defines in TARGET_MACHO conditional.
7184 (load_macho_picbase_di): Likewise.
7185 (reload_macho_picbase): Likewise.
7186 (reload_macho_picbase_si): Likewise.
7187 (reload_macho_picbase_di): Likewise.
7188 (nonlocal_goto_receiver): Likewise.
7190 2013-09-30 Nick Clifton <nickc@redhat.com>
7192 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
7193 that use the MSP430X ISA.
7194 (msp430_option_override): Scan -mmcu command line option for any
7195 MCU name that supports the MSP430X ISA.
7196 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
7197 -mmcu options which enable the MSP430X ISA.
7199 2013-09-30 Richard Biener <rguenther@suse.de>
7202 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
7203 before looking for setjmp-like calls.
7205 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7208 * config/darwin-protos.h (machopic_get_function_picbase): New.
7209 * config/darwin.c (machopic_get_function_picbase): New.
7210 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
7211 label for a new func. (load_macho_picbase_di): Likewise.
7212 (reload_macho_picbase): New expand.
7213 (reload_macho_picbase_si): New insn.
7214 (reload_macho_picbase_di): New insn.
7215 (nonlocal_goto_receiver): New define and split.
7216 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
7217 (unspecv enum): Add UNSPECV_NLGR.
7219 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7221 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
7222 that altivec registers are correctly sized on Darwin.
7224 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7226 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
7227 darwin-driver.o): Use COMPILE and POSTCOMPILE.
7228 * config/x-darwin (host-darwin.o): Likewise.
7229 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
7230 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
7231 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
7233 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
7235 * doc/invoke.texi: Fix usage of @tie{} command.
7237 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
7239 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
7241 2013-09-28 Jan Hubicka <jh@suse.cz>
7243 * config/i386/x86-tune.def: Add documentation for each of the options;
7246 2013-09-28 Jan Hubicka <jh@suse.cz>
7248 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
7250 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
7251 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
7252 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
7253 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
7254 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
7256 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
7258 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
7259 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
7260 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
7261 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
7262 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
7263 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
7264 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
7265 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
7266 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
7267 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
7268 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
7269 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
7270 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
7271 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
7272 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
7273 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
7274 gimple.h, godump.c, graphite-clast-to-gimple.c,
7275 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
7276 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
7277 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
7278 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
7279 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
7280 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
7281 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
7282 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
7283 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
7284 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
7285 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
7286 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
7287 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
7288 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
7289 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
7290 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
7291 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
7292 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
7293 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
7294 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
7295 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
7296 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
7297 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
7298 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
7299 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
7300 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
7301 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
7302 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
7303 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
7304 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
7305 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
7306 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
7307 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
7308 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
7309 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
7310 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
7311 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
7312 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
7313 whitespace before "(".
7315 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
7317 * expr.h (extract_bit_field): Remove packedp parameter.
7318 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
7319 from forward declaration.
7320 (store_split_bit_field): Remove packedp arg from calls to
7321 extract_fixed_bit_field.
7322 (extract_bit_field_1): Remove packedp parameter and packedp
7323 argument from recursive calls and calls to extract_fixed_bit_field.
7324 (extract_bit_field): Remove packedp parameter and corresponding
7325 arg to extract_bit_field_1.
7326 (extract_fixed_bit_field): Remove packedp parameter. Remove code
7328 (extract_split_bit_field): Remove packedp arg from call to
7329 extract_fixed_bit_field.
7330 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
7331 (copy_blkmode_from_reg): Likewise.
7332 (copy_blkmode_to_reg): Likewise.
7333 (read_complex_part): Likewise.
7334 (store_field): Likewise.
7335 (expand_expr_real_1): Likewise.
7336 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
7337 to extract_bit_field.
7338 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
7339 call to extract_bit_field.
7340 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
7341 call to extract_bit_field.
7342 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
7343 and special packedp behavior from -fstrict-volatile-bitfields
7346 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7348 * lra-eliminations.c (init_elim_table): Guard value_p.
7350 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7352 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
7353 DFmode, DImode, and SFmode in the upper VSX registers based on the
7354 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
7355 if -mpower8-vector. Combine -mvsx-timode handling with the rest
7356 of the VSX register handling.
7358 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
7360 (zero_extendsidi2_lfiwzx): Add support for loading into the
7361 Altivec registers with -mpower8-vector. Use wu/wv constraints to
7362 only do VSX memory options on Altivec registers.
7363 (extendsidi2_lfiwax): Likewise.
7364 (extendsfdf2_fpr): Likewise.
7365 (mov<mode>_hardfloat, SF/SD modes): Likewise.
7366 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
7367 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
7368 (movdi_internal64): Likewise.
7370 2013-09-27 Xinliang David Li <davidxl@google.com>
7372 * opts.c (finish_options): Adjust parameters
7373 according to vect cost model.
7374 (common_handle_option): Set dynamic vect cost
7376 targhooks.c (default_add_stmt_cost): Compute stmt cost
7378 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
7379 Use helper function.
7380 * tree-vectorizer.h (unlimited_cost_model): New function.
7381 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
7382 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
7384 (vect_enhance_data_refs_alignment): Ditto.
7385 * flag-types.h: New enum.
7386 * common/config/i386/i386-common.c (ix86_option_init_struct):
7387 No need to initialize vect_cost_model flag.
7388 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
7391 2013-09-27 Diego Novillo <dnovillo@google.com>
7393 * gimple.h (enum ssa_mode): Remove.
7395 2013-09-27 Paulo Matos <pmatos@broadcom.com>
7397 * cfgloop.h (number_of_loops): Fix typo in check for null.
7399 2013-09-27 Jakub Jelinek <jakub@redhat.com>
7402 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
7403 are children of outermost saved_cfun's loop, and set it up to
7404 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
7405 if loop != loop0 and SESE region contains bbs that belong to loop0.
7407 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
7409 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
7410 (binary_scale_code_p, get_base_term, get_index_term): New functions.
7411 (set_address_segment, set_address_base, set_address_index)
7412 (set_address_disp): Accept the argument unconditionally.
7413 (baseness): Remove must_be_base_p and must_be_index_p checks.
7414 (decompose_normal_address): Classify as much as possible in the
7417 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
7419 * cse.c (count_reg_usage): Handle INT_LIST.
7420 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
7421 * reginfo.c (reg_scan_mark_refs): Likewise.
7422 * reload1.c (eliminate_regs_1): Likewise.
7424 2013-09-27 Iain Sandoe <iain@codesourcery.com>
7427 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
7430 2013-09-27 Richard Biener <rguenther@suse.de>
7432 PR tree-optimization/58459
7433 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
7434 restriction not propagating into loops.
7436 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
7438 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
7439 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
7440 * doc/tree-ssa.texi (Walking use-def chains): Delete.
7442 2013-09-26 Richard Biener <rguenther@suse.de>
7444 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
7446 2013-09-26 Richard Biener <rguenther@suse.de>
7448 * alias.h (component_uses_parent_alias_set): Rename to ...
7449 (component_uses_parent_alias_set_from): ... this.
7450 * alias.c (component_uses_parent_alias_set): Rename to ...
7451 (component_uses_parent_alias_set_from): ... this and return
7453 (reference_alias_ptr_type_1): Use the result from
7454 component_uses_parent_alias_set_from instead of stripping
7455 components one at a time.
7456 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
7458 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7460 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
7461 Move prototypes to...
7462 * tree-ssa-ter.h: New File. Move prototypes here.
7463 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
7464 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
7465 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
7466 from is_replaceable_p.
7467 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
7468 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
7469 (process_replaceable): Use ter_is_replaceable_p.
7470 (find_replaceable_in_bb): Use ter_is_replaceable_p.
7471 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
7472 newly refactored ssa_is_replaceable_p.
7473 * cfgexpand.c: Include tree-outof-ssa.h.
7474 * ssaexpand.h: Delete.
7476 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7478 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
7479 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
7481 (create_gimple_tmp): Delete.
7482 (get_expr_type, build_assign, build_type_cast): Move to...
7483 * gimple-builder.c: New File.
7484 (get_expr_type): Relocate from gimple.c.
7485 (build_assign, build_type_cast): Change to only create ssanames.
7486 * gimple.h: Move prototypes to...
7487 * gimple-builder.h: New File. Here.
7488 * tree-ssa.h: And here.
7489 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
7490 count_uses_and_derefs): Relocate from gimple.c.
7491 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
7492 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
7493 * tree-ssa-math-opts (execute_cse_reciprocals): Use
7494 gimple_replace_ssa_lhs.
7495 * asan.c: Include gimple-builder.h.
7496 * Makefile.in: Add gimple-builder.o.
7498 2013-09-26 Richard Biener <rguenther@suse.de>
7500 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
7501 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
7502 (loe_visit_block): Use gcc_checking_assert.
7503 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
7504 gimple_assign_ssa_name_copy_p.
7505 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
7508 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
7510 * config/rs6000/t-rs6000 (rs6000.o): Remove.
7511 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
7513 2013-09-26 Richard Biener <rguenther@suse.de>
7515 PR tree-optimization/58539
7516 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
7517 the fact that debug statements are not taking part in loop-closed
7520 2013-09-26 Nick Clifton <nickc@redhat.com>
7522 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
7523 time warning message.
7524 (msp430_print_operand_raw): Delete unused letter parameter.
7525 (TARGET_PRINT_OPERAND_ADDRESS): Define.
7526 (msp430_print_operand_address): New function.
7527 (msp430_print_operand): Move address printing code from here to
7529 * config/msp430/msp430.md (movsipsi2): Add comment in generated
7531 (zero_extendpsisi2): Likewise.
7532 (extendpsisi2): New pattern.
7533 (andneghi3): New pattern.
7535 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
7537 * config/aarch64/aarch64.opt (mlra): New option.
7538 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
7539 (TARGET_LRA_P): Define.
7541 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
7543 * expr.c (expand_assignment): Remove obsolete comment.
7545 2013-09-25 Jeff Law <law@redhat.com>
7547 * tree-flow.h (thread_through_all_blocks): Prototype moved into
7548 tree-ssa-threadupdate.h.
7549 (register_jump_thread): Similarly.
7550 * tree-ssa-threadupdate.h: New header file.
7551 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
7552 * tree-vrp.c: Likewise.
7553 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
7554 (thread_around_empty_blocks): Change type of path vector argument to
7555 an edge,type pair from just an edge. Initialize both elements when
7556 appending to a jump threading path. Tweak references to elements
7558 (thread_across_edge): Similarly. Release memory for the elements
7560 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
7561 (dump_jump_thread_path): New function broken out from
7562 register_jump_thread.
7563 (register_jump_thread): Use dump_jump_thread_path. Change type of
7564 path vector entries. Search the path for NULL edges and dump
7565 the path if one is found. Tweak the conversion of path to 3-edge
7566 form to use the block copy type information embedded in the path.
7568 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7570 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
7572 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7573 Vladimir Makarov <vmakarov@redhat.com>
7575 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
7576 from the least significant bit.
7577 (strip_address_mutations): Add bitfield operations handling.
7578 (must_be_index_p): Add shifting and rotate operations handling.
7579 (set_address_base): Use must_be_base_p predicate.
7580 (set_address_index): Use must_be_index_p predicate.
7582 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
7583 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7584 Sergey Lega <sergey.s.lega@intel.com>
7585 Anna Tikhonova <anna.tikhonova@intel.com>
7586 Ilya Tocar <ilya.tocar@intel.com>
7587 Andrey Turetskiy <andrey.turetskiy@intel.com>
7588 Ilya Verbin <ilya.verbin@intel.com>
7589 Kirill Yukhin <kirill.yukhin@intel.com>
7590 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7592 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
7594 (ix86_expand_vector_move_misalign): Support new unaligned load and
7595 stores and use new names.
7596 (CODE_FOR_sse2_storedqu): Rename to ...
7597 (CODE_FOR_sse2_storedquv16qi): ... this.
7598 (CODE_FOR_sse2_loaddqu): Rename to ...
7599 (CODE_FOR_sse2_loaddquv16qi): ... this.
7600 (CODE_FOR_avx_loaddqu256): Rename to ...
7601 (CODE_FOR_avx_loaddquv32qi): ... this.
7602 (CODE_FOR_avx_storedqu256): Rename to ...
7603 (CODE_FOR_avx_storedquv32qi): ... this.
7604 * config/i386/i386.md (fpint_logic): New.
7605 * config/i386/sse.md (VMOVE): Extend for AVX512.
7609 (VI_UNALIGNED_LOADSTORE): Ditto.
7610 (sse2_avx_avx512f): Ditto.
7611 (sse2_avx2): Extend for AVX512.
7612 (sse4_1_avx2): Ditto.
7613 (avx2_avx512f): New.
7614 (sse): Extend for AVX512.
7617 (avxsizesuffix): Ditto.
7618 (sseintvecmode): Ditto.
7620 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
7621 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
7622 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7623 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
7624 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7625 (<sse2_avx_avx512f>_storedqu<mode): ... this.
7626 (<sse>_movnt<mode>): Replace constraint "x" with "v".
7627 (STORENT_MODE): Extend for AVX512.
7628 (*absneg<mode>2): Replace constraint "x" with "v".
7629 (*mul<mode>3): Ditto.
7630 (*ieee_smin<mode>3): Ditto.
7631 (*ieee_smax<mode>3): Ditto.
7632 (avx_cmp<mode>3): Replace VF with VF_128_256.
7633 (*<sse>_maskcmp<mode>3_comm): Ditto.
7634 (<sse>_maskcmp<mode>3): Ditto.
7635 (<sse>_andnot<mode>3): Extend for AVX512.
7636 (<code><mode>3, anylogic): Replace VF with VF_128_256.
7637 (<code><mode>3, fpint_logic): New.
7638 (*<code><mode>3): Extend for AVX512.
7639 (avx512flogicsuff): New.
7640 (avx512f_<logic><mode>): Ditto.
7641 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
7643 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
7644 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
7645 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
7646 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
7647 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
7648 (xop_vpermil2<mode>3): Ditto.
7649 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
7650 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
7651 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
7652 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
7654 2013-09-25 Tom Tromey <tromey@redhat.com>
7656 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
7657 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
7658 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
7659 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
7660 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
7661 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
7662 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
7663 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
7664 (GRAPHITE_HTAB_H): Remove.
7666 2013-09-25 Tom Tromey <tromey@redhat.com>
7668 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
7670 2013-09-25 Tom Tromey <tromey@redhat.com>
7672 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
7674 2013-09-25 Tom Tromey <tromey@redhat.com>
7676 * config/i386/t-i386 (i386.o): Remove.
7677 (i386-c.o): Use COMPILE and POSTCOMPILE.
7679 2013-09-25 Tom Tromey <tromey@redhat.com>
7681 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
7683 2013-09-25 Tom Tromey <tromey@redhat.com>
7685 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
7686 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
7687 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
7688 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
7689 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
7690 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
7691 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
7692 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
7693 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
7694 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
7695 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
7696 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
7697 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
7698 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
7699 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
7700 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
7701 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
7702 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
7703 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
7704 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
7705 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
7706 (tree-ssa-pre.o, tree-ssa-sccvn.o)
7707 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
7708 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
7709 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
7710 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
7711 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
7712 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
7713 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
7714 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
7715 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
7716 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
7717 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
7718 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
7719 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
7720 (graphite-clast-to-gimple.o, graphite-dependences.o)
7721 (graphite-interchange.o, graphite-poly.o)
7722 (graphite-scop-detection.o, graphite-sese-to-poly.o)
7723 (graphite-optimize-isl.o, tree-vect-loop.o)
7724 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
7725 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
7726 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
7727 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
7728 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
7729 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
7730 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
7731 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
7732 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
7733 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
7734 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
7735 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
7736 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
7737 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
7738 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
7739 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
7740 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
7741 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
7742 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
7743 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
7744 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
7745 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
7746 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
7747 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
7748 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
7749 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
7750 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
7751 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
7752 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
7753 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
7754 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
7755 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
7756 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
7757 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
7758 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
7759 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
7760 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
7761 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
7762 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
7763 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
7764 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
7765 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
7766 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
7767 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
7768 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
7769 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
7770 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
7771 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
7772 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
7773 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
7774 (gcov-dump.o): Remove.
7775 (default-c.o): Use COMPILE and POSTCOMPILE.
7776 (CFLAGS-gcc.o): New variable.
7777 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
7779 2013-09-25 Tom Tromey <tromey@redhat.com>
7781 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
7782 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
7783 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
7784 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
7785 (c-family/c-pragma.o, c-family/c-pretty-print.o)
7786 (c-family/c-semantics.o, c-family/c-ada-spec.o)
7787 (c-family/array-notation-common.o, c-family/stub-objc.o)
7788 (c-family/c-ubsan.o): Remove.
7790 2013-09-25 Tom Tromey <tromey@redhat.com>
7792 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
7794 2013-09-25 Tom Tromey <tromey@redhat.com>
7796 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
7797 to add -DENABLE_SHARED_LIBGCC.
7798 (gcc.o): Don't use subshell.
7800 2013-09-25 Tom Tromey <tromey@redhat.com>
7802 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
7803 * configure.ac: Don't invoke AM_PROG_CC_C_O.
7804 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
7805 * configure, config.in: Rebuild.
7807 2013-09-25 Tom Tromey <tromey@redhat.com>
7809 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
7810 (COMPILE, POSTCOMPILE): New variables.
7811 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
7812 (DEPFILES): New variable.
7813 Include ".Po" files.
7814 * configure.ac: Add checks for dependency checking.
7815 * configure, aclocal.m4: Regenerate.
7817 2013-09-25 Tom Tromey <tromey@redhat.com>
7819 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7820 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
7822 2013-09-25 Tom Tromey <tromey@redhat.com>
7824 * Makefile.in (generated_files): Add options.h,
7825 target-hooks-def.h, insn-opinit.h,
7826 common/common-target-hooks-def.h, pass-instances.def,
7827 c-family/c-target-hooks-def.h.
7829 2013-09-25 Jeff Law <law@redhat.com>
7831 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
7832 than foo[foo.length () - 1] to access last member in a vec.
7833 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
7835 2013-09-25 Richard Biener <rguenther@suse.de>
7838 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
7840 2013-09-25 Jan Hubicka <jh@suse.cz>
7842 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
7845 2013-09-25 Marek Polacek <polacek@redhat.com>
7848 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
7849 when determining the type name.
7851 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
7853 * config/sh/sh.md: Fix formatting.
7855 2013-09-24 Xinliang David Li <davidxl@google.com>
7857 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7858 max peel iterations parameter.
7859 * param.def: New parameter.
7860 * doc/invoke.texi: Document New parameter.
7862 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7864 * gimple-pretty-print.c: Various whitespace tweaks.
7865 * tree-core.h: Likewise.
7866 * tree-pretty-print.c: Likewise.
7867 * tree-ssa-alias.c: Likewise.
7868 * tree-ssa-copy.c: Likewise.
7869 * tree-ssanames.c: Likewise.
7870 * tree-ssanames.h: Likewise.
7871 * tree-vrp.c: Likewise.
7873 2013-09-24 Alan Modra <amodra@gmail.com>
7877 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7878 for output operands that disallow regs. Don't use EXPAND_WRITE on
7881 2013-09-24 Richard Biener <rguenther@suse.de>
7884 * tree.c (reference_alias_ptr_type): Move ...
7885 * alias.c (reference_alias_ptr_type): ... here and implement
7886 in terms of the new reference_alias_ptr_type_1.
7887 (ref_all_alias_ptr_type_p): New helper.
7888 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7889 use ref_all_alias_ptr_type_p.
7890 (get_deref_alias_set): Add flag_strict_aliasing check here.
7891 (reference_alias_ptr_type_1): New function, split out from ...
7892 (get_alias_set): ... here.
7893 (alias_ptr_types_compatible_p): New function.
7894 * alias.h (reference_alias_ptr_type): Declare.
7895 (alias_ptr_types_compatible_p): Likewise.
7896 * tree.h (reference_alias_ptr_type): Remove.
7897 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7898 to compare MEM_REF alias types.
7900 2013-09-24 Richard Biener <rguenther@suse.de>
7902 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7904 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7906 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7907 reload helper function arrays into a single array reg_addr.
7908 (reload_fpr_gpr): Likewise.
7909 (reload_gpr_vsx): Likewise.
7910 (reload_vsx_gpr): Likewise.
7911 (struct rs6000_reg_addr): Likewise.
7912 (reg_addr): Likewise.
7913 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7914 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7915 (rs6000_init_hard_regno_mode_ok): Likewise.
7916 (rs6000_secondary_reload_direct_move): Likewise.
7917 (rs6000_secondary_reload): Likewise.
7919 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7920 constraints: wu, ww, and wy. Repurpose wv constraint added during
7921 power8 changes. Put wg constraint in alphabetical order.
7923 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7924 for future work to add ISA 2.07 VSX single precision support.
7925 (-mvsx-scalar-double): Change default from -1 to 1, update
7926 documentation comment.
7927 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7928 (-mupper-regs-df): New debug switch to control whether DF values
7929 can go in the traditional Altivec registers.
7930 (-mupper-regs-sf): New debug switch to control whether SF values
7931 can go in the traditional Altivec registers.
7933 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7935 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7936 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7937 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7938 (rs6000_debug_legitimate_address_p): Print if we are running
7939 before, during, or after reload.
7940 (rs6000_secondary_reload): Add a comment.
7941 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7943 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7944 constraints. Update documentation string.
7945 (wd constraint): Likewise.
7946 (wf constraint): Likewise.
7947 (wg constraint): Likewise.
7948 (wn constraint): Likewise.
7949 (ws constraint): Likewise.
7950 (wt constraint): Likewise.
7951 (wx constraint): Likewise.
7952 (wz constraint): Likewise.
7953 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7955 (ww constraint): Likewise.
7956 (wy constraint): Likewise.
7957 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7958 in the previous submissions.
7959 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7961 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7963 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7964 (REG_BR_PROB): Say that the probability is stored in an int_list.
7965 * reg-notes.def: Update commentary to mention INT_LIST.
7966 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7967 (INT_LIST): New rtx.
7968 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7969 * rtlanal.c (int_reg_note_p): New function.
7970 (alloc_reg_note): Assert that the note does not have an int argument.
7971 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7972 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7973 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7974 rather than an INSN_LIST. Handle INT_LIST.
7975 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7976 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7977 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7978 Manipulate them as ints rather than rtxes.
7979 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7980 * regmove.c (copy_src_to_dest): Likewise.
7981 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7983 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7984 into the cases that need it.
7985 * config/arm/arm.c (arm_unwind_emit): Likewise.
7987 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7988 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7989 * loop-doloop.c (add_test, doloop_modify): Likewise.
7990 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7991 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7992 * predict.c (combine_predictions_for_insn): Likewise.
7993 * print-rtl.c (print_rtx): Handle INT_LIST.
7994 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7995 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7996 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7997 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7998 (ix86_split_fp_branch, predict_jump): Likewise.
7999 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
8000 * config/sh/sh.c (expand_cbranchsi4): Likewise.
8001 * config/spu/spu.c (ea_load_store_inline): Likewise.
8003 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
8004 value of a REG_BR_PROB note.
8005 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
8006 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
8007 * emit-rtl.c (try_split): Likewise.
8008 * predict.c (br_prob_note_reliable_p): Likewise.
8009 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
8010 * reorg.c (mostly_true_jump): Likewise.
8011 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
8012 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
8013 * config/i386/i386.c (ix86_print_operand): Likewise.
8014 * config/ia64/ia64.c (ia64_print_operand): Likewise.
8015 * config/mmix/mmix.c (mmix_print_operand): Likewise.
8016 * config/rs6000/rs6000.c (output_cbranch): Likewise.
8017 * config/s390/s390.c (s390_expand_tbegin): Likewise.
8018 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
8019 * config/sparc/sparc.c (output_cbranch): Likewise.
8020 * config/spu/spu.c (get_branch_target): Likewise.
8021 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
8022 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
8024 2013-09-23 Jan Hubicka <jh@suse.cz>
8026 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
8028 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
8029 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
8031 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
8032 (gimple_fold_call): Dump inconsistent devirtualizations; add
8033 sanity check for type based devirtualizations.
8034 * ipa-prop.c: Include ipa-utils.h
8035 (ipa_intraprocedural_devirtualization): Add sanity check.
8036 (try_make_edge_direct_virtual_call): Likewise.
8038 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
8040 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
8041 assignment statements.
8043 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
8045 * gimple-pretty-print.c (dump_ssaname_info): New function.
8046 (dump_gimple_phi): Call it.
8047 (pp_gimple_stmt_1): Likewise.
8048 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
8049 (range_info_def): Declare.
8050 * tree-pretty-print.c (pp_double_int): New function.
8051 (dump_generic_node): Call it.
8052 * tree-pretty-print.h (pp_double_int): Declare.
8053 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
8054 * tree-ssanames.h (range_info_def): New structure.
8055 (value_range_type): Move definition here.
8056 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
8058 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
8060 (set_range_info): New function.
8061 (get_range_info): Likewise.
8062 (duplicate_ssa_name_range_info): Likewise.
8063 (duplicate_ssa_name_fn): Check pointer type and call
8064 duplicate_ssa_name_range_info.
8065 * tree-ssa-copy.c (fini_copy_prop): Likewise.
8066 * tree-vrp.c (value_range_type): Remove definition, now in
8068 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
8069 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
8070 (SSA_NAME_RANGE_INFO): New macro.
8072 2013-09-23 Richard Biener <rguenther@suse.de>
8074 PR tree-optimization/58464
8075 * tree-ssa-pre.c (phi_trans_lookup): Remove.
8076 (phi_trans_add): Change to add conditionally on being not present.
8077 (phi_translate_1): Remove recursion detection here.
8078 (phi_translate): Pre-seed the cache with NULL to catch
8079 recursion here in a more generic way.
8080 (bitmap_find_leader): Adjust comment.
8081 (get_representative_for): Dump value-numbers.
8082 (create_expression_by_pieces): Likewise.
8083 (insert_into_preds_of_block): Likewise.
8085 2013-09-23 Christian Bruel <christian.bruel@st.com>
8088 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
8089 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
8091 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
8094 2013-09-20 Renlin Li <renlin.li@arm.com>
8096 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
8098 (aarch64_expand_epilogue): Likewise.
8099 (aarch64_legitimize_reload_address): Likewise.
8101 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
8103 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
8104 NULL_TREE before pushing them onto the vector. Likewise for labels.
8106 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
8108 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
8109 comparison operators when -fno-trapping-math is in effect.
8110 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
8111 comparison operators in TFmode and assert that unsupported operators
8113 (ia64_print_operand): Likewise.
8115 2013-09-21 Jan Hubicka <jh@suse.cz>
8117 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
8118 (sse_typeless_stores): Enable for core
8119 (sse_load0_by_pxor): Likewise.
8120 (four_jump_limit): Disable for core.
8121 (pad_returns): Likewise.
8122 (avoid_vector_decode): Likewise.
8123 (fuse_cmp_and_branch): Enable for cores.
8124 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
8126 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
8129 * config/pa/pa.c (pa_option_override): Disable auto increment and
8130 decrement instructions until reload is completed.
8132 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
8133 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
8134 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
8136 2013-09-20 DJ Delorie <dj@redhat.com>
8137 Nick Clifton <nickc@redhat.com>
8139 * config/rl78/rl78.c: Various whitespace and comment tweaks.
8140 (need_to_save): Save bank 0 on interrupts.
8141 (characterize_address): Strip far address wrappers.
8142 (rl78_as_legitimate_address): Likewise.
8143 (transcode_memory_rtx): Likewise.
8144 (rl78_peep_movhi_p): Disable this peephole after devirt.
8145 (rl78_propogate_register_origins): Forget all origins when a
8147 * config/rl78/rl78-virt.md: Various whitespace tweaks.
8148 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
8150 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
8151 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
8152 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
8154 * config/rl78/constraints.md: Various whitespace and paren tweaks.
8156 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
8158 * config/pa/pa.md: In "scc" insn patterns, change output template to
8159 handle const0_rtx in reg_or_0_operand operands.
8161 2013-09-20 Martin Husemann <martin@NetBSD.org>
8164 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
8165 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
8167 2013-09-20 Richard Biener <rguenther@suse.de>
8170 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
8171 remembering SSA name version and block index.
8172 (new_scev_info_str): Adjust.
8173 (hash_scev_info): Likewise. Also hash the block index.
8174 (eq_scev_info): Adjust.
8175 (find_var_scev_info): Likewise.
8176 (struct instantiate_cache_entry): Remove.
8177 (struct instantiate_cache_type): Use a htab to map name, block
8179 (instantiate_cache_type::~instantiate_cache_type): Adjust.
8180 (get_instantiated_value_entry): Likewise.
8181 (hash_idx_scev_info, eq_idx_scev_info): New functions.
8182 (instantiate_scev_name): Adjust.
8184 2013-09-20 Jeff Law <law@redhat.com>
8186 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
8188 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
8190 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
8191 Call aarch64_simd_expand_args to update op[argc].
8193 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
8195 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
8198 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
8200 * gengtype.c (file_rules): Added rule for *.cc files.
8201 (get_output_file_with_visibility): Give fatal message when no
8204 2013-09-20 Renlin Li <renlin.li@arm.com>
8206 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
8207 (aarch64_expand_epilogue): Likewise.
8208 (aarch64_legitimize_reload_address): Likewise.
8210 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
8213 * expr.c (expand_assignment): Remove misalignp code path.
8215 2013-09-20 Marek Polacek <polacek@redhat.com>
8218 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
8219 TYPE_PRECISION. Add asserts.
8221 2013-09-20 Richard Biener <rguenther@suse.de>
8223 PR tree-optimization/58453
8224 * tree-loop-distribution.c (distribute_loop): Apply the cost
8225 model for -ftree-loop-distribute-patterns, too.
8227 2013-09-20 Richard Biener <rguenther@suse.de>
8230 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
8231 make type comparison less strict.
8233 2013-09-20 Alan Modra <amodra@gmail.com>
8235 * configure: Regenerate.
8236 * aclocal.m4: Regenerate.
8238 2013-09-20 Marek Polacek <polacek@redhat.com>
8241 * doc/extend.texi: Document that attribute used is meant to be used
8242 on variables with static storage duration.
8244 2013-09-19 Jakub Jelinek <jakub@redhat.com>
8246 PR tree-optimization/58472
8247 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
8248 simd_lane_access set inv_p = false.
8249 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
8250 the simduid magic VAR_DECL.
8252 2013-09-19 Jan Hubicka <jh@suse.cz>
8254 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
8256 2013-09-17 Jeff Law <law@redhat.com>
8258 * tree-ssa-dom.c (record_temporary_equivalences): New function
8259 split out of dom_opt_dom_walker::after_dom_children.
8260 (dom_opt_dom_walker::thread_across_edge): Move common code
8261 in here from dom_opt_dom_walker::after_dom_children.
8262 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
8264 2013-09-19 Jan Hubicka <jh@suse.cz>
8266 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
8267 (TARGET_GENERIC): Use PROCESOR_GENERIC
8268 (enum processor_type): Unify generic32 and 64.
8269 * i386.md (cpu): Likewise.
8270 * x86-tune.def (use_leave): Enable for generic32.
8271 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
8272 * athlon.md: Change generic64 to generic in all occurences.
8273 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
8274 (ix86_target_macros_internal): Likewise.
8275 * driver-i386.c (host_detect_local_cpu): Likewise.
8276 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
8278 (generic_memcpy, generic_memset, generic_cost): This one.
8279 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
8280 (m_GENERIC32, m_GENERIC64): Remove.
8281 (m_GENERIC): Turn into one flag.
8282 (processor_target): Unify generic tunnings.
8283 (ix86_option_override_internal): Replace generic32/64 by generic.
8284 (ix86_issue_rate): Likewise.
8285 (ix86_adjust_cost): Likewise.
8287 2013-09-19 Jan Hubicka <jh@suse.cz>
8289 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
8290 of speculative flag.
8292 2013-09-19 Jakub Jelinek <jakub@redhat.com>
8294 * omp-low.c (expand_omp_sections): Always pass len - 1 to
8295 GOMP_sections_start, even if !exit_reachable.
8297 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
8299 * lra-constraints.c (need_for_all_save_p): Use macro
8300 HARD_REGNO_CALL_PART_CLOBBERED.
8301 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
8302 set up pseudo conflict hard regs.
8304 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8307 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
8310 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
8312 PR rtl-optimization/58438
8313 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
8314 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
8315 from previous subpasses.
8317 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
8319 * arm.c (arm_get_frame_offsets): Validate architecture supports
8320 LDRD/STRD before accepting the tuning preference.
8321 (arm_expand_prologue): Likewise.
8322 (arm_expand_epilogue): Likewise.
8324 2013-09-18 Richard Biener <rguenther@suse.de>
8326 PR tree-optimization/58417
8327 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
8328 have chrecs with symbols defined in the loop as operands.
8329 (chrec_fold_multiply): Likewise.
8330 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
8331 parameters before folding binary operations.
8332 (struct instantiate_cache_entry_hasher): Remove.
8333 (struct instantiate_cache_type): Use a pointer-map.
8334 (instantiate_cache_type::instantiate_cache_type): New function.
8335 (instantiate_cache_type::get): Likewise.
8336 (instantiate_cache_type::set): Likewise.
8337 (instantiate_cache_type::~instantiate_cache_type): Adjust.
8338 (get_instantiated_value_entry): Likewise.
8339 (global_cache): New global.
8340 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
8341 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
8342 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
8343 (instantiate_scev_name): Adjust.
8344 (instantiate_scev): Construct global instead of local cache.
8345 (resolve_mixers): Likewise.
8347 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
8348 Paolo Carlini <paolo.carlini@oracle.com>
8351 * doc/implement-cxx.texi: Fix references to the C++ standards.
8353 2013-09-18 Jakub Jelinek <jakub@redhat.com>
8355 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
8356 * tree-vect-data-refs.c (vect_analyze_data_refs): For
8357 simd_lane_access drs, update also DR_ALIGNED_TO.
8359 2013-09-18 Marek Polacek <polacek@redhat.com>
8362 * doc/extend.texi: Document no_sanitize_undefined attribute.
8363 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
8364 no_sanitize_undefined attribute.
8366 2013-09-18 Nick Clifton <nickc@redhat.com>
8368 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
8369 (ASM_DECLARE_FUNCTION_NAME): Define.
8371 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
8373 * compare-elim.c (find_comparison_dom_walker): New class
8374 (find_comparisons_in_bb): Rename to
8375 find_comparison_dom_walker::before_dom_children
8376 (find_comparisons): Adjust
8377 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
8379 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
8380 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
8381 (init_walk_dominator_tree): Remove declaration.
8382 (fini_walk_dominator_tree): Remove declaration.
8383 * fwprop.c (single_def_use_dom_walker): New class
8384 (single_def_use_enter_block): Convert to
8385 single_def_use_dom_walker::before_dom_children.
8386 (single_def_use_leave_block): Convert to
8387 single_def_use_dom_walker::after_dom_children.
8388 (build_single_def_use_links): Adjust.
8389 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
8391 (find_candidates_in_block): Convert to
8392 find_candidates_dom_walker::before_dom_children.
8393 (execute_strength_reduction): Adjust.
8394 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
8395 (sese_dom_walker): New class.
8396 (sese_dom_walker::sese_dom_walker): New constructor.
8397 (sese_dom_walker::~sese_dom_walker): New destructor.
8398 (build_sese_conditions_before): Convert to
8399 sese_dom_walker::before_dom_children.
8400 (build_sese_conditions_after): Convert to
8401 sese_dom_walker::after_dom_children.
8402 (build_poly_scop): Adjust
8403 * tree-into-ssa.c (rewrite_dom_walker): New class
8404 (rewrite_enter_block): Convert to
8405 rewrite_dom_walker::before_dom_children.
8406 (rewrite_leave_block): Convert to
8407 rewrite_dom_walker::after_dom_children.
8408 (rewrite_update_dom_walker): New class.
8409 (rewrite_update_enter_block): Convert to
8410 rewrite_update_dom_walker::before_dom_children.
8411 (rewrite_update_leave_block): Convert to
8412 rewrite_update_dom_walker::after_dom_children.
8413 (rewrite_blocks, rewrite_into_ssa): Adjust.
8414 (mark_def_dom_walker): New class.
8415 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
8416 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
8417 (mark_def_sites_blocks): Convert to
8418 mark_def_dom_walker::before_dom_children.
8419 (mark_def_site_blocks): Remove.
8420 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
8421 (tree_ssa_dominator_optimize): Adjust.
8422 (dom_thread_across_edge): Convert to method
8423 dom_opt_dom_walker::thread_across_edge.
8424 (dom_opt_enter_block): Convert to member function
8425 dom_opt_dom_walker::before_dom_children.
8426 (dom_opt_leave_block): Convert to member function
8427 dom_opt_dom_walker::after_dom_children.
8428 * tree-ssa-dse.c (dse_dom_walker): New class.
8429 (dse_enter_block): Convert to member function
8430 dse_dom_walker::before_dom_children.
8431 (tree_ssa_dse): Adjust.
8432 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
8433 (determine_invariantness_stmt): Convert to method
8434 invariantness_dom_walker::before_dom_children.
8435 (determine_invariantness): Remove
8436 (move_computations_dom_walker): New class.
8437 (move_computations_stmt): Convert to method
8438 move_computations_dom_walker::before_dom_children.
8439 (move_computations, tree_ssa_lim): Adjust.
8440 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
8441 (nt_init_block): Convert to method
8442 notrappping_dom_walker::before_dom_children.
8443 (nt_fini_block): Convert to method
8444 method nontrapping_dom_walker::after_dom_children.
8445 (get_non_trapping): Adjust.
8446 * tree-ssa-pre.c (eliminate_dom_walker): New class.
8447 (eliminate_bb): Convert to method
8448 eliminate_dom_walker::before_dom_children.
8449 (eliminate_leave_block): Convert to method
8450 eliminate_dom_walker::after_dom_children.
8451 (eliminate): Adjust.
8452 * tree-ssa-strlen.c (strlen_dom_walker): New class.
8453 (strlen_enter_block): Convert to method
8454 strlen_dom_walker::before_dom_children.
8455 (strlen_leave_block): Convert to method
8456 method strlen_dom_walker::after_dom_children.
8457 (tree_ssa_strlen): Adjust.
8458 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
8459 (tree_ssa_uncprop): Adjust.
8460 (uncprop_leave_block): Convert to method
8461 uncprop_dom_walker::after_dom_children.
8462 (uncprop_leave_block): Convert to method
8463 uncprop_dom_walker::before_dom_children.
8465 2013-09-18 Bin Cheng <bin.cheng@arm.com>
8467 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
8468 branch in same basic block. Check both src and dest of the move insn.
8470 2013-09-17 Nick Clifton <nickc@redhat.com>
8472 * config/rl78/rl78-real.md (bf): New pattern.
8474 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
8475 (rl78_print_operand): Do not put a # before a %B.
8476 * config/rl78/rl78.opt: Tweak doc strings.
8478 2013-09-17 DJ Delorie <dj@redhat.com>
8480 * config/rl78/constraints.md (Wcv): Allow up to $r31.
8481 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
8482 (rl78_option_override): Likewise, if -mallregs.
8483 (is_virtual_register): Likewise.
8484 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
8485 (REGNO_OK_FOR_BASE_P): Likewise.
8486 * config/rl78/rl78.opt (-mallregs): New.
8488 2013-09-17 Nick Clifton <nickc@redhat.com>
8490 * config/rl78/rl78.c (need_to_save): Change return type to bool.
8491 For interrupt functions: save all call clobbered registers if the
8492 interrupt handler is not a leaf function.
8493 (rl78_expand_prologue): Always recompute the frame information.
8494 For interrupt functions: only select bank 0 if one of the bank 0
8495 registers is going to be psuhed.
8497 2013-09-17 DJ Delorie <dj@redhat.com>
8499 * config/rl78/constraints.md: For each W* constraint, rename to C*
8500 and create a W* constraint that checks for an optional ES: prefix
8502 * config/rl78/rl78.md (UNS_ES_ADDR): New.
8503 (es_addr): New. Used to wrap far addresses.
8504 * config/rl78/rl78-protos.h (rl78_es_addr): New.
8505 (rl78_es_base): New.
8506 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
8507 wrapped far addresses.
8508 (rl78_print_operand_1): Unwrap far addresses before processing.
8509 (rl78_lo16): Wrap far addresses in unspecs.
8510 (rl78_es_addr): New.
8511 (rl78_es_base): New.
8512 (insn_ok_now): Check for not-yet-wrapped far addresses.
8513 (transcode_memory_rtx): Properly re-wrap far addresses.
8515 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
8517 * config/sparc/t-rtems: Add leon3 multilibs.
8519 2013-09-17 Cong Hou <congh@google.com>
8521 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
8522 when checking the dot production pattern. The type of rhs operand
8523 of multiply is now checked correctly.
8525 2013-09-17 Jeff Law <law@redhat.com>
8527 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8528 edge implied equivalences into successor phis.
8529 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
8530 here from tree-ssa-threadedge.c.
8531 (mark_threaded_blocks): When threading through a joiner, if both
8532 successors of the joiner's clone reach the same block, verify the
8533 PHI arguments are equal. If not, cancel the jump threading request.
8534 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
8535 tree-ssa-threadupdate.c
8536 (thread_across_edge): Don't check PHI argument equality when
8537 threading through joiner block here.
8539 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
8541 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
8542 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
8543 (warn_uninit, warn_uninitialized_vars,
8544 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8545 Move to tree-ssa-uninit.c.
8546 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
8547 (has_undefined_value_p): New. Pass dependant parts of
8548 ssa_undefined_value_p.
8549 (uninit_undefined_value_p): Use has_undefined_value_p.
8550 (warn_uninit, warn_uninitialized_vars,
8551 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8552 Move from tree-ssa.c.
8553 * tree-ssa.h: Adjust prototypes.
8555 2013-09-17 Jan Hubicka <jh@suse.cz>
8558 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
8559 * ipa-inline.c (can_inline_edge_p): Do not downgrade
8560 FUNCTION_NOT_OPTIMIZED.
8561 * ipa-inline-analysis.c (compute_inline_parameters): Function
8562 not optimized is not inlinable unless it is alwaysinline.
8563 (inline_analyze_function): Force calls in not optimized
8564 function not inlinable.
8566 2013-09-17 Jan Hubicka <jh@suse.cz>
8569 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
8571 * ipa.c (function_and_variable_visibility): Likewise.
8572 * ipa-profile.c (ipa_profile): Likewise.
8574 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
8577 * cgraph.c (cgraph_function_body_availability): Check for ifunc
8578 attribute, and don't inline the resolver in this case.
8580 2013-09-17 Teresa Johnson <tejohnson@google.com>
8582 * coverage.c (get_coverage_counts): Add missing newline.
8584 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8586 PR tree-optimization/58088
8587 * fold-const.c (mask_with_trailing_zeros): New function.
8588 (fold_binary_loc): Make sure we don't recurse infinitely
8589 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
8590 Use mask_with_trailing_zeros where appropriate.
8592 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
8594 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
8595 of 'prev' var to get better distance estimation.
8597 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
8599 * tree-inline.h (struct copy_body_data): Add transform_parameter.
8600 * tree-inline.c (is_parameter_of): New predicate.
8601 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
8602 a parameter has been remapped.
8603 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
8604 (optimize_inline_calls): Initialize transform_parameter.
8605 (copy_gimple_seq_and_replace_locals): Likewise.
8606 (tree_function_versioning): Likewise.
8607 (maybe_inline_call_in_expr): Likewise.
8609 2013-09-17 Nick Clifton <nickc@redhat.com>
8611 * config/msp430/msp430-protos.h: Add prototypes for new functions.
8612 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
8614 (is_attr_func): New function.
8615 (msp430_is_interrupt_func): New function.
8616 (is_naked_func): New function.
8617 (is_reentrant_func): New function.
8618 (is_critical_func): New function.
8619 (msp430_start_function): Add annotations for function attributes.
8620 (msp430_attr): New function.
8621 (msp430_attribute_table): New.
8622 (msp430_function_section): New function.
8623 (TARGET_ASM_FUNCTION_SECTION): Define.
8624 (msp430_builtin): New enum.
8625 (msp430_init_builtins): New function.
8626 (msp430_builtin_devl): New function.
8627 (msp430_expand_builtin): New function.
8628 (TARGET_INIT_BUILTINS): Define.
8629 (TARGET_EXPAND_BUILTINS): Define.
8630 (TARGET_BUILTIN_DECL): Define.
8631 (msp430_expand_prologue): Add support for naked, interrupt,
8632 critical and reentrant functions.
8633 (msp430_expand_epilogue): Likewise.
8634 (msp430_print_operand): Handle 'O' character.
8635 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
8637 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
8638 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
8639 (pushm): Use a 'n' rather than an 'i' constraint.
8640 (msp_return): Add generation of the interrupt return instruction.
8641 (disable_interrupts): New pattern.
8642 (enable_interrupts): New pattern.
8643 (push_intr_state): New pattern.
8644 (pop_intr_state): New pattern.
8645 (bic_SR): New pattern.
8646 (bis_SR): New pattern.
8647 * doc/extend.texi: Document MSP430 function attributes and builtin
8650 2013-09-17 Richard Biener <rguenther@suse.de>
8652 PR tree-optimization/58432
8653 * tree-loop-distribution.c (tree_loop_distribution): Also
8654 scan PHIs for outside loop uses and seed a partition from them.
8656 2013-09-17 Bin Cheng <bin.cheng@arm.com>
8658 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
8659 (restructure_reference): Call backtrace_base_for_ref.
8661 2013-09-17 Alan Modra <amodra@gmail.com>
8664 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
8667 2013-09-16 DJ Delorie <dj@redhat.com>
8669 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
8670 vregs location for RL78/G10.
8671 (rl78_expand_prologue): Avoid SEL on G10.
8672 (rl78_expand_epilogue): Likewise.
8673 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
8674 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
8675 __RL78_G10__ when appropriate.
8676 (ASM_SPEC): Pass -mg10 along to the assembler.
8677 * config/rl78/rl78.md (sel_rb): Disable for G10.
8678 * config/rl78/rl78.opt: Add -mg10 option.
8679 * config/rl78/t-rl78: Add -mg10 multilib.
8681 2013-09-16 Xinliang David Li <davidxl@google.com>
8683 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
8684 * omp-low.c (omp_max_vf): Ditto.
8685 (expand_omp_simd): Ditto.
8686 * tree-vectorizer.c (vectorize_loops): Ditto.
8687 (gate_vect_slp): Ditto.
8688 (gate_increase_alignment): Ditto.
8689 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
8690 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
8691 (gate_tree_vectorize): Name change.
8692 (tree_vectorize): Ditto.
8693 (pass_vectorize::gate): Call new function.
8694 (pass_vectorize::execute): Ditto.
8695 * opts.c: O3 default setting change.
8696 (finish_options): Check new flag.
8697 * doc/invoke.texi: Document new flags.
8698 * common.opt: New flags.
8700 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
8702 * doc/tm.texi.in (Cond Exec Macros): Remove node.
8703 (Condition Code): Don't reference it.
8704 * doc/tm.texi: Regenerate.
8706 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
8709 * lra-constraints.c (undo_optional_reloads): Consider all optional
8710 reload even if it did not get a hard reg.
8712 2013-09-16 Teresa Johnson <tejohnson@google.com>
8714 * dumpfile.c (dump_loc): Remove newline emission.
8715 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
8716 emission to dump_printf_loc calls where missing.
8717 (vect_mark_for_runtime_alias_test): Ditto.
8718 (vect_analyze_data_ref_dependence): Ditto.
8719 (vect_analyze_data_ref_dependences): Ditto.
8720 (vect_slp_analyze_data_ref_dependence): Ditto.
8721 (vect_slp_analyze_data_ref_dependences): Ditto.
8722 (vect_compute_data_ref_alignment): Ditto.
8723 (vect_update_misalignment_for_peel): Ditto.
8724 (vect_verify_datarefs_alignment): Ditto.
8725 (vector_alignment_reachable_p): Ditto.
8726 (vect_get_data_access_cost): Ditto.
8727 (vect_enhance_data_refs_alignment): Ditto.
8728 (vect_find_same_alignment_drs): Ditto.
8729 (vect_analyze_data_refs_alignment): Ditto.
8730 (vect_analyze_group_access): Ditto.
8731 (vect_analyze_data_ref_access): Ditto.
8732 (vect_analyze_data_ref_accesses): Ditto.
8733 (vect_prune_runtime_alias_test_list): Ditto.
8734 (vect_analyze_data_refs): Ditto.
8735 (vect_create_addr_base_for_vector_ref): Ditto.
8736 (vect_create_data_ref_ptr): Ditto.
8737 (vect_grouped_store_supported): Ditto.
8738 (vect_grouped_load_supported): Ditto.
8739 * value-prof.c (check_counter): Ditto.
8740 (check_ic_target): Ditto.
8741 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
8742 (vect_recog_widen_mult_pattern): Ditto.
8743 (vect_recog_widen_sum_pattern): Ditto.
8744 (vect_recog_over_widening_pattern): Ditto.
8745 (vect_recog_widen_shift_pattern): Ditto.
8746 (vect_recog_rotate_pattern): Ditto.
8747 (vect_recog_vector_vector_shift_pattern): Ditto.
8748 (vect_recog_divmod_pattern): Ditto.
8749 (vect_recog_mixed_size_cond_pattern): Ditto.
8750 (vect_recog_bool_pattern): Ditto.
8751 (vect_pattern_recog_1): Ditto.
8752 (vect_pattern_recog): Ditto.
8753 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
8754 (vect_is_simple_iv_evolution): Ditto.
8755 (vect_analyze_scalar_cycles_1): Ditto.
8756 (vect_get_loop_niters): Ditto.
8757 (vect_analyze_loop_1): Ditto.
8758 (vect_analyze_loop_form): Ditto.
8759 (vect_analyze_loop_operations): Ditto.
8760 (vect_analyze_loop_2): Ditto.
8761 (vect_analyze_loop): Ditto.
8762 (report_vect_op): Ditto.
8763 (vect_is_slp_reduction): Ditto.
8764 (vect_is_simple_reduction_1): Ditto.
8765 (vect_get_known_peeling_cost): Ditto.
8766 (vect_estimate_min_profitable_iters): Ditto.
8767 (vect_model_reduction_cost): Ditto.
8768 (vect_model_induction_cost): Ditto.
8769 (get_initial_def_for_induction): Ditto.
8770 (vect_create_epilog_for_reduction): Ditto.
8771 (vectorizable_reduction): Ditto.
8772 (vectorizable_induction): Ditto.
8773 (vectorizable_live_operation): Ditto.
8774 (vect_loop_kill_debug_uses): Ditto.
8775 (vect_transform_loop): Ditto.
8776 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
8777 (vect_stmt_relevant_p): Ditto.
8778 (process_use): Ditto.
8779 (vect_mark_stmts_to_be_vectorized): Ditto.
8780 (vect_model_simple_cost): Ditto.
8781 (vect_model_promotion_demotion_cost): Ditto.
8782 (vect_model_store_cost): Ditto.
8783 (vect_get_store_cost): Ditto.
8784 (vect_model_load_cost): Ditto.
8785 (vect_get_load_cost): Ditto.
8786 (vect_init_vector_1): Ditto.
8787 (vect_get_vec_def_for_operand): Ditto.
8788 (vect_finish_stmt_generation): Ditto.
8789 (vectorizable_call): Ditto.
8790 (vectorizable_conversion): Ditto.
8791 (vectorizable_assignment): Ditto.
8792 (vectorizable_shift): Ditto.
8793 (vectorizable_operation): Ditto.
8794 (vectorizable_store): Ditto.
8795 (vectorizable_load): Ditto.
8796 (vectorizable_condition): Ditto.
8797 (vect_analyze_stmt): Ditto.
8798 (vect_transform_stmt): Ditto.
8799 (vect_is_simple_use): Ditto.
8800 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
8801 (vect_can_advance_ivs_p): Ditto.
8802 (vect_update_ivs_after_vectorizer): Ditto.
8803 (vect_do_peeling_for_loop_bound): Ditto.
8804 (vect_gen_niters_for_prolog_loop): Ditto.
8805 (vect_update_inits_of_drs): Ditto.
8806 (vect_create_cond_for_alias_checks): Ditto.
8807 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
8808 (vect_build_slp_tree_1): Ditto.
8809 (vect_supported_load_permutation_p): Ditto.
8810 (vect_analyze_slp_instance): Ditto.
8811 (vect_analyze_slp): Ditto.
8812 (vect_make_slp_decision): Ditto.
8813 (vect_detect_hybrid_slp): Ditto.
8814 (vect_bb_vectorization_profitable_p): Ditto.
8815 (vect_slp_analyze_bb_1): Ditto.
8816 (vect_update_slp_costs_according_to_vf): Ditto.
8817 (vect_get_mask_element): Ditto.
8818 (vect_transform_slp_perm_load): Ditto.
8819 (vect_schedule_slp_instance): Ditto.
8820 (vect_schedule_slp): Ditto.
8821 (vect_slp_transform_bb): Ditto.
8822 * profile.c (read_profile_edge_counts): Ditto.
8823 (compute_branch_probabilities): Ditto.
8824 * coverage.c (get_coverage_counts): Ditto.
8826 2013-09-16 Diego Novillo <dnovillo@google.com>
8828 * tree-core.h: Add missing comment lines from refactoring of tree.h.
8830 2013-09-16 Jan Hubicka <jh@suse.cz>
8832 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
8833 abstract functions; for static functions check the presence of body.
8835 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8837 * config/aarch64/aarch64-simd-builtins.def (fma): New.
8838 * config/aarch64/aarch64-simd.md
8839 (aarch64_mla_elt<mode>): New.
8840 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8841 (aarch64_mls_elt<mode>): Likewise.
8842 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8843 (aarch64_fma4_elt<mode>): Likewise.
8844 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8845 (aarch64_fma4_elt_to_128v2df): Likewise.
8846 (aarch64_fma4_elt_to_64df): Likewise.
8847 (fnma<mode>4): Likewise.
8848 (aarch64_fnma4_elt<mode>): Likewise.
8849 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8850 (aarch64_fnma4_elt_to_128v2df): Likewise.
8851 (aarch64_fnma4_elt_to_64df): Likewise.
8852 * config/aarch64/iterators.md (VDQSF): New.
8853 * config/aarch64/arm_neon.h
8854 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
8855 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8857 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8859 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8860 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8861 (aarch64_mul3_elt_to_128df): Likewise.
8862 (aarch64_mul3_elt_to_64v2df): Likewise.
8863 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8865 (VMUL_CHANGE_NLANES) Likewise.
8868 * config/aarch64/arm_neon.h
8869 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8871 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8873 * config/aarch64/arm_neon.h
8874 (vcvtx_high_f32_f64): Fix parameters.
8876 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8877 Uros Bizjak <ubizjak@gmail.com>
8879 * config/alpha.c: Include tree-ssanames.h.
8881 2013-09-16 Richard Biener <rguenther@suse.de>
8883 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8884 (dot_rdg_1): Handle control_dd.
8885 (create_edge_for_control_dependence): New function.
8886 (create_rdg_edges): Add control dependences if asked for.
8887 (build_rdg): Likewise.
8888 (generate_loops_for_partition): If there are not necessary
8889 control stmts remove all their dependencies.
8890 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8891 (distribute_loop): Pass on control dependences.
8892 (tree_loop_distribution): Compute control dependences and remove
8893 restriction on number of loop nodes.
8895 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8897 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8900 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8902 * cse.c (try_const_anchors): Punt on CC modes.
8904 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8906 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8908 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8911 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8914 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8917 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8919 (darwin_objc1_section): Likewise.
8920 (darwin_file_end): Emit Image Info section when required.
8922 2013-09-14 Jan Hubicka <jh@suse.cz>
8924 * tree-into-ssa.c (gate_into_ssa): New.
8925 (pass_data_build_ssa): Use it.
8926 * cgraph.h (expand_thunk): Update prototype.
8927 * cgraphunit.c (analyze_function): Expand thunks early.
8928 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8929 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8930 set lowered flag; do not add new function.
8931 (assemble_thunks_and_aliases): Update.
8932 * tree-ssa.c (gate_init_datastructures): New gate.
8933 (pass_data_init_datastructures): Use it.
8935 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8938 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8939 xmm register set described in the psABI.
8941 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8943 * dwarf2out.c (should_emit_struct_debug): Add check
8944 for type_decl variable is not NULL.
8946 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8948 * config.gcc: Use new winnt-c.c target hooks
8949 * config/t-winnt: New file
8950 * config/winnt-c.c: New file
8951 * doc/tm.texi.in: Document new hook
8952 * doc/tm.texi: Regenerated
8954 2013-09-13 Jan Hubicka <jh@suse.cz>
8957 * ipa-inline.c (check_callers): New function.
8958 (check_caller_edge): Remove.
8959 (want_inline_function_to_all_callers_p): Also permit alises that are
8961 (inline_to_all_callers): Terminate the walk when devirtualization
8962 introduce new calls.
8964 2013-09-13 Jan Hubicka <jh@suse.cz>
8966 * ipa-inline-analysis.c (struct growth_data): Add node.
8967 (do_estimate_growth_1): Fix detection of recursion.
8969 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8971 PR tree-optimization/58392
8972 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8973 to avoid shadowing of outer loop variable. If
8974 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8975 replace_by_duplicate_decl simduid of loops that have it set and
8976 set dest_cfun->has_simduid_loops and/or
8977 dest_cfun->has_force_vect_loops.
8978 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8979 instead of maybe_lookup_decl.
8980 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8981 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8982 call. Copy over force_vect and copy and remap simduid. Set
8983 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8984 (copy_cfg_body): Remove blocks_to_copy argument. Use
8985 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8986 caller. Don't set cfun->has_simduid_loops and/or
8987 cfun->has_force_vect_loops here.
8988 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8990 (expand_call_inline, tree_function_versioning): Adjust copy_body
8993 2013-09-13 Martin Jambor <mjambor@suse.cz>
8996 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8997 the assert if the edge was a speculative one.
8999 2013-09-13 Richard Biener <rguenther@suse.de>
9001 * tree-data-ref.h (known_dependences_p): Move here ...
9002 * tree-loop-distribution.c (known_dependences_p): ... from here.
9003 (dump_rdg_component, debug_rdg_component): Remove.
9005 (generate_loops_for_partition): Use gimple_uid instead of
9006 relying on matching stmt visit order.
9007 (rdg_build_partitions): Take starting stmt vector.
9008 (ldist_gen): Merge into ...
9009 (distribute_loop): ... this function. Do not compute starting
9011 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
9013 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9015 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
9016 Set type attribute correctly. Set predicable_short_it attribute.
9017 (cmpsi_shiftsi): Remove %? from output template.
9019 2013-09-13 Richard Biener <rguenther@suse.de>
9021 * tree-loop-distribution.c (struct rdg_component,
9022 rdg_defs_used_in_other_loops_p, free_rdg_components,
9023 rdg_build_components): Remove.
9024 (stmts_from_loop): Do not record virtual PHIs.
9025 (generate_loops_for_partition): Skip virtual PHIs.
9026 (build_rdg_partition_for_component): Rename to ...
9027 (build_rdg_partition_for_vertex): ... this and adjust.
9028 (rdg_build_partitions): Take a vector of starting vertices
9029 instead of components. Remove unnecessary leftover handling.
9030 (ldist_gen): Do not build components or record other stores.
9031 (distribute_loop): Do not distribute loops containing stmts
9034 2013-09-13 Christian Bruel <christian.bruel@st.com>
9037 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
9039 2013-09-13 Kai Tietz <ktietz@redhat.com>
9041 * config.gcc: Separate cases for mingw and cygwin targets,
9042 and add 64-bit cygwin target case.
9044 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
9045 dll-export inline-functions.
9046 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
9048 2013-09-13 Jeff Law <law@redhat.com>
9052 2013-09-06 Jeff Law <law@redhat.com>
9054 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9055 edge implied equivalences into successor phis.
9057 2013-09-12 DJ Delorie <dj@redhat.com>
9059 * config/rl78/rl78-virt.md: Change from | to \; for asm line
9062 2013-09-12 Brooks Moses <bmoses@google.com>
9065 * Makefile.in: Do not install driver binaries in $(target)/bin.
9067 2013-09-12 DJ Delorie <dj@redhat.com>
9069 * config/rl78/rl78.opt (mrelax): New.
9070 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
9071 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
9073 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
9075 (rl78_expand_epilogue): Likewise.
9077 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
9080 * lra-eliminations.c (remove_reg_equal_offset_note): New.
9081 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
9082 pointer elimination with using remove_reg_equal_offset_note.
9084 2013-09-12 DJ Delorie <dj@redhat.com>
9086 * config/msp430/: New port.
9087 * config.gcc (msp430): Added.
9088 * doc/invoke.texi: Document MSP430 options.
9089 * doc/install.texi: Document msp430-elf
9090 * doc/md.texi: Document msp430-elf
9091 * doc/contrib.texi: Document msp430-elf
9093 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
9096 2013-09-12 Martin Jambor <mjambor@suse.cz>
9099 * ipa-prop.c (remove_described_reference): Give up if the edge in the
9100 reference descriptor is NULL.
9101 (ipa_edge_removal_hook): If owning a reference descriptor, set its
9104 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
9106 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
9107 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
9108 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
9109 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
9110 make_temp_ssa_name): move to tree-ssanames.h
9111 * tree-ssa-alias.h: Move prototype.
9112 * tree-ssa.h: Include tree-ssanames.h.
9113 * tree-ssanames.c (FREE_SSANAMES): Move to here.
9114 * tree-ssanames.h: New. Move items from tree-flow*.h
9115 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
9117 2013-09-12 Richard Biener <rguenther@suse.de>
9119 PR tree-optimization/58404
9120 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
9121 propagate non-invariant addresses into dereferences wrapped
9122 in component references.
9124 2013-09-12 Richard Biener <rguenther@suse.de>
9126 PR tree-optimization/58402
9127 * passes.def: Move pass_late_warn_uninitialized later.
9129 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
9131 * tree-ssa.h: New. Move content from tree-flow.h and
9133 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
9134 Move prototypes belonging to tree-ssa.c.
9135 * tree-flow-inline.h (redirect_edge_var_map_def,
9136 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
9138 * gimple.h: Adjust prototypes.
9139 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
9141 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
9142 * tree.h: Move prototype to tree-ssa.h.
9143 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
9144 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
9145 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
9146 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
9147 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
9148 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
9149 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
9150 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
9151 graphite-blocking.c, graphite-clast-to-gimple.c,
9152 graphite-dependences.c, graphite-interchange.c,
9153 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
9154 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
9155 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
9156 ipa-reference.c, ipa-split.c, ipa-utils.c,
9157 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
9158 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
9159 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
9160 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
9161 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
9162 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
9163 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
9164 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
9165 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
9166 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
9167 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
9168 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
9169 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
9170 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
9171 value-prof.c, var-tracking.c,
9172 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
9174 2013-09-12 Richard Biener <rguenther@suse.de>
9176 PR tree-optimization/58396
9177 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
9178 (build_rdg): Take a loop-nest parameter, fix memleaks.
9179 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
9181 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9183 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
9186 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
9188 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
9189 The FMA instruction names should have a 'v' prefix.
9191 2013-09-12 Richard Biener <rguenther@suse.de>
9193 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
9194 (dot_rdg): Use popen instead of system in optional code.
9195 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
9196 (already_processed_vertex_p): Adjust.
9197 (has_anti_or_output_dependence, predecessor_has_mem_write,
9198 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
9199 rdg_flag_uses): Remove.
9200 (rdg_flag_vertex): Simplify.
9201 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
9203 (build_rdg_partition_for_component): Process the first vertex
9204 of a component only.
9205 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
9207 2013-09-12 Alan Modra <amodra@gmail.com>
9209 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
9211 2013-09-11 DJ Delorie <dj@redhat.com>
9212 Nick Clifton <nickc@redhat.com>
9214 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
9215 (rl78_stack_based_mem): New.
9216 * config/rl78/constraints.md (Iv08): New.
9225 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
9227 (movsi): Change from expand to insn-and-split.
9228 (ashrsi3): Clobber AX.
9232 * config/rl78/rl78.md (CC_REG): Fix.
9233 (addsi3): Allow memory and immediate operands.
9234 (addsi3_internal): Split into...
9235 (addsi3_internal_virt): ...new, and ...
9236 (addsi3_internal_real): ...new.
9238 (subsi3_internal_virt): New.
9239 (subsi3_internal_real): New.
9240 (mulsi3): Add memory operand.
9241 (mulsi3_rl78): Likewise.
9242 (mulsi3_g13): Likewise.
9243 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
9244 (cbranchqi4_real): Add more constraint options.
9245 (cbranchhi4_real): Expand pattern.
9246 (cbranchhi4_real_signed): New.
9247 (cbranchhi4_real_inverted): New.
9248 (cbranchsi4_real_lt): New.
9249 (cbranchsi4_real_ge): New.
9250 (cbranchsi4_real_signed): New.
9251 (cbranchsi4_real): New.
9253 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
9255 (lshrsi3_virt): Likewise.
9256 (ashlsi3_virt): Likewise.
9257 (cbranchqi4_virt_signed): New.
9258 (cbranchhi4_virt_signed): New.
9259 (cbranchsi4_virt): New.
9260 * config/rl78/rl78.c: Whitespace fixes throughout.
9261 (move_elim_pass): New.
9262 (pass_data_rl78_move_elim): New.
9263 (pass_rl78_move_elim): New.
9264 (make_pass_rl78_move_elim): New.
9265 (rl78_devirt_info): Run devirt earlier.
9266 (rl78_move_elim_info): New.
9267 (rl78_asm_file_start): Register it.
9268 (rl78_split_movsi): New.
9269 (rl78_as_legitimate_address): Allow virtual base registers when
9271 (rl78_addr_space_convert): Remove spurious debug stuff.
9272 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
9273 (rl78_print_operand): More cases for not printing '#'.
9274 (rl78_expand_compare): Remove most of the logic.
9275 (content_memory): New.
9276 (clear_content_memory): New.
9277 (get_content_index): New.
9278 (get_content_name): New.
9279 (display_content_memory): New.
9280 (update_content): New.
9281 (record_content): New.
9282 (already_contains): New.
9283 (insn_ok_now): Re-recog insns with virtual registers.
9284 (add_postponed_content_update): New.
9285 (process_postponed_content_update): New.
9286 (gen_and_emit_move): New.
9287 (transcode_memory_rtx): Record new location content.
9288 Use gen_and_emit_move.
9289 (force_into_acc): New.
9290 (move_to_acc): Use gen_and_emit_move.
9291 (move_from_acc): Likewise.
9292 (move_acc_to_reg): Likewise.
9293 (move_to_x): Likewise.
9294 (move_to_hl): Likewise.
9295 (move_to_de): Likewise.
9296 (rl78_alloc_physical_registers_op1): Record location content.
9297 (has_constraint): New.
9298 (rl78_alloc_physical_registers_op2): Record location content.
9300 (rl78_alloc_physical_registers_ro1): Likewise.
9301 (rl78_alloc_physical_registers_cmp): Likewise.
9302 (rl78_alloc_physical_registers_umul): Likewise.
9303 (rl78_alloc_address_registers_macax): New.
9304 (rl78_alloc_physical_registers): Initialize and set location
9305 content memory as needed.
9306 (rl78_reorg): Make sure split2 is called.
9307 (rl78_rtx_costs): New.
9309 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
9311 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
9312 for (not (neg ...)) and (neg (not ...)) cases.
9314 2013-09-11 Richard Biener <rguenther@suse.de>
9317 * passes.def: Split critical edges before late uninit warning passes.
9318 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
9320 2013-09-11 Jakub Jelinek <jakub@redhat.com>
9322 PR tree-optimization/58385
9323 * fold-const.c (build_range_check): If both low and high are NULL,
9324 use omit_one_operand_loc to preserve exp side-effects.
9326 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9328 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
9330 2013-09-11 Richard Biener <rguenther@suse.de>
9332 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
9333 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
9334 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
9335 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
9336 stmts_from_loop, known_dependences_p, build_empty_rdg,
9337 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
9338 * tree-loop-distribution.c: ... here.
9339 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
9340 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
9341 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
9342 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
9343 * tree-loop-distribution.c: ... here.
9344 * tree-loop-distribution.c: Include gimple-pretty-print.h.
9345 (struct partition_s): Add loops member.
9346 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
9347 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
9348 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
9350 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
9351 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
9352 Sergey Lega <sergey.s.lega@intel.com>
9353 Anna Tikhonova <anna.tikhonova@intel.com>
9354 Ilya Tocar <ilya.tocar@intel.com>
9355 Andrey Turetskiy <andrey.turetskiy@intel.com>
9356 Ilya Verbin <ilya.verbin@intel.com>
9357 Kirill Yukhin <kirill.yukhin@intel.com>
9358 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9360 * config/i386/constraints.md (k): New.
9362 * config/i386/i386.c (const regclass_map): Add new mask registers.
9363 (dbx_register_map): Ditto.
9364 (dbx64_register_map): Ditto.
9365 (svr4_dbx_register_map): Ditto.
9366 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
9368 (ix86_preferred_reload_class): Disable constants for mask registers.
9369 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
9370 (ix86_hard_regno_mode_ok): Support new mask registers.
9371 (x86_order_regs_for_local_alloc): Ditto.
9372 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
9373 (FIXED_REGISTERS): Add new mask registers.
9374 (CALL_USED_REGISTERS): Ditto.
9375 (REG_ALLOC_ORDER): Ditto.
9376 (VALID_MASK_REG_MODE): New.
9377 (FIRST_MASK_REG): Ditto.
9378 (LAST_MASK_REG): Ditto.
9379 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
9380 (MAYBE_MASK_CLASS_P): New.
9381 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
9382 (REG_CLASS_CONTENTS): Ditto.
9383 (MASK_REGNO_P): New.
9384 (ANY_MASK_REG_P): Ditto.
9385 (HI_REGISTER_NAMES): Add new mask registers.
9386 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
9387 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
9389 (attribute "type"): Add mskmov, msklog.
9390 (attribute "length_immediate"): Support them.
9391 (attribute "memory"): Ditto.
9392 (attribute "prefix_0f"): Ditto.
9393 (*movhi_internal): Support new mask registers.
9394 (*movqi_internal): Ditto.
9395 (define_split): Split out clobber pattern is a logic
9396 insn on mask registers.
9397 (*k<logic><mode>): New.
9398 (*andhi_1): Extend to support mask regs.
9399 (*andqi_1): Extend to support mask regs.
9401 (define_split): Split and-not to and and not if operands
9403 (*<code><mode>_1): Separate HI mode to new pattern...
9404 (*<code>hi_1): This.
9405 (*<code>qi_1): Extend to support mask regs.
9407 (kortestzhi): Ditto.
9408 (kortestchi): Ditto.
9410 (*one_cmpl<mode>2_1): Remove HImode and handle it...
9411 (*one_cmplhi2_1): ...Here, now with mask registers support.
9412 (*one_cmplqi2_1): Support new mask registers.
9413 (HI/QImode arithmetics splitter): Don't split if mask registers
9415 (HI/QImode not splitter): Ditto.
9416 * config/i386/predicated.md (mask_reg_operand): New.
9417 (general_reg_operand): Ditto.
9419 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
9421 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
9422 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
9424 2013-09-10 Jeff Law <law@redhat.com>
9426 PR tree-optimization/58380
9427 * tree-ssa-threadupdate.c (thread_block): Recognize another case
9428 of threading through a buried loop header.
9430 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
9431 return value for single successor case.
9433 2013-09-10 Jan Hubicka <jh@suse.cz>
9435 * ipa-devirt.c (ipa_devirt): Enable with LTO.
9437 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
9440 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
9441 support conditional execution.
9442 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
9444 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
9446 * lra.c (lra): Clear lra_optional_reload_pseudos before every
9448 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
9449 Check destination too to check move insn.
9450 (undo_optional_reloads): Add check that the original peudo did not
9451 changed its allocation and the optional reload was inherited on last
9452 inheritance pass. Break loop after deciding to keep optional reload.
9453 (lra_undo_inheritance): Add check that inherited pseudo still in
9456 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
9458 * config/aarch64/aarch64.md (generic_sched): New.
9459 * config/aarch64/aarch64-generic.md (load): Make conditional
9460 on generic_sched attribute.
9461 (nonload): Likewise.
9463 2013-09-10 Jan Hubicka <jh@suse.cz>
9465 * lto-cgraph.c: Include ipa-utils.h.
9466 (compute_ltrans_boundary): Also add possible targets into the boundary.
9468 2013-09-10 Jan Hubicka <jh@suse.cz>
9470 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
9471 VAR_DECL of vtable rather than full expression.
9473 2013-09-10 Jan Hubicka <jh@suse.cz>
9474 Paolo Carlini <paolo.carlini@oracle.com>
9476 * cgraphunit.c (analyze_functions): Save input_location, set it
9477 to UNKNOWN_LOCATION and restore it at the end.
9479 2013-09-10 Martin Jambor <mjambor@suse.cz>
9481 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
9482 represented by a thunk.
9484 2013-09-10 Jeff Law <law@redhat.com>
9486 PR tree-optimization/58343
9487 * tree-ssa-threadupdate.c (thread_block): Identify and disable
9488 jump threading requests through loop headers buried in the middle
9489 of a jump threading path.
9491 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
9492 in return value/type.
9494 2013-09-10 Jakub Jelinek <jakub@redhat.com>
9496 PR rtl-optimization/58365
9497 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
9498 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
9501 2013-09-10 Richard Biener <rguenther@suse.de>
9503 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
9504 * tree-data-ref.c (create_rdg_vertices): Collect all data
9505 references, signal failure to the caller, use data-ref API.
9506 (build_rdg): Compute data references only once. Maintain lifetime
9507 of data references and data dependences from within RDG.
9508 (free_rdg): Free dependence relations.
9509 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
9510 inventing extra dependences.
9511 (distribute_loop): Update for RDG API changes.
9513 2013-09-10 Kai Tietz <ktietz@redhat.com>
9515 * doc/invoke.texi (fms-extensions): Document changed
9516 behavior for ms-abi targets.
9517 * config/i386/i386.c (ix86_option_override_internal):
9518 Set default value of option -fms-extension for ms-abi targets.
9520 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9522 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
9524 2013-09-10 Alan Modra <amodra@gmail.com>
9527 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
9529 2013-09-10 Alan Modra <amodra@gmail.com>
9531 * config/rs6000/predicates.md (add_cint_operand): New.
9532 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
9533 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
9534 using add_cint_operand.
9535 (largetoc_high_plus_aix): Likewise.
9537 2013-09-09 Jakub Jelinek <jakub@redhat.com>
9539 PR tree-optimization/58364
9540 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
9541 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
9542 the current range can't be an unconditional true or false.
9544 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
9546 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
9548 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
9550 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
9552 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
9555 * doc/invoke.texi (-Wdelete-incomplete): Document it.
9557 2013-09-09 Ian Bolton <ian.bolton@arm.com>
9559 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
9560 NO_REGS for immediate that can't be moved directly into FP_REGS.
9562 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9564 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
9565 comparison with negated operand.
9566 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
9569 2013-09-09 Richard Biener <rguenther@suse.de>
9572 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
9573 of a subloop record all its block as affecting loop-closed SSA form.
9575 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9577 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
9578 of an rtx/bitpos pair.
9579 (store_fixed_bit_field): Update accordingly.
9581 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9583 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
9585 * builtins.c (expand_errno_check): Likewise.
9586 * dwarf2cfi.c (init_return_column_size): Likewise.
9587 * except.c (sjlj_mark_call_sites): Likewise.
9588 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
9589 * lra-constraints.c (emit_inc): Likewise.
9590 * ree.c (combine_set_extension): Likewise.
9591 * regmove.c (fixup_match_2): Likewise.
9592 * reload1.c (inc_for_reload): Likewise.
9594 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9596 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
9597 (force_to_mode, simplify_shift_const_1, simplify_comparison):
9598 Use gen_int_mode with the mode of the associated simplify_* call.
9599 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
9600 * expmed.c (expand_shift_1): Likewise.
9601 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9602 * loop-iv.c (iv_number_of_iterations): Likewise.
9603 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9604 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
9606 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9608 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
9609 of the associated expand_* call.
9610 (asan_emit_stack_protection): Likewise.
9611 * builtins.c (round_trampoline_addr): Likewise.
9612 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
9613 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
9614 (emit_store_flag): Likewise.
9615 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
9616 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
9618 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9619 * ifcvt.c (noce_try_store_flag_constants): Likewise.
9620 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9621 * modulo-sched.c (generate_prolog_epilog): Likewise.
9622 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
9623 (expand_ctz, expand_ffs, expand_unop): Likewise.
9625 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9627 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
9628 of the associated gen_rtx_* call.
9629 * caller-save.c (init_caller_save): Likewise.
9630 * combine.c (find_split_point, make_extraction): Likewise.
9631 (make_compound_operation): Likewise.
9632 * dwarf2out.c (mem_loc_descriptor): Likewise.
9633 * explow.c (plus_constant, probe_stack_range): Likewise.
9634 * expmed.c (expand_mult_const): Likewise.
9635 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
9636 * reload1.c (init_reload): Likewise.
9637 * valtrack.c (cleanup_auto_inc_dec): Likewise.
9638 * var-tracking.c (adjust_mems): Likewise.
9639 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
9640 rather than gen_rtx_fmt_ee.
9642 2013-09-09 Jan Hubicka <jh@suse.cz>
9645 * value-prof.c (gimple_ic): Copy also abnormal edges.
9647 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9649 * asan.c (asan_shadow_cst): Use gen_int_mode.
9651 2013-09-08 Jan Hubicka <jh@suse.cz>
9653 * ipa-profile.c: Add toplevel comment.
9654 (ipa_propagate_frequency_1): Be more conservative when profile is read.
9655 (contains_hot_call_p): New function.
9656 (ipa_propagate_frequency): Set frequencies based on counts when
9658 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
9659 profile; do not tamper with profile after inlining if it is read.
9661 2013-09-08 Jan Hubicka <jh@suse.cz>
9663 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
9666 2013-09-08 Jan Hubicka <jh@suse.cz>
9668 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
9671 2013-09-08 Jeff Law <law@redhat.com>
9674 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
9677 2013-09-08 Andi Kleen <ak@linux.intel.com>
9679 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
9681 2013-09-08 Jan Hubicka <jh@suse.cz>
9683 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
9685 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
9687 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
9688 for non-debug insns.
9689 * lra.c (new_insn_reg): Take the containing insn as a parameter.
9690 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
9691 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
9694 2013-09-08 Jan Hubicka <jh@suse.cz>
9696 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
9697 targets and devirtualize to BUILT_IN_UNREACHABLE.
9698 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
9699 * ipa.c (walk_polymorphic_call_targets): New function.
9700 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
9701 functions; use the new timevar.
9702 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
9703 was removed from the program.
9704 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
9705 not consider it in the walk when its vtable is dead.
9706 (possible_polymorphic_call_targets_1): Pass anonymous flag to
9708 (devirt_variable_node_removal_hook): New function.
9709 (possible_polymorphic_call_targets): Also register
9710 devirt_variable_node_removal_hook.
9711 (ipa_devirt): Do not do non-speculative devirtualization.
9712 (gate_ipa_devirt): One execute if devirtualizing speculatively.
9714 2013-09-08 Jan Hubicka <jh@suse.cz>
9716 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
9717 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
9718 varpool_remove_variable_insertion_hook): Declare.
9719 * varpool.c (varpool_node_hook_list): New structure.
9720 (first_varpool_node_removal_hook,
9721 first_varpool_variable_insertion_hook): New variables.
9722 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
9723 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
9724 varpool_remove_variable_insertion_hook,
9725 varpool_call_variable_insertion_hooks): New functions.
9726 (varpool_remove_node): Use it.
9728 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
9731 * diagnostic.c (diagnostic_build_prefix): When s.file is
9732 "<built-in>" don't output line and column numbers.
9734 2013-09-06 Jan Hubicka <jh@suse.cz>
9736 * cgraphunit.c (expand_thunk): Get body before touching arguments.
9737 * lto-streamer-out.c: Stream thunks, too.
9738 * lto-streamer-in.c (input_function): Pop cfun here
9739 (lto_read_body): Instead of here.
9741 2013-09-06 Caroline Tice <cmtice@google.com>
9743 * doc/install.texi: Add documentation for the --enable-vtable-verify
9744 and the --disable-libvtv configure options.
9746 2013-09-06 Jeff Law <law@redhat.com>
9748 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9749 edge implied equivalences into successor phis.
9751 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
9753 * resource.c (mark_referenced_resources): Handle COND_EXEC.
9755 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
9757 * resource.c (mark_target_live_regs): Compute resources taking
9758 into account if a call is predicated or not.
9760 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9762 * toplev.c (output_stack_usage): Be prepared for suffixes created by
9763 the compiler in the function names.
9765 2013-09-06 Jan Hubicka <jh@suse.cz>
9768 * ipa-inline.c (has_caller_p): New function.
9769 (want_inline_function_to_all_callers_p): Use it.
9770 (sum_callers, inline_to_all_callers): Break out from ...
9771 (ipa_inline): ... here.
9773 2013-09-06 Jan Hubicka <jh@suse.cz>
9775 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
9776 only when AVX is enabled.
9778 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9780 * config/aarch64/aarch64.md
9781 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
9782 is fpsimd_<load/store>2.
9783 (load_pair<mode>): Likewise.
9784 (store_pair<mode>): Likewise.
9786 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9788 * config/arm/types.md (type): Add "mrs" type.
9789 * config/aarch64/aarch64.md
9790 (aarch64_load_tp_hard): Make type "mrs".
9792 (load_tp_hard): Make type "mrs".
9793 * config/arm/cortex-a15.md: Update with new attributes.
9794 * config/arm/cortex-a5.md: Update with new attributes.
9795 * config/arm/cortex-a53.md: Update with new attributes.
9796 * config/arm/cortex-a7.md: Update with new attributes.
9797 * config/arm/cortex-a8.md: Update with new attributes.
9798 * config/arm/cortex-a9.md: Update with new attributes.
9799 * config/arm/cortex-m4.md: Update with new attributes.
9800 * config/arm/cortex-r4.md: Update with new attributes.
9801 * config/arm/fa526.md: Update with new attributes.
9802 * config/arm/fa606te.md: Update with new attributes.
9803 * config/arm/fa626te.md: Update with new attributes.
9804 * config/arm/fa726te.md: Update with new attributes.
9806 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9808 * config/aarch64/aarch64.md
9809 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
9810 (*movtf_aarch64): Likewise.
9812 (thumb1_movdi_insn): Use "multiple" for type where more than one
9813 instruction is used for a move.
9814 (*arm32_movhf): Likewise.
9815 (*thumb_movdf_insn): Likewise.
9817 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9819 * config/arm/types.md (type): Rename fcpys to fmov.
9821 (*arm_movsi_vfp): Rename type fcpys as fmov.
9822 (*thumb2_movsi_vfp): Likewise
9823 (*movhf_vfp_neon): Likewise
9824 (*movhf_vfp): Likewise
9825 (*movsf_vfp): Likewise
9826 (*thumb2_movsf_vfp): Likewise
9827 (*movsfcc_vfp): Likewise
9828 (*thumb2_movsfcc_vfp): Likewise
9829 * config/aarch64/aarch64-simd.md
9830 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
9831 * config/aarch64/aarch64.md
9832 (*movsi_aarch64): Replace type mov_reg with fmovs.
9833 (*movdi_aarch64): Likewise
9834 (*movsf_aarch64): Likewise
9835 (*movdf_aarch64): Likewise
9837 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
9838 * config/arm/iwmmxt.md
9839 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
9840 * config/arm/arm1020e.md: Update with new attributes.
9841 * config/arm/cortex-a15-neon.md: Update with new attributes.
9842 * config/arm/cortex-a5.md: Update with new attributes.
9843 * config/arm/cortex-a53.md: Update with new attributes.
9844 * config/arm/cortex-a7.md: Update with new attributes.
9845 * config/arm/cortex-a8-neon.md: Update with new attributes.
9846 * config/arm/cortex-a9.md: Update with new attributes.
9847 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9848 * config/arm/cortex-r4f.md: Update with new attributes.
9849 * config/arm/marvell-pj4.md: Update with new attributes.
9850 * config/arm/vfp11.md: Update with new attributes.
9852 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9854 * config/aarch64/aarch64.md
9855 (*madd<mode>): Fix type attribute.
9856 (*maddsi_uxtw): Likewise.
9857 (*msub<mode>): Likewise.
9858 (*msubsi_uxtw): Likewise.
9859 (<su_optab>maddsidi4): Likewise.
9860 (<su_optab>msubsidi4): Likewise.
9862 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9864 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9865 * config/arm/arm.md (core_cycles): Remove fdiv.
9866 * config/arm/vfp.md:
9867 (*sqrtsf2_vfp): Update for attribute changes.
9868 (*sqrtdf2_vfp): Likewise.
9869 * config/aarch64/aarch64.md:
9870 (sqrt<mode>2): Update for attribute changes.
9871 * config/arm/arm1020e.md: Update with new attributes.
9872 * config/arm/cortex-a15-neon.md: Update with new attributes.
9873 * config/arm/cortex-a5.md: Update with new attributes.
9874 * config/arm/cortex-a53.md: Update with new attributes.
9875 * config/arm/cortex-a7.md: Update with new attributes.
9876 * config/arm/cortex-a8-neon.md: Update with new attributes.
9877 * config/arm/cortex-a9.md: Update with new attributes.
9878 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9879 * config/arm/cortex-r4f.md: Update with new attributes.
9880 * config/arm/marvell-pj4.md: Update with new attributes.
9881 * config/arm/vfp11.md: Update with new attributes.
9883 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9885 * config/arm/types.md
9886 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9887 * config/aarch64/aarch64.md
9888 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9890 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9891 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9892 (float<GPI:mode><GPF:mode>2): Likewise.
9894 (*truncsisf2_vfp): Update with new attributes.
9895 (*truncsidf2_vfp): Likewise.
9896 (fixuns_truncsfsi2): Likewise.
9897 (fixuns_truncdfsi2): Likewise.
9898 (*floatsisf2_vfp): Likewise.
9899 (*floatsidf2_vfp): Likewise.
9900 (floatunssisf2): Likewise.
9901 (floatunssidf2): Likewise.
9902 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9903 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9904 * config/arm/arm1020e.md: Update with new attributes.
9905 * config/arm/cortex-a15-neon.md: Update with new attributes.
9906 * config/arm/cortex-a5.md: Update with new attributes.
9907 * config/arm/cortex-a53.md: Update with new attributes.
9908 * config/arm/cortex-a7.md: Update with new attributes.
9909 * config/arm/cortex-a8-neon.md: Update with new attributes.
9910 * config/arm/cortex-a9.md: Update with new attributes.
9911 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9912 * config/arm/cortex-r4f.md: Update with new attributes.
9913 * config/arm/marvell-pj4.md: Update with new attributes.
9914 * config/arm/vfp11.md: Update with new attributes.
9916 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9918 * config/aarch64/arm_neon.h
9919 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9920 (vqtbx<1,2,3,4><q>_s8): Likewise.
9922 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9924 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9925 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9926 (add<mode>3): Add type attribute.
9927 (add<mode>3): Likewise.
9928 (usadd<mode>3): Likewise.
9929 (ssadd<mode>3): Likewise.
9930 (sub<mode>3): Likewise.
9931 (sub<mode>3): Likewise.
9932 (ussub<mode>3): Likewise.
9933 (sssub<mode>3): Likewise.
9934 (ssmulsa3): Likewise.
9935 (usmulusa3): Likewise.
9936 (arm_usatsihi): Likewise.
9938 (*movdi_vfp): Add types for all instructions.
9939 (*movdi_vfp_cortexa8): Likewise.
9940 (*movhf_vfp_neon): Likewise.
9941 (*movhf_vfp): Likewise.
9942 (*movdf_vfp): Likewise.
9943 (*thumb2_movdf_vfp): Likewise.
9944 (*thumb2_movdfcc_vfp): Likewise.
9945 * config/arm/arm.md: Add type attribute to all insn patterns.
9946 (*thumb1_adddi3): Add type attribute.
9947 (*arm_adddi3): Likewise.
9948 (*adddi_sesidi_di): Likewise.
9949 (*adddi_zesidi_di): Likewise.
9950 (*thumb1_addsi3): Likewise.
9951 (addsi3_compare0): Likewise.
9952 (*addsi3_compare0_scratch): Likewise.
9953 (*compare_negsi_si): Likewise.
9954 (cmpsi2_addneg): Likewise.
9955 (*addsi3_carryin_<optab>): Likewise.
9956 (*addsi3_carryin_alt2_<optab>): Likewise.
9957 (*addsi3_carryin_clobercc_<optab>): Likewise.
9958 (*subsi3_carryin): Likewise.
9959 (*subsi3_carryin_const): Likewise.
9960 (*subsi3_carryin_compare): Likewise.
9961 (*subsi3_carryin_compare_const): Likewise.
9962 (*arm_subdi3): Likewise.
9963 (*thumb_subdi3): Likewise.
9964 (*subdi_di_zesidi): Likewise.
9965 (*subdi_di_sesidi): Likewise.
9966 (*subdi_zesidi_di): Likewise.
9967 (*subdi_sesidi_di): Likewise.
9968 (*subdi_zesidi_ze): Likewise.
9969 (thumb1_subsi3_insn): Likewise.
9970 (*arm_subsi3_insn): Likewise.
9971 (*anddi3_insn): Likewise.
9972 (*anddi_zesidi_di): Likewise.
9973 (*anddi_sesdi_di): Likewise.
9974 (*ne_zeroextracts): Likewise.
9975 (*ne_zeroextracts): Likewise.
9976 (*ite_ne_zeroextr): Likewise.
9977 (*ite_ne_zeroextr): Likewise.
9978 (*anddi_notdi_di): Likewise.
9979 (*anddi_notzesidi): Likewise.
9980 (*anddi_notsesidi): Likewise.
9981 (andsi_notsi_si): Likewise.
9982 (thumb1_bicsi3): Likewise.
9983 (*iordi3_insn): Likewise.
9984 (*iordi_zesidi_di): Likewise.
9985 (*iordi_sesidi_di): Likewise.
9986 (*thumb1_iorsi3_insn): Likewise.
9987 (*xordi3_insn): Likewise.
9988 (*xordi_zesidi_di): Likewise.
9989 (*xordi_sesidi_di): Likewise.
9990 (*arm_xorsi3): Likewise.
9991 (*andsi_iorsi3_no): Likewise.
9992 (*smax_0): Likewise.
9993 (*smax_m1): Likewise.
9994 (*arm_smax_insn): Likewise.
9995 (*smin_0): Likewise.
9996 (*arm_smin_insn): Likewise.
9997 (*arm_umaxsi3): Likewise.
9998 (*arm_uminsi3): Likewise.
9999 (*minmax_arithsi): Likewise.
10000 (*minmax_arithsi_): Likewise.
10001 (*satsi_<SAT:code>): Likewise.
10002 (arm_ashldi3_1bit): Likewise.
10003 (arm_ashrdi3_1bit): Likewise.
10004 (arm_lshrdi3_1bit): Likewise.
10005 (*arm_negdi2): Likewise.
10006 (*thumb1_negdi2): Likewise.
10007 (*arm_negsi2): Likewise.
10008 (*thumb1_negsi2): Likewise.
10009 (*negdi_extendsid): Likewise.
10010 (*negdi_zero_extend): Likewise.
10011 (*arm_abssi2): Likewise.
10012 (*thumb1_abssi2): Likewise.
10013 (*arm_neg_abssi2): Likewise.
10014 (*thumb1_neg_abss): Likewise.
10015 (one_cmpldi2): Likewise.
10016 (extend<mode>di2): Likewise.
10017 (*compareqi_eq0): Likewise.
10018 (*arm_extendhisi2addsi): Likewise.
10019 (*arm_movdi): Likewise.
10020 (*thumb1_movdi_insn): Likewise.
10021 (*arm_movt): Likewise.
10022 (*thumb1_movsi_insn): Likewise.
10023 (pic_add_dot_plus_four): Likewise.
10024 (pic_add_dot_plus_eight): Likewise.
10025 (tls_load_dot_plus_eight): Likewise.
10026 (*thumb1_movhi_insn): Likewise.
10027 (*thumb1_movsf_insn): Likewise.
10028 (*movdf_soft_insn): Likewise.
10029 (*thumb_movdf_insn): Likewise.
10030 (cbranchsi4_insn): Likewise.
10031 (cbranchsi4_scratch): Likewise.
10032 (*negated_cbranchsi4): Likewise.
10033 (*tbit_cbranch): Likewise.
10034 (*tlobits_cbranch): Likewise.
10035 (*tstsi3_cbranch): Likewise.
10036 (*cbranchne_decr1): Likewise.
10037 (*addsi3_cbranch): Likewise.
10038 (*addsi3_cbranch_scratch): Likewise.
10039 (*arm_cmpdi_insn): Likewise.
10040 (*arm_cmpdi_unsig): Likewise.
10041 (*arm_cmpdi_zero): Likewise.
10042 (*thumb_cmpdi_zero): Likewise.
10043 (*deleted_compare): Likewise.
10044 (*mov_scc): Likewise.
10045 (*mov_negscc): Likewise.
10046 (*mov_notscc): Likewise.
10047 (*cstoresi_eq0_thumb1_insn): Likewise.
10048 (cstoresi_nltu_thumb1): Likewise.
10049 (cstoresi_ltu_thu): Likewise.
10050 (thumb1_addsi3_addgeu): Likewise.
10051 (*arm_jump): Likewise.
10052 (*thumb_jump): Likewise.
10053 (*check_arch2): Likewise.
10054 (arm_casesi_internal): Likewise.
10055 (thumb1_casesi_dispatch): Likewise.
10056 (*arm_indirect_jump): Likewise.
10057 (*thumb1_indirect_jump): Likewise.
10059 (*and_scc): Likewise.
10060 (*ior_scc): Likewise.
10061 (*compare_scc): Likewise.
10062 (*cond_move): Likewise.
10063 (*cond_arith): Likewise.
10064 (*cond_sub): Likewise.
10065 (*cmp_ite0): Likewise.
10066 (*cmp_ite1): Likewise.
10067 (*cmp_and): Likewise.
10068 (*cmp_ior): Likewise.
10069 (*ior_scc_scc): Likewise.
10070 (*ior_scc_scc_cmp): Likewise.
10071 (*and_scc_scc): Likewise.
10072 (*and_scc_scc_cmp): Likewise.
10073 (*and_scc_scc_nod): Likewise.
10074 (*negscc): Likewise.
10075 (movcond_addsi): Likewise.
10076 (movcond): Likewise.
10077 (*ifcompare_plus_move): Likewise.
10078 (*if_plus_move): Likewise.
10079 (*ifcompare_move_plus): Likewise.
10080 (*if_move_plus): Likewise.
10081 (*ifcompare_arith_arith): Likewise.
10082 (*if_arith_arith): Likewise.
10083 (*ifcompare_arith_move): Likewise.
10084 (*if_arith_move): Likewise.
10085 (*ifcompare_move_arith): Likewise.
10086 (*if_move_arith): Likewise.
10087 (*ifcompare_move_not): Likewise.
10088 (*if_move_not): Likewise.
10089 (*ifcompare_not_move): Likewise.
10090 (*if_not_move): Likewise.
10091 (*ifcompare_shift_move): Likewise.
10092 (*if_shift_move): Likewise.
10093 (*ifcompare_move_shift): Likewise.
10094 (*if_move_shift): Likewise.
10095 (*ifcompare_shift_shift): Likewise.
10096 (*ifcompare_not_arith): Likewise.
10097 (*ifcompare_arith_not): Likewise.
10098 (*if_arith_not): Likewise.
10099 (*ifcompare_neg_move): Likewise.
10100 (*if_neg_move): Likewise.
10101 (*ifcompare_move_neg): Likewise.
10102 (*if_move_neg): Likewise.
10103 (prologue_thumb1_interwork): Likewise.
10104 (*cond_move_not): Likewise.
10105 (*sign_extract_onebit): Likewise.
10106 (*not_signextract_onebit): Likewise.
10107 (stack_tie): Likewise.
10108 (align_4): Likewise.
10109 (align_8): Likewise.
10110 (consttable_end): Likewise.
10111 (consttable_1): Likewise.
10112 (consttable_2): Likewise.
10113 (consttable_4): Likewise.
10114 (consttable_8): Likewise.
10115 (consttable_16): Likewise.
10116 (*thumb1_tablejump): Likewise.
10117 (prefetch): Likewise.
10118 (force_register_use): Likewise.
10119 (thumb_eh_return): Likewise.
10120 (load_tp_hard): Likewise.
10121 (load_tp_soft): Likewise.
10122 (tlscall): Likewise.
10123 (*arm_movtas_ze): Likewise.
10124 (*arm_rev): Likewise.
10125 (*arm_revsh): Likewise.
10126 (*arm_rev16): Likewise.
10127 * config/arm/thumb2.md
10128 (*thumb2_smaxsi3): Likewise.
10129 (*thumb2_sminsi3): Likewise.
10130 (*thumb32_umaxsi3): Likewise.
10131 (*thumb2_uminsi3): Likewise.
10132 (*thumb2_negdi2): Likewise.
10133 (*thumb2_abssi2): Likewise.
10134 (*thumb2_neg_abss): Likewise.
10135 (*thumb2_movsi_insn): Likewise.
10136 (tls_load_dot_plus_four): Likewise.
10137 (*thumb2_movhi_insn): Likewise.
10138 (*thumb2_mov_scc): Likewise.
10139 (*thumb2_mov_negs): Likewise.
10140 (*thumb2_mov_negs): Likewise.
10141 (*thumb2_mov_nots): Likewise.
10142 (*thumb2_mov_nots): Likewise.
10143 (*thumb2_movsicc_): Likewise.
10144 (*thumb2_movsfcc_soft_insn): Likewise.
10145 (*thumb2_indirect_jump): Likewise.
10146 (*thumb2_and_scc): Likewise.
10147 (*thumb2_ior_scc): Likewise.
10148 (*thumb2_ior_scc_strict_it): Likewise.
10149 (*thumb2_cond_move): Likewise.
10150 (*thumb2_cond_arith): Likewise.
10151 (*thumb2_cond_ari): Likewise.
10152 (*thumb2_cond_sub): Likewise.
10153 (*thumb2_negscc): Likewise.
10154 (*thumb2_movcond): Likewise.
10155 (thumb2_casesi_internal): Likewise.
10156 (thumb2_casesi_internal_pic): Likewise.
10157 (*thumb2_alusi3_short): Likewise.
10158 (*thumb2_mov<mode>_shortim): Likewise.
10159 (*thumb2_addsi_short): Likewise.
10160 (*thumb2_subsi_short): Likewise.
10161 (thumb2_addsi3_compare0): Likewise.
10162 (*thumb2_cbz): Likewise.
10163 (*thumb2_cbnz): Likewise.
10164 (*thumb2_one_cmplsi2_short): Likewise.
10165 (*thumb2_negsi2_short): Likewise.
10166 (*orsi_notsi_si): Likewise.
10167 * config/arm/arm1020e.md: Update with new attributes.
10168 * config/arm/arm1026ejs.md: Update with new attributes.
10169 * config/arm/arm1136jfs.md: Update with new attributes.
10170 * config/arm/arm926ejs.md: Update with new attributes.
10171 * config/arm/cortex-a15.md: Update with new attributes.
10172 * config/arm/cortex-a5.md: Update with new attributes.
10173 * config/arm/cortex-a53.md: Update with new attributes.
10174 * config/arm/cortex-a7.md: Update with new attributes.
10175 * config/arm/cortex-a8.md: Update with new attributes.
10176 * config/arm/cortex-a9.md: Update with new attributes.
10177 * config/arm/cortex-m4.md: Update with new attributes.
10178 * config/arm/cortex-r4.md: Update with new attributes.
10179 * config/arm/fa526.md: Update with new attributes.
10180 * config/arm/fa606te.md: Update with new attributes.
10181 * config/arm/fa626te.md: Update with new attributes.
10182 * config/arm/fa726te.md: Update with new attributes.
10184 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10186 * config/aarch64/aarch64-simd.md
10187 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
10188 <vwx> iterator to ensure correct register choice.
10189 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
10190 (aarch64_sqdmull_n<mode>): Likewise.
10191 (aarch64_sqdmull2_n<mode>_internal): Likewise.
10192 * config/aarch64/arm_neon.h
10193 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
10194 (vml<as><q>_n_<su>16): Likewise.
10195 (vml<as>l_high_lane<q>_<su>16): Likewise.
10196 (vml<as>l_high_n_<su>16): Likewise.
10197 (vml<as>l_lane<q>_<su>16): Likewise.
10198 (vml<as>l_n_<su>16): Likewise.
10199 (vmul<q>_lane<q>_<su>16): Likewise.
10200 (vmul<q>_n_<su>16): Likewise.
10201 (vmull_lane<q>_<su>16): Likewise.
10202 (vmull_n_<su>16): Likewise.
10203 (vmull_high_lane<q>_<su>16): Likewise.
10204 (vmull_high_n_<su>16): Likewise.
10205 (vqrdmulh<q>_n_s16): Likewise.
10207 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
10209 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
10210 have the correct lane parameter.
10212 2013-09-06 Richard Biener <rguenther@suse.de>
10214 * cfganal.c (control_dependences::~control_dependences):
10215 Properly free all of the vector.
10217 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
10220 * config/i386/i386.c (ix86_conditional_register_usage):
10221 Proper initialize extended SSE registers.
10223 2013-09-06 Jan Hubicka <jh@suse.cz>
10225 PR tree-optimization/58311
10226 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
10228 2013-09-06 Jan Hubicka <jh@suse.cz>
10230 * Makefile.in (tree-sra.o): Update dependencies.
10231 * tree-sra.c: Include ipa-utils.h
10232 (scan_function): Use recursive_call_p.
10233 (has_caller_p): New function.
10234 (cgraph_for_node_and_aliases): Count also callers of aliases.
10236 2013-09-06 Jan Hubicka <jh@suse.cz>
10238 PR middle-end/58094
10239 * cgraph.h (symtab_semantically_equivalent_p): Declare.
10240 * tree-tailcall.c: Include ipa-utils.h.
10241 (find_tail_calls): Use it.
10242 * ipa-pure-const.c (check_call): Likewise.
10243 * ipa-utils.c (recursive_call_p): New function.
10244 * ipa-utils.h (recursive_call_p): Dclare.
10245 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
10246 (symtab_semantically_equivalent_p): New function.
10247 * Makefile.in (tree-tailcall.o): Update dependencies.
10249 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
10251 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
10252 non-inlinable part.
10254 2013-09-06 Richard Biener <rguenther@suse.de>
10256 * lto-streamer.h (lto_global_var_decls): Remove.
10257 * Makefile.in (OBJS): Remove lto-symtab.o.
10258 (lto-symtab.o): Remove.
10259 (GTFILES): Remove lto-symtab.c
10260 * lto-symtab.c: Move to lto/
10262 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10264 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
10265 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
10266 (UNSPEC_FPINT_RINT): New constant definitions.
10267 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
10268 definition with 2 attributes.
10269 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
10270 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
10273 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10275 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
10276 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
10277 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
10278 attribute to "z196_alone".
10279 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
10282 2013-09-06 Richard Biener <rguenther@suse.de>
10284 * basic-block.h (class control_dependences): New.
10285 * tree-ssa-dce.c (control_dependence_map): Remove.
10287 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
10288 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
10289 find_pdom, find_control_dependence, find_all_control_dependences):
10291 (mark_control_dependent_edges_necessary,
10292 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
10293 tree_dce_done, perform_tree_ssa_dce): Adjust.
10294 * cfganal.c (set_control_dependence_map_bit,
10295 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
10296 find_all_control_dependences): Move from tree-ssa-dce.c and
10297 implement as methods of control_dependences class.
10298 (control_dependences::control_dependences): New.
10299 (control_dependences::~control_dependences): Likewise.
10300 (control_dependences::get_edges_dependent_on): Likewise.
10301 (control_dependences::get_edge): Likewise.
10303 2013-09-04 Jan Hubicka <jh@suse.cz>
10305 * tree.c (types_same_for_odr): Drop overactive check.
10306 * ipa-devirt.c (hash_type_name): Likewise.
10308 2013-09-04 Jan Hubicka <jh@suse.cz>
10310 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
10311 (analyze_functions): ... here.
10313 2013-09-04 Jan Hubicka <jh@suse.cz>
10315 PR middle-end/58201
10316 * cgraphunit.c (analyze_functions): Clear AUX fields
10317 after processing; initialize assembler name has.
10319 2013-09-05 Jeff Law <law@redhat.com>
10321 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
10322 from thread_around_empty_block. Record threading path into PATH.
10323 Recurse if threading through the initial block is successful.
10324 (thread_across_edge): Corresponding changes to slightly simplify.
10326 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10328 * config/aarch64/aarch64.md
10329 (type): Remove frecpe, frecps, frecpx.
10330 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
10331 fix to be a TARGET_SIMD instruction.
10332 (aarch64_frecps): Remove.
10333 * config/aarch64/aarch64-simd.md
10334 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
10335 (aarch64_frecps<mode>): Handle all float/vector of float modes.
10337 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10338 Sofiane Naci <sofiane.naci@arm.com>
10340 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
10341 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
10342 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
10343 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
10344 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
10345 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
10346 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
10347 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
10348 into "clz, "rbit". Rename "shift" to "shift_imm".
10349 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
10350 changes. Update for attribute changes all occurrences of arlo_* and
10352 * config/arm/arm-fixed.md: Update for attribute changes
10353 all occurrences of arlo_* types.
10354 * config/arm/thumb2.md: Update for attribute changes all occurrences
10356 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
10357 (cortexa7_older_only): Likewise.
10358 (cortexa7_younger): Likewise.
10359 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
10360 (1020alu_shift_op): Likewise.
10361 (1020alu_shift_reg_op): Likewise.
10362 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
10363 (alu_shift_op): Likewise.
10364 (alu_shift_reg_op): Likewise.
10365 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
10366 (11_alu_shift_op): Likewise.
10367 (11_alu_shift_reg_op): Likewise.
10368 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
10369 (9_alu_shift_reg_op): Likewise.
10370 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
10372 (cortex_a15_alu_shift): Likewise.
10373 (cortex_a15_alu_shift_reg): Likewise.
10374 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
10376 (cortex_a5_alu_shift): Likewise.
10377 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
10379 (cortex_a53_alu_shift): Likewise.
10380 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
10382 (cortex_a7_alu_reg): Likewise.
10383 (cortex_a7_alu_shift): Likewise.
10384 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
10386 (cortex_a8_alu_shift): Likewise.
10387 (cortex_a8_alu_shift_reg): Likewise.
10388 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
10389 (cortex_a9_dp_shift): Likewise.
10390 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
10392 * config/arm/cortex-r4.md
10393 (cortex_r4_alu): Update for attribute changes.
10394 (cortex_r4_mov): Likewise.
10395 (cortex_r4_alu_shift_reg): Likewise.
10396 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
10397 (526_alu_shift_op): Likewise.
10398 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
10399 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
10400 (626te_alu_shift_op): Likewise.
10401 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
10402 (726te_alu_shift_op): Likewise.
10403 (726te_alu_shift_reg_op): Likewise.
10404 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
10405 (mp626_alu_shift_op): Likewise.
10406 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
10407 (pj4_alu_conds): Likewise.
10408 (pj4_shift): Likewise.
10409 (pj4_shift_conds): Likewise.
10410 (pj4_alu_shift): Likewise.
10411 (pj4_alu_shift_conds): Likewise.
10412 * config/aarch64/aarch64.md: Update for attribute change
10413 all occurrences of arlo_* and shift* types.
10415 2013-09-05 Mike Stump <mikestump@comcast.net>
10417 * tree.h: Move documentation for tree_function_decl to tree-core.h
10418 with the declaration.
10420 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
10423 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
10424 looking for widest mode.
10426 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
10428 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
10430 2013-09-05 Richard Biener <rguenther@suse.de>
10432 PR tree-optimization/58137
10433 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
10434 Do not create vectors of pointers.
10435 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
10436 types for the components of the vector initializer.
10437 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
10438 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
10440 2013-09-05 Martin Jambor <mjambor@suse.cz>
10442 * ipa-prop.c (remove_described_reference): Accept missing references,
10443 return false if that hppens, otherwise return true.
10444 (cgraph_node_for_jfunc): New function.
10445 (try_decrement_rdesc_refcount): Likewise.
10446 (try_make_edge_direct_simple_call): Use them.
10447 (ipa_edge_removal_hook): Remove references from rdescs.
10448 (ipa_edge_duplication_hook): Clone rdescs and their references
10449 when the new edge has the same caller as the old one.
10450 * cgraph.c (cgraph_resolve_speculation): Remove speculative
10451 reference before removing any edges.
10453 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
10455 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
10457 * thumb2.md (thumb2_storewb_parisi): New pattern.
10459 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
10461 * config/aarch64/aarch64-option-extensions.def: Add
10462 AARCH64_OPT_EXTENSION of 'crc'.
10463 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
10464 (AARCH64_ISA_CRC): Ditto.
10465 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
10466 description of the CRC extension.
10468 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
10470 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
10471 * config/rs6000/linux.h: Ditto.
10472 * alpha/linux.h: Ditto.
10473 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
10474 no_c99_libc_has_function.
10475 * config/c6x/uclinux-elf.h: Ditto.
10476 * config/lm32/uclinux-elf.h: Ditto.
10477 * config/m68k/uclinux.h: Ditto.
10478 * config/moxie/uclinux.h: Ditto.
10479 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
10480 (crisv32-*-linux*, cris-*-linux*): Ditto.
10481 * config/bfin/bfin.c: Include "tm_p.h".
10483 2013-09-05 Richard Biener <rguenther@suse.de>
10485 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
10486 check for a definition without a basic-block.
10488 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10489 Sofiane Naci <sofiane.naci@arm.com>
10491 * config/aarch64/aarch64.md
10492 (*movti_aarch64): Rename r_2_f and f_2_r.
10493 (*movsf_aarch64): Likewise.
10494 (*movdf_aarch64): Likewise.
10495 (*movtf_aarch64): Likewise.
10496 (aarch64_movdi_<mode>low): Likewise.
10497 (aarch64_movdi_<mode>high): Likewise.
10498 (aarch64_mov<mode>high_di): Likewise.
10499 (aarch64_mov<mode>low_di): Likewise.
10500 (aarch64_movtilow_tilow): Likewise.
10501 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
10502 values to config/arm/types.md
10503 (attribute "conds"): Update for attribute change.
10504 (anddi3_insn): Likewise.
10505 (iordi3_insn): Likewise.
10506 (xordi3_insn): Likewise.
10507 (one_cmpldi2): Likewise.
10508 * config/arm/types.md (type): Add Neon types.
10509 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
10510 use "type" attribute.
10511 (movmisalign<mode>_neon_store): Likewise.
10512 (movmisalign<mode>_neon_load): Likewise.
10513 (vec_set<mode>_internal): Likewise.
10514 (vec_setv2di_internal): Likewise.
10515 (vec_extract<mode>): Likewise.
10516 (vec_extractv2di): Likewise.
10517 (add<mode>3_neon): Likewise.
10518 (adddi3_neon): Likewise.
10519 (sub<mode>3_neon): Likewise.
10520 (subdi3_neon): Likewise.
10521 (mul<mode>3_neon): Likewise.
10522 (mul<mode>3add<mode>_neon): Likewise.
10523 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
10524 (fma<VCVTF:mode>4)): Likewise.
10525 (fma<VCVTF:mode>4_intrinsic): Likewise.
10526 (fmsub<VCVTF:mode>4)): Likewise.
10527 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
10528 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
10529 (ior<mode>3): Likewise.
10530 (and<mode>3): Likewise.
10531 (anddi3_neon): Likewise.
10532 (orn<mode>3_neon): Likewise.
10533 (orndi3_neon): Likewise.
10534 (bic<mode>3_neon): Likewise.
10535 (bicdi3_neon): Likewise.
10536 (xor<mode>3): Likewise.
10537 (one_cmpl<mode>2): Likewise.
10538 (abs<mode>2): Likewise.
10539 (neg<mode>2): Likewise.
10540 (umin<mode>3_neon): Likewise.
10541 (umax<mode>3_neon): Likewise.
10542 (smin<mode>3_neon): Likewise.
10543 (smax<mode>3_neon): Likewise.
10544 (vashl<mode>3): Likewise.
10545 (vashr<mode>3_imm): Likewise.
10546 (vlshr<mode>3_imm): Likewise.
10547 (ashl<mode>3_signed): Likewise.
10548 (ashl<mode>3_unsigned): Likewise.
10549 (neon_load_count): Likewise.
10550 (ashldi3_neon_noclobber): Likewise.
10551 (signed_shift_di3_neon): Likewise.
10552 (unsigned_shift_di3_neon): Likewise.
10553 (ashrdi3_neon_imm_noclobber): Likewise.
10554 (lshrdi3_neon_imm_noclobber): Likewise.
10555 (widen_ssum<mode>3): Likewise.
10556 (widen_usum<mode>3): Likewise.
10557 (quad_halves_<code>v4si): Likewise.
10558 (quad_halves_<code>v4sf): Likewise.
10559 (quad_halves_<code>v8hi): Likewise.
10560 (quad_halves_<code>v16qi): Likewise.
10561 (reduc_splus_v2di): Likewise.
10562 (neon_vpadd_internal<mode>): Likewise.
10563 (neon_vpsmin<mode>): Likewise.
10564 (neon_vpsmax<mode>): Likewise.
10565 (neon_vpumin<mode>): Likewise.
10566 (neon_vpumax<mode>): Likewise.
10567 (ss_add<mode>_neon): Likewise.
10568 (us_add<mode>_neon): Likewise.
10569 (ss_sub<mode>_neon): Likewise.
10570 (us_sub<mode>_neon): Likewise.
10571 (neon_vadd<mode>_unspec): Likewise.
10572 (neon_vaddl<mode>): Likewise.
10573 (neon_vaddw<mode>): Likewise.
10574 (neon_vhadd<mode>): Likewise.
10575 (neon_vqadd<mode>): Likewise.
10576 (neon_vaddhn<mode>): Likewise.
10577 (neon_vmul<mode>): Likewise.
10578 (neon_vmla<mode>): Likewise.
10579 (neon_vmlal<mode>): Likewise.
10580 (neon_vmls<mode>): Likewise.
10581 (neon_vmlsl<mode>): Likewise.
10582 (neon_vqdmulh<mode>): Likewise.
10583 (neon_vqdmlal<mode>): Likewise.
10584 (neon_vqdmlsl<mode>): Likewise.
10585 (neon_vmull<mode>): Likewise.
10586 (neon_vqdmull<mode>): Likewise.
10587 (neon_vsub<mode>_unspec): Likewise.
10588 (neon_vsubl<mode>): Likewise.
10589 (neon_vsubw<mode>): Likewise.
10590 (neon_vqsub<mode>): Likewise.
10591 (neon_vhsub<mode>): Likewise.
10592 (neon_vsubhn<mode>): Likewise.
10593 (neon_vceq<mode>): Likewise.
10594 (neon_vcge<mode>): Likewise.
10595 (neon_vcgeu<mode>): Likewise.
10596 (neon_vcgt<mode>): Likewise.
10597 (neon_vcgtu<mode>): Likewise.
10598 (neon_vcle<mode>): Likewise.
10599 (neon_vclt<mode>): Likewise.
10600 (neon_vcage<mode>): Likewise.
10601 (neon_vcagt<mode>): Likewise.
10602 (neon_vtst<mode>): Likewise.
10603 (neon_vabd<mode>): Likewise.
10604 (neon_vabdl<mode>): Likewise.
10605 (neon_vaba<mode>): Likewise.
10606 (neon_vabal<mode>): Likewise.
10607 (neon_vmax<mode>): Likewise.
10608 (neon_vmin<mode>): Likewise.
10609 (neon_vpaddl<mode>): Likewise.
10610 (neon_vpadal<mode>): Likewise.
10611 (neon_vpmax<mode>): Likewise.
10612 (neon_vpmin<mode>): Likewise.
10613 (neon_vrecps<mode>): Likewise.
10614 (neon_vrsqrts<mode>): Likewise.
10615 (neon_vqabs<mode>): Likewise.
10616 (neon_vqneg<mode>): Likewise.
10617 (neon_vcls<mode>): Likewise.
10618 (clz<mode>2): Likewise.
10619 (popcount<mode>2): Likewise.
10620 (neon_vrecpe): Likewise.
10621 (neon_vrsqrte): Likewise.
10622 (neon_vget_lane<mode>_sext_internal): Likewise.
10623 (neon_vget_lane<mode>_zext_internal): Likewise.
10624 (neon_vdup_n<mode>): Likewise.
10625 (neon_vdup_nv2di): Likewise.
10626 (neon_vdpu_lane<mode>_internal): Likewise.
10627 (neon_vswp<mode>): Likewise.
10628 (float<mode><V_cvtto>2): Likewise.
10629 (floatuns<mode><V_cvtto>2): Likewise.
10630 (fix_trunc<mode><V_cvtto>)2): Likewise
10631 (fixuns_trunc<mode><V_cvtto)2): Likewise.
10632 (neon_vcvt<mode>): Likewise.
10633 (neon_vcvtv4sfv4hf): Likewise.
10634 (neon_vcvtv4hfv4sf): Likewise.
10635 (neon_vcvt_n<mode>): Likewise.
10636 (neon_vmovn<mode>): Likewise.
10637 (neon_vqmovn<mode>): Likewise.
10638 (neon_vqmovun<mode>): Likewise.
10639 (neon_vmovl<mode>): Likewise.
10640 (neon_vmul_lane<mode>): Likewise.
10641 (neon_vmull_lane<mode>): Likewise.
10642 (neon_vqdmull_lane<mode>): Likewise.
10643 (neon_vqdmulh_lane<mode>): Likewise.
10644 (neon_vmla_lane<mode>): Likewise.
10645 (neon_vmlal_lane<mode>): Likewise.
10646 (neon_vqdmlal_lane<mode>): Likewise.
10647 (neon_vmls_lane<mode>): Likewise.
10648 (neon_vmlsl_lane<mode>): Likewise.
10649 (neon_vqdmlsl_lane<mode>): Likewise.
10650 (neon_vext<mode>): Likewise.
10651 (neon_vrev64<mode>): Likewise.
10652 (neon_vrev32<mode>): Likewise.
10653 (neon_vrev16<mode>): Likewise.
10654 (neon_vbsl<mode>_internal): Likewise.
10655 (neon_vshl<mode>): Likewise.
10656 (neon_vqshl<mode>): Likewise.
10657 (neon_vshr_n<mode>): Likewise.
10658 (neon_vshrn_n<mode>): Likewise.
10659 (neon_vqshrn_n<mode>): Likewise.
10660 (neon_vqshrun_n<mode>): Likewise.
10661 (neon_vshl_n<mode>): Likewise.
10662 (neon_vqshl_n<mode>): Likewise.
10663 (neon_vqshlu_n<mode>): Likewise.
10664 (neon_vshll_n<mode>): Likewise.
10665 (neon_vsra_n<mode>): Likewise.
10666 (neon_vsri_n<mode>): Likewise.
10667 (neon_vsli_n<mode>): Likewise.
10668 (neon_vtbl1v8qi): Likewise.
10669 (neon_vtbl2v8qi): Likewise.
10670 (neon_vtbl3v8qi): Likewise.
10671 (neon_vtbl4v8qi): Likewise.
10672 (neon_vtbx1v8qi): Likewise.
10673 (neon_vtbx2v8qi): Likewise.
10674 (neon_vtbx3v8qi): Likewise.
10675 (neon_vtbx4v8qi): Likewise.
10676 (neon_vtrn<mode>_internal): Likewise.
10677 (neon_vzip<mode>_internal): Likewise.
10678 (neon_vuzp<mode>_internal): Likewise.
10679 (neon_vld1<mode>): Likewise.
10680 (neon_vld1_lane<mode>): Likewise.
10681 (neon_vld1_dup<mode>): Likewise.
10682 (neon_vld1_dupv2di): Likewise.
10683 (neon_vst1<mode>): Likewise.
10684 (neon_vst1_lane<mode>): Likewise.
10685 (neon_vld2<mode>): Likewise.
10686 (neon_vld2_lane<mode>): Likewise.
10687 (neon_vld2_dup<mode>): Likewise.
10688 (neon_vst2<mode>): Likewise.
10689 (neon_vst2_lane<mode>): Likewise.
10690 (neon_vld3<mode>): Likewise.
10691 (neon_vld3qa<mode>): Likewise.
10692 (neon_vld3qb<mode>): Likewise.
10693 (neon_vld3_lane<mode>): Likewise.
10694 (neon_vld3_dup<mode>): Likewise.
10695 (neon_vst3<mode>): Likewise.
10696 (neon_vst3qa<mode>): Likewise.
10697 (neon_vst3qb<mode>): Likewise.
10698 (neon_vst3_lane<mode>): Likewise.
10699 (neon_vld4<mode>): Likewise.
10700 (neon_vld4qa<mode>): Likewise.
10701 (neon_vld4qb<mode>): Likewise.
10702 (neon_vld4_lane<mode>): Likewise.
10703 (neon_vld4_dup<mode>): Likewise.
10704 (neon_vst4<mode>): Likewise.
10705 (neon_vst4qa<mode>): Likewise.
10706 (neon_vst4qb<mode>): Likewise.
10707 (neon_vst4_lane<mode>): Likewise.
10708 (neon_vec_unpack<US>_lo_<mode>): Likewise.
10709 (neon_vec_unpack<US>_hi_<mode>): Likewise.
10710 (neon_vec_<US>mult_lo_<mode>): Likewise.
10711 (neon_vec_<US>mult_hi_<mode>): Likewise.
10712 (neon_vec_<US>shiftl_<mode>): Likewise.
10713 (neon_unpack<US>_<mode>): Likewise.
10714 (neon_vec_<US>mult_<mode>): Likewise.
10715 (vec_pack_trunc_<mode>): Likewise.
10716 (neon_vec_pack_trunk_<mode>): Likewise.
10717 (neon_vabd<mode>_2): Likewise.
10718 (neon_vabd<mode>_3): Likewise.
10719 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
10720 (thumb2_movsi_vfp): Likewise.
10721 (movdi_vfp): Likewise.
10722 (movdi_vfp_cortexa8): Likewise.
10723 (movhf_vfp_neon): Likewise.
10724 (movhf_vfp): Likewiwse.
10725 (movsf_vfp): Likewiwse.
10726 (thumb2_movsf_vfp): Likewiwse.
10727 (movdf_vfp): Likewise.
10728 (thumb2_movdf_vfp): Likewise.
10729 (movsfcc_vfp): Likewise.
10730 (thumb2_movsfcc_vfp): Likewise.
10731 (movdfcc_vfp): Likewise.
10732 (thumb2_movdfcc_vfp): Likewise.
10733 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
10734 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
10735 (v10_v2c): Likewise.
10736 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
10738 (cortex_a15_neon_int_2): Likewise.
10739 (cortex_a15_neon_int_3): Likewise.
10740 (cortex_a15_neon_int_4): Likewise.
10741 (cortex_a15_neon_int_5): Likewise.
10742 (cortex_a15_neon_vqneg_vqabs): Likewise.
10743 (cortex_a15_neon_vmov): Likewise.
10744 (cortex_a15_neon_vaba): Likewise.
10745 (cortex_a15_neon_vaba_qqq): Likewise.
10746 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10747 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10748 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
10749 scalar_64_32_long_scalar): Likewise.
10750 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10751 (cortex_a15_neon_mla_qqq_8_16): Likewise.
10752 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10753 lotype_qdd_64_32_long): Likewise.
10754 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10755 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10756 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
10757 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10758 (cortex_a15_neon_shift_1): Likewise.
10759 (cortex_a15_neon_shift_2): Likewise.
10760 (cortex_a15_neon_shift_3): Likewise.
10761 (cortex_a15_neon_vshl_ddd): Likewise.
10762 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10763 (cortex_a15_neon_vsra_vrsra): Likewise.
10764 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
10765 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
10766 (cortex_a15_neon_fp_vmul_ddd): Likewise.
10767 (cortex_a15_neon_fp_vmul_qqd): Likewise.
10768 (cortex_a15_neon_fp_vmla_ddd): Likewise.
10769 (cortex_a15_neon_fp_vmla_qqq): Likewise.
10770 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
10771 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
10772 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10773 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10774 (cortex_a15_neon_bp_simple): Likewise.
10775 (cortex_a15_neon_bp_2cycle): Likewise.
10776 (cortex_a15_neon_bp_3cycle): Likewise.
10777 (cortex_a15_neon_vld1_1_2_regs): Likewise.
10778 (cortex_a15_neon_vld1_3_4_regs): Likewise.
10779 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10780 (cortex_a15_neon_vld2_4_regs): Likewise.
10781 (cortex_a15_neon_vld3_vld4): Likewise.
10782 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10783 (cortex_a15_neon_vst1_3_4_regs): Likewise.
10784 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
10785 (cortex_a15_neon_vst3_vst4): Likewise.
10786 (cortex_a15_neon_vld1_vld2_lane): Likewise.
10787 (cortex_a15_neon_vld3_vld4_lane" 10
10788 (cortex_a15_neon_vst1_vst2_lane): Likewise.
10789 (cortex_a15_neon_vst3_vst4_lane): Likewise.
10790 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
10791 (cortex_a15_neon_ldm_2): Likewise.0
10792 (cortex_a15_neon_stm_2): Likewise.
10793 (cortex_a15_neon_mcr): Likewise.
10794 (cortex_a15_neon_mcr_2_mcrr): Likewise.
10795 (cortex_a15_neon_mrc): Likewise.
10796 (cortex_a15_neon_mrrc): Likewise.
10797 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10799 (cortex_a15_alu_shift): Likewise.
10800 (cortex_a15_alu_shift_reg): Likewise.
10801 (cortex_a15_mult32): Likewise.
10802 (cortex_a15_mult64): Likewise.
10803 (cortex_a15_block): Likewise.
10804 (cortex_a15_branch): Likewise.
10805 (cortex_a15_load1): Likewise.
10806 (cortex_a15_load3): Likewise.
10807 (cortex_a15_store1): Likewise.
10808 (cortex_a15_store3): Likewise.
10809 (cortex_a15_call): Likewise.
10810 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
10811 (cortex_a5_f2r): Likewise.
10812 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
10814 (cortex_a53_f2r): Likewise.
10815 * config/arm/cortex-a7.md
10816 (cortex_a7_branch): Update for attribute change.
10817 (cortex_a7_call): Likewise.
10818 (cortex_a7_alu_imm): Likewise.
10819 (cortex_a7_alu_reg): Likewise.
10820 (cortex_a7_alu_shift): Likewise.
10821 (cortex_a7_mul): Likewise.
10822 (cortex_a7_load1): Likewise.
10823 (cortex_a7_store1): Likewise.
10824 (cortex_a7_load2): Likewise.
10825 (cortex_a7_store2): Likewise.
10826 (cortex_a7_load3): Likewise.
10827 (cortex_a7_store3): Likewise.
10828 (cortex_a7_load4): Likewise.
10829 (cortex_a7_store4): Likewise.
10830 (cortex_a7_fpalu): Likewise.
10831 (cortex_a7_fconst): Likewise.
10832 (cortex_a7_fpmuls): Likewise.
10833 (cortex_a7_neon_mul): Likewise.
10834 (cortex_a7_fpmacs): Likewise.
10835 (cortex_a7_neon_mla: Likewise.
10836 (cortex_a7_fpmuld: Likewise.
10837 (cortex_a7_fpmacd: Likewise.
10838 (cortex_a7_fpfmad: Likewise.
10839 (cortex_a7_fdivs: Likewise.
10840 (cortex_a7_fdivd: Likewise.
10841 (cortex_a7_r2f: Likewise.
10842 (cortex_a7_f2r: Likewise.
10843 (cortex_a7_f_flags: Likewise.
10844 (cortex_a7_f_loads: Likewise.
10845 (cortex_a7_f_loadd: Likewise.
10846 (cortex_a7_f_stores: Likewise.
10847 (cortex_a7_f_stored: Likewise.
10848 (cortex_a7_neon): Likewise.
10849 * config/arm/cortex-a8-neon.md
10850 (cortex_a8_neon_mrc): Update for attribute change.
10851 (cortex_a8_neon_mrrc): Likewise.
10852 (cortex_a8_neon_int_1): Likewise.
10853 (cortex_a8_neon_int_2): Likewise.
10854 (cortex_a8_neon_int_3): Likewise.
10855 (cortex_a8_neon_int_4): Likewise.
10856 (cortex_a8_neon_int_5): Likewise.
10857 (cortex_a8_neon_vqneg_vqabs): Likewise.
10858 (cortex_a8_neon_vmov): Likewise.
10859 (cortex_a8_neon_vaba): Likewise.
10860 (cortex_a8_neon_vaba_qqq): Likewise.
10861 (cortex_a8_neon_vsma): Likewise.
10862 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10863 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10864 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10865 long_scalar): Likewise.
10866 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10867 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10868 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10869 long_scalar_qdd_64_32_long): Likewise.
10870 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10871 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10872 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10873 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10874 (cortex_a8_neon_shift_1): Likewise.
10875 (cortex_a8_neon_shift_2): Likewise.
10876 (cortex_a8_neon_shift_3): Likewise.
10877 (cortex_a8_neon_vshl_ddd): Likewise.
10878 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10879 (cortex_a8_neon_vsra_vrsra): Likewise.
10880 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10881 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10882 (cortex_a8_neon_fp_vsum): Likewise.
10883 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10884 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10885 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10886 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10887 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10888 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10889 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10890 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10891 (cortex_a8_neon_bp_simple): Likewise.
10892 (cortex_a8_neon_bp_2cycle): Likewise.
10893 (cortex_a8_neon_bp_3cycle): Likewise.
10894 (cortex_a8_neon_ldr): Likewise.
10895 (cortex_a8_neon_str): Likewise.
10896 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10897 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10898 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10899 (cortex_a8_neon_vld2_4_regs): Likewise.
10900 (cortex_a8_neon_vld3_vld4): Likewise.
10901 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10902 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10903 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10904 (cortex_a8_neon_vst3_vst4): Likewise.
10905 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10906 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10907 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10908 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10909 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10910 (cortex_a8_neon_mcr): Likewise.
10911 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10912 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10913 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10915 (ca9_neon_mrrc): Likewise.
10916 (cortex_a9_neon_int_1): Likewise.
10917 (cortex_a9_neon_int_2): Likewise.
10918 (cortex_a9_neon_int_3): Likewise.
10919 (cortex_a9_neon_int_4): Likewise.
10920 (cortex_a9_neon_int_5): Likewise.
10921 (cortex_a9_neon_vqneg_vqabs): Likewise.
10922 (cortex_a9_neon_vmov): Likewise.
10923 (cortex_a9_neon_vaba): Likewise.
10924 (cortex_a9_neon_vaba_qqq): Likewise.
10925 (cortex_a9_neon_vsma): Likewise.
10926 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10927 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10928 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10929 long_scalar): Likewise.
10930 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10931 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10932 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10933 long_scalar_qdd_64_32_long): Likewise.
10934 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10935 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10936 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10937 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10938 (cortex_a9_neon_shift_1): Likewise.
10939 (cortex_a9_neon_shift_2): Likewise.
10940 (cortex_a9_neon_shift_3): Likewise.
10941 (cortex_a9_neon_vshl_ddd): Likewise.
10942 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10943 (cortex_a9_neon_vsra_vrsra): Likewise.
10944 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10945 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10946 (cortex_a9_neon_fp_vsum): Likewise.
10947 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10948 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10949 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10950 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10951 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10952 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10953 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10954 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10955 (cortex_a9_neon_bp_simple): Likewise.
10956 (cortex_a9_neon_bp_2cycle): Likewise.
10957 (cortex_a9_neon_bp_3cycle): Likewise.
10958 (cortex_a9_neon_ldr): Likewise.
10959 (cortex_a9_neon_str): Likewise.
10960 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10961 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10962 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10963 (cortex_a9_neon_vld2_4_regs): Likewise.
10964 (cortex_a9_neon_vld3_vld4): Likewise.
10965 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10966 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10967 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10968 (cortex_a9_neon_vst3_vst4): Likewise.
10969 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10970 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10971 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10972 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10973 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10974 (cortex_a9_neon_mcr): Likewise.
10975 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10976 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10977 (cortex_a9_fps): Likewise.
10978 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10980 (cortex_m4_fmuls): Likewise.
10981 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10983 (cortex_r4_mrc): Likewise.
10984 * config/arm/iterators.md: Update comment referring to neon_type.
10985 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10986 (iwmmxt_movsi_insn): Likewise.
10987 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10989 (pj4_core_to_vfp): Likewise.
10990 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10992 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10993 (vfp_fstore): Likewise.
10994 * doc/md.texi: Change references to neon_type to refer to type.
10996 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10998 * tree.h (DECL_BUILT_IN): Fix typo in comment.
11000 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
11002 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
11003 lglobl if not weak.
11005 2013-09-04 Easwaran Raman <eraman@google.com>
11007 PR middle-end/57370
11008 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
11009 (build_and_add_sum): Use it.
11010 (appears_later_in_bb): Simplify code.
11012 2013-09-04 Teresa Johnson <tejohnson@google.com>
11014 * dumpfile.c (dump_finish): Don't close stderr/stdout.
11016 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
11018 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
11020 2013-09-04 Jan Hubicka <jh@suse.cz>
11022 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
11023 * ipa-devirt.c: Include diganostic.h
11024 (odr_type_d): Add types and types_set.
11025 (hash_type_name): Work for types with vtables during LTO.
11026 (odr_hasher::remove): Fix comment; destroy types_set.
11027 (add_type_duplicate): New function,
11028 (get_odr_type): Use it.
11029 (dump_type_inheritance_graph): Dump type duplicates.
11030 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
11032 * tree.c (types_same_for_odr): Give exact answers on types with
11035 2013-09-04 Dodji Seketeli <dodji@redhat.com>
11037 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
11038 explaining their differences.
11040 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
11042 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
11044 2013-09-03 Jeff Law <law@redhat.com>
11046 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
11047 when not threading through a joiner block. Pass joiner/no joiner
11048 state to register_jump_thread.
11049 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
11050 state from argument rather than implying on path length.
11051 Dump the entire jump thread path into debugging dump.
11052 * tree-flow.h (register_jump_thread): Update prototype.
11054 2013-08-29 Xinliang David Li <davidxl@google.com>
11056 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
11057 Remove a trivial gcc_assert.
11059 2013-08-29 Xinliang David Li <davidxl@google.com>
11061 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
11062 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
11063 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
11064 Delay base decl alignment adjustment.
11065 * tree-vectorizer.c (vect_destroy_datarefs): New function.
11066 * tree-vectorizer.h: New data structure.
11067 (set_dr_misalignment): New function.
11068 (dr_misalignment): Ditto.
11069 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
11070 (vectorizable_load): Ditto.
11071 (ensure_base_align): New function.
11072 (vectorize_loops): Add dbg_cnt support.
11073 (execute_vect_slp): Ditto.
11074 * dbgcnt.def: New debug counter.
11075 * Makefile: New dependency.
11077 2013-09-03 Meador Inge <meadori@codesourcery.com>
11081 2013-08-30 Meador Inge <meadori@codesourcery.com>
11083 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
11085 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
11087 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
11088 function descriptor.
11090 2013-09-03 Richard Biener <rguenther@suse.de>
11092 * tree-affine.c (add_elt_to_tree): Fix association issue,
11093 avoid useless converts and make sure to always return a
11094 properly typed result.
11096 2013-09-03 Richard Biener <rguenther@suse.de>
11098 PR middle-end/57656
11099 * fold-const.c (negate_expr_p): Fix division case.
11100 (negate_expr): Likewise.
11102 2013-09-03 Richard Biener <rguenther@suse.de>
11105 * tree-streamer-out.c: Include tm.h.
11106 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
11108 2013-09-03 Jan Hubicka <jh@suse.cz>
11110 * tree-profile.c (tree_profiling): Cleanup CFG when done.
11112 2013-09-03 Alan Modra <amodra@gmail.com>
11114 * config.gcc (powerpc*-*-linux*): Add support for little-endian
11115 multilibs to big-endian target and vice versa.
11116 * config/rs6000/t-linux64: Use := assignment on all vars.
11117 (MULTILIB_EXTRA_OPTS): Remove fPIC.
11118 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
11119 * config/rs6000/t-linux64le: New file.
11120 * config/rs6000/t-linux64bele: New file.
11121 * config/rs6000/t-linux64lebe: New file.
11123 2013-09-02 Jan Hubicka <jh@suse.cz>
11125 * ipa-inline-transform.c (inline_transform): Do not
11126 optimize_inline_calls when not optimizing.
11128 2013-09-02 Jan Hubicka <jh@suse.cz>
11130 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
11131 duplicated nodes for assembler names.
11132 * symtab.c (symtab_unregister_node): Do not attempt to unlink
11133 hard registers from assembler name hash.
11135 2013-09-02 Jan Hubicka <jh@suse.cz>
11137 * ipa-split.c (execute_split_functions): Split externally visible
11138 functions called once.
11140 2013-09-02 Martin Jambor <mjambor@suse.cz>
11143 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
11144 linked list. When finding the correct duplicate, also consider also
11145 the caller in additon to its inlined_to node.
11147 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
11149 * config/aarch64/aarch64-simd-builtins.def
11150 (dup_lane_scalar): Remove.
11151 * config/aarch64/aarch64-simd.md
11152 (aarch64_simd_dup): Add 'w->w' alternative.
11153 (aarch64_dup_lane<mode>): Allow for VALL.
11154 (aarch64_dup_lane_scalar<mode>): Remove.
11155 (aarch64_dup_lane_<vswap_width_name><mode>): New.
11156 (aarch64_get_lane_signed<mode>): Add w->w altenative.
11157 (aarch64_get_lane_unsigned<mode>): Likewise.
11158 (aarch64_get_lane<mode>): Likewise.
11159 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
11160 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
11161 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
11162 (VCON): Change container of V2SF.
11163 (vswap_width_name): Likewise.
11164 * config/aarch64/arm_neon.h
11165 (__aarch64_vdup_lane_any): New.
11166 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
11167 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
11168 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
11170 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
11172 PR middle-end/56382
11173 * expr.c (emit_move_complex): Do not move complex FP values as parts if
11174 the source or the destination is a single hard register.
11176 2013-09-02 Richard Biener <rguenther@suse.de>
11178 PR middle-end/57511
11179 * tree-scalar-evolution.c (instantiate_scev_name): Allow
11182 2013-09-02 Richard Biener <rguenther@suse.de>
11184 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
11185 arithmetic to sizetype.
11187 2013-09-02 Bin Cheng <bin.cheng@arm.com>
11189 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
11190 Find auto-increment use both before and after candidate.
11192 2013-09-02 Marek Polacek <polacek@redhat.com>
11194 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
11196 2013-09-01 Jan Hubicka <jh@suse.cz>
11198 * Makefile.in: Add ipa-profile.o
11199 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
11200 * cgraph.c (struct cgraph_propagate_frequency_data,
11201 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
11202 ipa-profile.c; replace cgraph_ by ipa_ prefix.
11203 * cgraph.h (cgraph_propagate_frequency): Remove.
11204 * ipa-inline-analysis.c: Include ipa-utils.h;
11205 drop duplicated cfgloop.h.
11206 (inline_update_callee_summaries): Update.
11207 * ipa-profile.c: New file.
11208 * ipa-utils.h (ipa_propagate_frequency): Declare.
11209 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
11210 data-streamer.h, value-prof.h.
11211 (symtab_remove_unreachable_nodes): Update profile.
11212 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
11213 account_time_size, cmp_counts, dump_histogram,
11214 ipa_profile_generate_summary, ipa_profile_write_summary,
11215 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
11216 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
11217 Move to ipa-profile.c.
11219 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
11221 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
11223 2013-09-01 Jan Hubicka <jh@suse.cz>
11225 * common.opt (fdevirtualize-speculatively): New function.
11226 * invoke.texi (fdevirtualize-speculatively): Document.
11227 * ipa-devirt.c: Include ipa-inline.h
11228 (likely_target_p): New function.
11229 (ipa_devirt): New function.
11230 (gate_ipa_devirt): New function.
11231 (pass_data_ipa_devirt): New static var.
11232 (pass_ipa_devirt): Likewise.
11233 (make_pass_ipa_devirt): New function.
11234 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
11235 (common_handle_option): Disable devirtualization when
11236 value range profiling is available.
11237 * passes.def (pass_ipa_devirt): Add.
11238 * timever.def (TV_IPA_DEVIRT): New timevar.
11239 * tree-pass.h (make_pass_ipa_devirt):
11241 2013-09-01 Iain Sandoe <iain@codesourcery.com>
11243 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
11244 include sanitize(undefined).
11246 2013-08-31 Diego Novillo <dnovillo@google.com>
11248 * Makefile.in (TREE_CORE_H): Define.
11250 (GTFILES): Add tree-core.h.
11251 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
11253 * tree-core.h: New file.
11254 Move all data structures, enum, typedefs, global
11255 declarations and constants from ...
11256 * tree.h: ... here.
11258 2013-08-31 Jan Hubicka <jh@suse.cz>
11260 * bulitins.c (expand_builtin): Do not early exit for gcov
11261 instrumented functions.
11263 2013-08-31 Marek Polacek <polacek@redhat.com>
11265 * ubsan.c: Include tm_p.h.
11267 2013-08-31 Jan Hubicka <jh@suse.cz>
11269 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
11272 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
11273 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
11274 * tree-cfg.c (verify_gimple_label): ... here.
11275 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
11276 (ipa_merge_profiles): New function.
11277 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
11278 (lto_input_function_body): Likewise.
11279 * ipa-utils.h (ipa_merge_profiles): Declare.
11280 * lto-streamer.h (lto_input_function_body): Update prototype.
11281 (emit_label_in_global_context_p): Remove.
11282 * lto-symtab.c: Include ipa-utils.h
11283 (lto_cgraph_replace_node): Use ipa_merge_profiles.
11285 2013-08-31 Jan Hubicka <jh@suse.cz>
11287 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
11289 2013-08-31 Jan Hubicka <jh@suse.cz>
11291 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
11293 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
11295 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
11296 "cmp" RTX before signed_comparison_operator check to account
11297 for "code" changes.
11299 2013-08-30 Jan Hubicka <jh@suse.cz>
11301 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
11302 (detect_type_change_1): Rename to ...
11303 (detect_type_change): ... this one; early return on non-polymorphic
11305 (detect_type_change_ssa): Add comp_type parameter; update
11306 use of detect_type_change.
11307 (compute_complex_assign_jump_func): Add param_type parameter;
11308 update use of detect_type_change_ssa.
11309 (compute_complex_ancestor_jump_func): Likewise.
11310 (ipa_get_callee_param_type): New function.
11311 (ipa_compute_jump_functions_for_edge): Compute parameter type;
11312 update calls to the jump function computation functions.
11314 2013-08-30 Teresa Johnson <tejohnson@google.com>
11315 Steven Bosscher <steven@gcc.gnu.org>
11317 * cfgrtl.c (fixup_new_cold_bb): New routine.
11318 (commit_edge_insertions): Invoke fixup_partitions.
11319 (find_partition_fixes): New routine.
11320 (fixup_partitions): Ditto.
11321 (verify_hot_cold_block_grouping): Update comments.
11322 (rtl_verify_edges): Invoke find_partition_fixes.
11323 (rtl_verify_bb_pointers): Update comments.
11324 (rtl_verify_bb_layout): Ditto.
11325 * basic-block.h (probably_never_executed_edge_p): Declare.
11326 (fixup_partitions): Ditto.
11327 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
11328 * bb-reorder.c (sanitize_hot_paths): New function.
11329 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
11330 sanitize_hot_paths.
11331 * predict.c (probably_never_executed_edge_p): New routine.
11332 * cfg.c (check_bb_profile): Add partition insanity warnings.
11334 2013-08-30 Meador Inge <meadori@codesourcery.com>
11336 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
11338 2013-08-30 Marek Polacek <polacek@redhat.com>
11340 * Makefile.in (ubsan.o): Add.
11341 (c-family/c-ubsan.o): Add.
11342 (builtins.o): Add ubsan.h dependency.
11343 * ubsan.h: New file.
11344 * ubsan.c: New file.
11345 * common.opt: Add -fsanitize=undefined option.
11346 (flag_sanitize): Add variable.
11347 (fsanitize=): Add option. Add Driver.
11348 (fsanitize=thread): Remove option.
11349 (fsanitize=address): Likewise.
11350 (static-libubsan): New option.
11351 * doc/invoke.texi: Document the new flag and -static-libubsan.
11352 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
11353 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
11354 * builtin-attrs.def (ATTR_COLD): Define.
11355 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
11356 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
11357 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
11358 * flag-types.h (sanitize_code): New enum.
11359 * opts.c (common_handle_option): Parse command line arguments
11360 of -fsanitize=. Add -fsanitize=unreachable option.
11361 * varasm.c (get_variable_section): Adjust.
11362 (assemble_noswitch_variable): Likewise.
11363 (assemble_variable): Likewise.
11364 (output_constant_def_contents): Likewise.
11365 (categorize_decl_for_section): Likewise.
11366 (place_block_symbol): Likewise.
11367 (output_object_block): Likewise.
11368 * builtins.def: Likewise.
11369 * toplev.c (compile_file): Likewise.
11370 (process_options): Likewise.
11371 * cppbuiltin.c: Likewise.
11372 * tsan.c (tsan_pass): Likewise.
11373 (tsan_gate): Likewise.
11374 (tsan_gate_O0): Likewise.
11375 * cfgexpand.c (partition_stack_vars): Likewise.
11376 (expand_stack_vars): Likewise.
11377 (defer_stack_allocation): Likewise.
11378 (expand_used_vars): Likewise.
11379 * cfgcleanup.c (old_insns_match_p): Likewise.
11380 * asan.c (asan_finish_file): Likewise.
11381 (asan_instrument): Likewise.
11382 (gate_asan): Likewise.
11383 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
11384 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
11385 (asan_global_struct): Use pointer_sized_int_node instead
11386 calling build_nonstandard_integer_type.
11387 (initialize_sanitizer_builtins): Likewise.
11388 (asan_finish_file): Likewise.
11389 * gcc.c: Document %{%:function(args):X}.
11390 (static_spec_functions): Add sanitize.
11391 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
11392 store funcval != NULL there.
11393 (do_spec_1): Adjust handle_spec_function caller.
11394 (handle_braces): Allow %:function(args) as condition.
11395 (sanitize_spec_function): New function.
11396 (ADD_STATIC_LIBUBSAN_LIBS): Define.
11397 (LIBUBSAN_SPEC): Likewise.
11398 (LIBUBSAN_EARLY_SPEC): Likewise.
11399 (SANITIZER_SPEC): Handle libubsan.
11400 (SANITIZER_EARLY_SPEC): Likewise.
11401 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
11402 instead of fsanitize=address.
11403 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
11404 instead of fsanitize=address*.
11405 * builtins.c: Include ubsan.h.
11406 (fold_builtin_0): Instrument __builtin_unreachable.
11407 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
11408 instead of flag_asan.
11409 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
11410 (pointer_sized_int_node): Define.
11411 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
11413 2013-08-30 Mike Stump <mikestump@comcast.net>
11415 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
11418 2013-08-29 Jan Hubicka <jh@suse.cz>
11420 * cgraph.c (cgraph_function_body_availability): Handle weakref
11422 * passes.def: Remove pass_fixup_cfg.
11423 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
11424 track when we need to remove functions.
11425 (gate_ipa_inline): Execute inlining always; add comment why.
11426 (pass_data_ipa_inline): Remove TODO_remove_functions.
11427 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
11428 do not produce summaries.
11429 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
11430 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
11431 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
11432 constant pool and vtable.
11434 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
11436 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
11437 New arm_neon.h's internal macros to specify 64-bit constants.
11438 Avoid using stdint.h's macros.
11440 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
11442 * recog.c (verify_changes): Verify that changes[i].old is non-zero
11443 before applying REG_P.
11445 2013-08-30 Jakub Jelinek <jakub@redhat.com>
11447 PR tree-optimization/58277
11448 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
11449 after seeing too many stmts with vdef in between dombb and current
11450 bb, invalidate everything.
11452 2013-08-30 Richard Biener <rguenther@suse.de>
11454 * fold-const.c (fold_single_bit_test): Fix overflow test.
11456 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
11458 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
11459 and which can live in a register, always retrieve the value on entry.
11460 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
11461 passed by invisible reference specially.
11462 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
11463 (vt_add_function_parameter): Correctly deal with a parameter passed by
11464 invisible reference.
11466 2013-08-30 Jan Hubicka <jh@suse.cz>
11468 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
11470 2013-08-30 Richard Biener <rguenther@suse.de>
11472 PR tree-optimization/58228
11473 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
11474 allow invariant loads in nested loop vectorization.
11476 2013-08-30 Richard Biener <rguenther@suse.de>
11478 PR tree-optimization/58223
11479 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
11480 (has_anti_or_output_dependence): ... this and adjust to also
11481 look for output dependences.
11482 (mark_nodes_having_upstream_mem_writes): Adjust.
11483 (rdg_flag_uses): Likewise.
11485 2013-08-30 Richard Biener <rguenther@suse.de>
11487 PR tree-optimization/58010
11488 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
11489 assert that we have a loop-closed PHI.
11491 2013-08-29 Jan Hubicka <jh@suse.cz>
11493 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
11494 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
11495 * lto-section-in.c (lto_free_function_in_decl_state): New function.
11496 (lto_free_function_in_decl_state_for_node): New function.
11498 2013-08-29 Xinliang David Li <davidxl@google.com>
11500 * loop-unroll.c (report_unroll_peel): Minor message change.
11501 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
11502 Emit alignment peeling message with default -fopt-info.
11503 (vect_loop_versioning): Emit loop version info message.
11504 * tree-vectorizer.c (vectorize_loops): Minor message change.
11505 (execute_vect_slp): Ditto.
11507 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
11509 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
11510 of the clone from the DECL_NAME of the original function.
11512 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
11514 * passes.c (register_pass): Add overload.
11515 * tree-pass.h (register_pass): Forward declare it. Add comment.
11517 2013-08-29 Jan Hubicka <jh@suse.cz>
11519 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
11520 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
11521 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
11522 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11523 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
11524 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
11525 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11526 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
11528 2013-08-29 Teresa Johnson <tejohnson@google.com>
11530 * dumpfile.c (dump_loc): Output column number.
11531 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
11532 * doc/invoke.texi: Document optall -fopt-info flag.
11533 * profile.c (read_profile_edge_counts): Use new dump framework.
11534 (compute_branch_probabilities): Ditto.
11535 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
11536 when pass not in any opt group.
11537 * pass_manager.h (pass_manager::get_pass_profile): New method.
11538 * value-prof.c (check_counter): Use new dump framework.
11539 (check_ic_target): Ditto.
11540 * coverage.c (get_coverage_counts): Ditto.
11541 (coverage_init): Setup new dump framework.
11543 2013-08-29 Richard Biener <rguenther@suse.de>
11545 PR tree-optimization/58246
11546 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
11547 handle the dominance check inside a basic-block.
11549 2013-08-29 Richard Biener <rguenther@suse.de>
11551 PR middle-end/57287
11552 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
11553 of default defs that appear in abnormal PHI nodes.
11555 2013-08-29 Richard Biener <rguenther@suse.de>
11557 PR tree-optimization/57685
11558 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
11559 single-use operands to avoid exponential complexity.
11561 2013-08-28 Dehao Chen <dehao@google.com>
11563 * ipa-inline.c (edge_badness): Fix integer underflow.
11565 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
11567 * gtm-builtins.def (_ITM_free): Declare leaf.
11569 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11572 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
11573 (*tls_local_dynamic_base_64_largepic): Likewise.
11574 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
11575 Remove predicate from call operand.
11576 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
11577 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
11579 2013-08-28 Jeff Law <law@redhat.com>
11581 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
11582 checks for the number of predecessors and successors allowed.
11583 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
11584 which require copying a joiner block if there is a request which
11585 is a subpath that requires no joiner block copying.
11587 2013-08-28 Jan Hubicka <jh@suse.cz>
11589 * lto-streamer-out.c (DFS_write_tree_body): Drop
11590 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
11591 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
11592 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
11593 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
11594 Do not read DECL_ERROR_ISSUED.
11595 (unpack_ts_decl_with_vis_value_fields): Do not read
11597 (lto_input_ts_binfo_tree_pointers): Do not read
11598 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
11599 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
11600 write DECL_ERROR_ISSUED..
11601 (pack_ts_decl_with_vis_value_fields): Do not write
11603 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
11604 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
11605 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
11606 * tree.h (tree_decl_common): Update comment.
11607 (DECL_ERROR_ISSUED): Remove.
11609 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11611 PR middle-end/58257
11612 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
11614 2013-08-28 Jan Hubicka <jh@suse.cz>
11616 * builtins.def (free): Declare leaf.
11618 2013-08-27 David Malcolm <dmalcolm@redhat.com>
11620 * gdbhooks.py: New.
11621 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
11622 * configure: Regenerate.
11624 2013-08-27 Martin Jambor <mjambor@suse.cz>
11626 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
11627 (ipa_ancestor_jf_data): Likewise.
11628 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
11629 (ipa_get_jf_pass_through_type_preserved): New function.
11630 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
11631 (ipa_get_jf_ancestor_type_preserved): New function.
11632 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
11633 (ipa_get_jf_ancestor_result): Likewise.
11634 (propagate_vals_accross_pass_through): Use
11635 ipa_get_jf_pass_through_result to do all the value mappings.
11636 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
11637 type_preserved flag.
11638 (ipa_set_jf_cst_copy): New function.
11639 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
11640 (ipa_set_jf_arith_pass_through): Likewise.
11641 (ipa_set_ancestor_jf): Likewise.
11642 (compute_complex_assign_jump_func): Set type_preserved instead of
11644 (ipa_compute_jump_functions_for_edge): Likewise.
11645 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
11646 (update_jump_functions_after_inlining): Update type_preserved.
11647 Explicitely create jump functions when combining one with pass_through.
11648 (ipa_write_jump_function): Stream the type_preserved flags.
11649 (ipa_read_jump_function): Likewise.
11651 2013-08-27 Jakub Jelinek <jakub@redhat.com>
11652 Aldy Hernandez <aldyh@redhat.com>
11654 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
11655 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
11656 * function.h (struct function): Add has_force_vect_loops and
11658 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
11659 * gimple.c (gimple_build_omp_critical): Add KIND argument and
11661 * gimple.def: Update CLAUSES comments.
11662 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
11663 (gimple_build_omp_for): Add argument to prototype.
11664 (gimple_omp_for_kind): New.
11665 (gimple_omp_for_set_kind): New.
11666 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
11667 GOVD_DATA_SHARE_CLASS.
11668 (enum omp_region_type): Add ORT_SIMD.
11669 (gimple_add_tmp_var): Handle ORT_SIMD.
11670 (gimplify_var_or_parm_decl): Same.
11671 (is_gimple_stmt): Same.
11672 (omp_firstprivatize_variable): Same.
11673 (omp_add_variable): Only use splay_tree_insert if lookup failed.
11674 (omp_notice_variable): Handle ORT_SIMD.
11675 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
11676 (omp_check_private): Handle ORT_SIMD.
11677 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11678 OMP_CLAUSE_SAFELEN.
11679 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
11680 Handle OMP_CLAUSE_LASTPRIVATE.
11681 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11682 OMP_CLAUSE_SAFELEN.
11683 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
11684 (gimplify_expr): Handle OMP_SIMD.
11685 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
11686 (expand_GOMP_SIMD_VF): New.
11687 (expand_GOMP_SIMD_LAST_LANE): New.
11688 * internal-fn.def (GOMP_SIMD_LANE): New.
11689 (GOMP_SIMD_VF): New.
11690 (GOMP_SIMD_LAST_LANE): New.
11691 * omp-low.c: Include target.h.
11692 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
11693 OMP_CLAUSE_SAFELEN.
11694 (check_omp_nesting_restrictions): Same.
11696 (lower_rec_simd_input_clauses): New.
11697 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
11699 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
11700 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
11701 (expand_omp_build_assign): New.
11702 (expand_omp_for_init_counts): New.
11703 (expand_omp_for_init_vars): New.
11704 (extract_omp_for_update_vars): New.
11705 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
11706 and rewrite accordingly.
11707 (expand_omp_simd): New.
11708 (expand_omp_for): Use expand_omp_simd.
11709 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
11710 (lower_omp_for): Do not lower the body.
11711 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
11712 in their own loops.
11713 * tree-flow.h (find_omp_clause): Remove prototype.
11714 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
11715 forcing vectorization of the loop, or if flag_tree_vectorize.
11716 (gate_tree_if_conversion): Similarly.
11717 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
11718 gimple_build_omp_for.
11719 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
11720 * tree-parloops (create_parallel_loop): Pass kind argument to
11721 gimple_build_omp_for.
11722 * tree-pretty-print.c (dump_omp_clause): Add cases for
11723 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
11724 OMP_CLAUSE__SIMDUID_.
11725 (dump_generic_node): Handle OMP_SIMD.
11726 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
11727 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
11728 unroll OMP_SIMD loops here.
11729 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
11730 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
11732 (vect_analyze_data_refs): Handle simd loops.
11733 * tree-vect-loop.c (vectorizable_live_operation): Handle
11735 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
11736 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
11737 (vectorizable_load): Same.
11738 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
11739 (struct simduid_to_vf): New.
11740 (simduid_to_vf::hash): New.
11741 (simduid_to-vf::equal): New.
11742 (struct simd_array_to_simduid): New.
11743 (simd_array_to_simduid::hash): New.
11744 (simd_array_to_simduid::equal): New.
11745 (adjust_simduid_builtins): New.
11746 (struct note_simd_array_uses_struct): New.
11747 (note_simd_array_uses_cb): New.
11748 (note_simd_array_uses): New.
11749 (vectorize_loops): Handle simd hints and adjust simd builtins
11751 * tree-vectorizer.h (struct _stmt_vec_info): Add
11752 simd_lane_access_p field.
11753 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
11754 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
11755 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
11756 (omp_clause_code_name): Same.
11757 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
11758 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
11759 * tree.def (OMP_SIMD): New entry.
11760 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
11761 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
11762 (OMP_CLAUSE_DECL): Adjust range for new clauses.
11763 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
11764 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
11765 (OMP_CLAUSE_LINEAR_STEP): New.
11766 (OMP_CLAUSE_SAFELEN_EXPR): New.
11767 (OMP_CLAUSE__SIMDUID__DECL): New.
11768 (find_omp_clause): New prototype.
11770 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
11772 * config/i386/driver-i386.c (host_detect_local_cpu): Update
11773 Haswell processor detection.
11775 2013-08-27 Christian Widmer <shadow@umbrox.de>
11778 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
11779 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
11780 AVX2 capable processors.
11782 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
11784 * config/aarch64/arm_neon.h: Replace all inline asm implementations
11785 of vget_low_* with implementations in terms of other intrinsics.
11787 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
11789 PR middle-end/57219
11790 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
11791 values to -1, 0 and 1.
11793 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
11795 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
11796 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
11797 (<optab><mode>3_insn): Remove.
11798 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
11799 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
11800 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
11801 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
11802 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
11803 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
11804 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
11805 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
11806 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
11807 (ror<mode>3_insn): Likewise.
11808 * config/aarch64/predicates.md (aarch64_simd_register): New.
11810 2013-08-27 Richard Biener <rguenther@suse.de>
11812 PR tree-optimization/57521
11813 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
11814 one edge is non-critical.
11815 (find_phi_replacement_condition): Make sure to use a non-critical
11816 edge. Cleanup and remove old bug workarounds.
11817 (bb_postdominates_preds): Remove.
11818 (if_convertible_loop_p_1): Do not compute post-dominators.
11819 (combine_blocks): Do not free post-dominators.
11820 (main_tree_if_conversion): Likewise.
11821 (pass_data_if_conversion): Add TODO_verify_ssa.
11823 2013-08-27 DJ Delorie <dj@redhat.com>
11825 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
11827 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
11829 * function.c (assign_parm_find_data_types): Set passed_mode and
11830 nominal_mode to the TYPE_MODE of nominal_type for the built
11831 pointer type in case of the struct-pass-by-reference.
11833 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11835 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
11836 (UINT16_TYPE): Change default to "unsigned int".
11838 * config/avr/avr.opt (mfract-convert-truncate): New option.
11839 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
11840 is set, round negative fractional integers according to n1169
11841 when converting to integer types.
11843 2013-08-26 Jan Hubicka <jh@suse.cz>
11845 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
11846 methods can not be called indirectly when their address is not taken.
11848 2013-08-26 Jan Hubicka <jh@suse.cz>
11850 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
11853 2013-08-26 Jan Hubicka <jh@suse.cz>
11855 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11858 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11860 * reload.c (find_valid_class): Allow classes that do not include
11861 FIRST_PSEUDO_REGISTER - 1.
11863 2013-08-26 Jan Hubicka <jh@suse.cz>
11865 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11866 fix edge count/frequency when speculation failed; fix type check
11867 for the direct call.
11869 2013-08-26 Jan Hubicka <jh@suse.cz>
11871 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11873 2013-08-26 Jan Hubicka <jh@suse.cz>
11875 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11876 to be changed by edge redirection.
11878 2013-08-26 Jan Hubicka <jh@suse.cz>
11880 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11881 formating; add sanity check.
11882 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11883 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11884 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11885 (resolve_noninline_speculation): Update callee keys, too.
11887 2013-08-26 Jan Hubicka <jh@suse.cz>
11889 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11890 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11892 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11894 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11897 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11899 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11900 (x86_64_elf_select_section): Ditto.
11901 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11903 (ix86_pass_by_reference): Ditto.
11904 (output_set_got): Ditto.
11905 (ix86_unary_operator_ok): Ditto.
11906 (ix86_expand_builtin): Ditto.
11908 2013-08-23 Jan Hubicka <jh@suse.cz>
11910 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11912 2013-08-23 Jan Hubicka <jh@suse.cz>
11914 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11915 (tree_decl_with_vis): Add FINAL field.
11917 2013-08-23 Jeff Law <law@redhat.com>
11919 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11920 the debugging dump when the expression is fully redundant.
11922 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11924 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11925 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11926 * pretty-print.c (pp_formatted_text_data): Likewise.
11927 (pp_write_text_to_stream): Likewise.
11928 (pp_write_text_as_dot_label_to_stream): Likewise.
11929 (pp_append_r): Likewise.
11930 (pp_format): Likewise.
11931 (pp_flush): Likewise.
11932 (pp_clear_output_area): Likewise.
11933 (pp_append_text): Likewise.
11934 (pp_formatted_text): Likewise.
11935 (pp_remaining_character_count_for_line): Likewise.
11936 (pp_newline): Likewise.
11937 (pp_character): Likewise.
11938 (output_buffer::~output_buffer): Define.
11939 (pretty_printer::~pretty_printer): Destruct output buffer.
11940 * pretty-print.h (output_buffer::~output_buffer): Declare.
11941 (pretty_printer::~pretty_printer): Declare virtual.
11943 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11946 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11947 HOST_WIDE_INT_M1U): New macros.
11948 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11949 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11950 unsigned -1 for lshift.
11951 * cse.c (cse_insn): Likewise.
11952 * double-int.c (rshift_double, lshift_double): Likewise.
11953 * builtins.c (fold_builtin_bitop): Likewise.
11954 * combine.c (force_to_mode): Likewise.
11955 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11956 * simplify-rtx.c (simplify_const_unary_operation,
11957 simplify_const_binary_operation): Likewise.
11958 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11959 check_va_list_escapes): Likewise.
11960 * rtlanal.c (nonzero_bits1): Likewise.
11961 * expmed.c (expand_smod_pow2): Likewise.
11962 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11964 2013-08-23 Jan Hubicka <jh@suse.cz>
11966 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11967 as having address taken.
11969 2013-08-23 Jan Hubicka <jh@suse.cz>
11971 * ipa-utils.h (method_class_type): Declare.
11972 * ipa-devirt.c (method_class_type): Export.
11974 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11975 do not walk base classes of anonymous types.
11977 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11979 PR rtl-optimization/58220
11980 PR regression/58221
11981 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11982 handle SEQUENCE insns properly.
11984 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11986 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11988 (pp_newline_and_indent): Likewise.
11989 (pp_separate_with): Likewise.
11990 * pretty-print.c (pp_newline_and_flush): Define.
11991 (pp_newline_and_indent): Likewise.
11992 (pp_separate_with): Likewise.
11994 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11997 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11998 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
12000 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
12002 * config/i386/predicates.md (ext_sse_reg_operand): New.
12003 * config/i386/i386.md (*movti_internal): Use
12004 predicate to determine if EVEX is needed.
12005 (*movsi_internal): Ditto.
12006 (*movdf_internal): Ditto.
12007 (*movsf_internal): Ditto.
12008 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
12010 2013-08-23 Jakub Jelinek <jakub@redhat.com>
12012 PR tree-optimization/58209
12013 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
12014 (find_tail_calls): Give up for pointer result types if m is non-NULL.
12015 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
12016 emit POINTER_PLUS_EXPR.
12017 (create_tailcall_accumulator): For pointer result type accumulate in
12020 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
12022 * configure.ac: Add backslashes missing from the last change.
12023 * configure: Regenerate.
12025 2013-08-22 Jan Hubicka <jh@susue.cz>
12027 * ipa.c (function_and_variable_visibility): First remember function
12028 was global and then make it local.
12030 2013-08-22 Julian Brown <julian@codesourcery.com>
12032 * configure.ac: Add aarch64 to list of arches which use "nop" in
12034 * configure: Regenerate.
12036 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12038 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
12039 gnu_libc_has_function.
12040 * config/s390/tpf.h: Likewise.
12042 2013-08-22 Jan Hubicka <jh@susue.cz>
12044 * timevar.c (validate_phases): Add cast.
12046 2013-08-22 Jan Hubicka <jh@susue.cz>
12048 * timevar.c (validate_phases): Use size_t for memory.
12049 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
12051 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
12053 * pretty-print.h (output_buffer::output_buffer): Declare.
12054 (pretty_printer::pretty_printer): Likewise.
12055 (pp_construct): Remove.
12056 * pretty-print.c (output_buffer::output_buffer): Define.
12057 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
12058 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
12059 (print_gimple_expr): Likewise.
12060 (print_gimple_seq): Likewise.
12061 (gimple_dump_bb): Likewise.
12062 * sched-vis.c (dump_value_slim): Likewise.
12063 (dump_insn_slim): Likewise.
12064 (dump_rtl_slim): Likewise.
12065 (str_pattern_slim): Likewise.
12066 * tree-mudflap.c (mf_varname_tree): Likewise.
12067 * graph.c (print_graph_cfg): Likewise.
12068 (start_graph_dump): Likewise.
12069 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
12071 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
12072 pretty printer initialization.
12073 * coretypes.h (diagnostic_context): Remove superflous type alias
12075 (pretty_printer): Likewise. Declare directly as a class.
12076 (pretty_print_info): Remove declaration as class.
12077 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
12078 and pp_clear_output_area.
12079 (asan_add_global): Likewise.
12081 2013-08-22 Jan Hubicka <jh@suse.cz>
12083 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
12084 * ipa-utils.h (update_type_inheritance_graph): Declare.
12085 (possible_polymorphic_call_target_p): Declare.
12086 (possible_polymorphic_call_target_p): New.
12087 * ipa-devirt.c: Update toplevel comments.
12088 (cached_polymorphic_call_targets): Move up.
12089 (odr_type_d): Move ID down.
12090 (polymorphic_type_binfo_p): Update comment.
12091 (odr_hasher::remove): Likewise;
12092 (get_odr_type): Set anonymous_namespace.
12093 (dump_odr_type): Dump it.
12094 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
12095 (maybe_record_node): Record node in cached_polymorphic_call_targets.
12096 (record_binfo): Add comment.
12097 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
12099 (devirt_node_removal_hook): Do not iCE when cache is freed.
12100 (possible_polymorphic_call_target_p): New predicate.
12101 (update_type_inheritance_graph): New function.
12103 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
12104 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
12105 Sergey Lega <sergey.s.lega@intel.com>
12106 Anna Tikhonova <anna.tikhonova@intel.com>
12107 Ilya Tocar <ilya.tocar@intel.com>
12108 Andrey Turetskiy <andrey.turetskiy@intel.com>
12109 Ilya Verbin <ilya.verbin@intel.com>
12110 Kirill Yukhin <kirill.yukhin@intel.com>
12111 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12113 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
12114 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
12115 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
12116 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
12117 (OPTION_MASK_ISA_AVX2_UNSET): Update.
12118 (OPTION_MASK_ISA_AVX512F_UNSET): New.
12119 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
12120 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
12121 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
12122 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
12123 OPT_mavx512pf, OPT_mavx512er cases.
12124 * config/i386/constraints.md (v): New constraint.
12125 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
12126 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
12127 (bit_AVX512CD): New.
12128 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
12129 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
12130 * config/i386/i386-c.c (ix86_target_macros_internal):
12131 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
12133 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
12134 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
12135 * config/i386/i386.c (regclass_map, dbx_register_map)
12136 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
12137 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
12138 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
12139 -mavx512pf options.
12140 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
12141 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
12142 -mavx512pf options. Fix formatting.
12143 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
12144 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
12145 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
12146 -mavx512cd, -mavx512pf options.
12147 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
12148 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
12149 (ix86_preferred_output_reload_class): Replace SSE_REGS with
12151 (ix86_hard_regno_mode_ok): Support 512-bit registers.
12152 (ix86_set_reg_reg_cost): Ditto.
12153 (x86_order_regs_for_local_alloc): Ditto.
12154 (MAX_VECT_LEN): Extend to 64-byte.
12155 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
12156 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
12157 (TARGET_AVX512ER, TARGET_AVX512CD): New.
12158 (BIGGEST_ALIGNMENT): Extend to 512-bits.
12159 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
12160 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
12161 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
12162 (SSE_REG_MODE_P): Support new modes.
12163 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
12164 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
12165 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
12166 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
12167 (REG_CLASS_CONTENTS): Add new registers.
12168 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
12169 (EXT_REX_SSE_REGNO_P): New.
12170 (HI_REGISTER_NAMES): Add new registers.
12171 * config/i386/i386.md: Define constants for new registers.
12172 (mode): Add new 512-bit modes.
12173 (prefix): Support evex prefix.
12174 (isa): Support avx512f, noavx512f, fma_avx512f.
12175 (ssemodesuffix): Add new 512-bit modes.
12177 (*movxi_internal_avx512f): Ditto.
12178 (*movdi_internal): Replace constraint "x" with the new constraint "v".
12180 (*movsi_internal): Likewise.
12181 (*movdf_internal): Likewise.
12182 (*movsf_internal): Likewise.
12183 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
12184 (<code><mode>3): Likewise.
12185 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
12187 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
12188 with the new constraint "v".
12189 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
12191 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
12192 with the new constraint "v".
12193 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
12194 (<sse2>_storedqu<avxsizesuffix>): Likewise.
12195 (*<plusminus_insn><mode>3): Likewise.
12196 (<sse>_vm<plusminus_insn><mode>3): Likewise.
12197 (*mul<mode>3): Likewise.
12198 (<sse>_vmmul<mode>3): Likewise.
12199 (<sse>_div<mode>3): Likewise.
12200 (<sse>_vmdiv<mode>3): Likewise.
12201 (<sse>_sqrt<mode>2): Likewise.
12202 (<sse>_vmsqrt<mode>2): Likewise.
12203 (*<code><mode>3_finite): Likewise.
12204 (*<code><mode>3) <smaxmin>: Likewise.
12205 (<sse>_vm<code><mode>3): Likewise.
12206 (*<code><mode>3) <any_logic>: Likewise.
12207 (*fma_fmadd_<mode>): Likewise.
12208 (*fma_fmsub_<mode>): Likewise.
12209 (*fma_fnmadd_<mode>): Likewise.
12210 (*fma_fnmsub_<mode>): Likewise.
12211 (*fma_fmaddsub_<mode>): Likewise.
12212 (*fma_fmsubadd_<mode>): Likewise.
12213 (*fmai_fmadd_<mode>): Likewise.
12214 (*fmai_fmsub_<mode>): Likewise.
12215 (*fmai_fnmadd_<mode>): Likewise.
12216 (*fmai_fnmsub_<mode>): Likewise.
12217 (sse_cvtsi2ss): Likewise.
12218 (sse_cvtsi2ssq): Likewise.
12219 (sse_cvtss2si): Likewise.
12220 (sse_cvtss2si_2): Likewise.
12221 (sse_cvtss2siq): Likewise.
12222 (sse_cvtss2siq_2): Likewise.
12223 (sse_cvttss2si): Likewise.
12224 (sse_cvtss2siq_2): Likewise.
12225 (float<sseintvecmodelower><mode>2): Likewise.
12226 (sse2_cvtsd2si_2): Likewise.
12227 (sse2_cvtsd2siq_2): Likewise.
12228 (*<plusminus_insn><mode>3): Likewise.
12229 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
12230 (*<sse4_1_avx2>_mul<mode>3): Likewise.
12231 (ashr<mode>3): Likewise.
12232 (<shift_insn><mode>3): Likewise.
12233 (avx2_<code><mode>3): Likewise.
12234 (*avx2_<code><mode>3): Likewise.
12235 (*andnot<mode>3): Likewise.
12236 (*<code><mode>3) <any_logic>: Likewise.
12237 (abs<mode>2): Likewise.
12238 (avx2_permvar<mode>): Likewise.
12239 (avx2_perm<mode>_1): Likewise.
12240 (*avx_vpermilp<mode>): Likewise.
12241 (avx_vpermilvar<mode>3): Likewise.
12242 (avx2_ashrv<mode>): Likewise.
12243 (avx2_<shift_insn>v<mode>): Likewise.
12244 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
12246 * doc/rtl.texi: Document XImode.
12248 2013-08-21 Jeff Law <law@redhat.com>
12250 * tree-flow.h (register_jump_thread): Pass vector of edges
12251 instead of each important edge.
12252 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
12253 thread path into a vector and pass that to register_jump_thread.
12254 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
12255 passed in edge vector to the current 3-edge form.
12258 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
12260 * dce.c (fini_dce): Call df_analyze again just in case
12261 delete_unmarked_insns removed anything.
12263 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
12265 * reload.h (struct reg_equivs): Rename to ..
12266 (struct reg_equivs_s): .. this.
12268 2013-08-20 Martin Liska <marxin.liska@gmail.com>
12270 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
12272 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12274 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
12276 2013-08-21 Jeff Law <law@redhat.com>
12278 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
12279 simplify assignments too. If the RHS collapses to a singleton
12280 range, then return the value for the range.
12282 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
12284 * config/i386/sse.md (V16): Rename to...
12286 (mov<mode>): Update iterator name.
12287 (*mov<mode>_internal): Ditto.
12288 (push<mode>1): Ditto.
12289 (movmisalign<mode>): Ditto.
12291 2013-08-20 Jan Hubicka <jh@suse.cz>
12294 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
12295 entry for direct edges.
12296 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
12298 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12300 Revert my last two changes, r201865 and r201864:
12303 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12305 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
12306 instances can own GC refs.
12308 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
12309 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
12310 (gcc::context::gt_pch_nx): Likewise.
12311 (gcc::context::gt_pch_nx): Likewise.
12312 * ggc.h (gt_ggc_mx <T>): New.
12313 (gt_pch_nx_with_op <T>): New.
12314 (gt_pch_nx <T>): New.
12315 * passes.c (opt_pass::gt_ggc_mx): New.
12316 (opt_pass::gt_pch_nx): New.
12317 (opt_pass::gt_pch_nx_with_op): New.
12318 (pass_manager::gt_ggc_mx): New.
12319 (pass_manager::gt_pch_nx): New.
12320 (pass_manager::gt_pch_nx_with_op): New.
12321 (pass_manager::operator new): Use
12322 ggc_internal_cleared_alloc_stat rather than xcalloc.
12323 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
12324 (pass_manager::gt_ggc_mx): New.
12325 (pass_manager::gt_pch_nx): New.
12326 (pass_manager::gt_pch_nx_with_op): New.
12327 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
12328 (opt_pass::operator new): New.
12329 (opt_pass::gt_ggc_mx): New.
12330 (opt_pass::gt_pch_nx): New.
12331 (opt_pass::gt_pch_nx_with_op): New.
12334 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12336 * Makefile.in (GTFILES): Add context.h.
12337 * context.c (gcc::context::operator new): New.
12338 (gcc::context::gt_ggc_mx): New.
12339 (gcc::context::gt_pch_nx): New.
12340 (gcc::context::gt_pch_nx): New.
12341 * context.h (gcc::context): Add GTY((user)) marking.
12342 (gcc::context::operator new): New.
12343 (gcc::context::gt_ggc_mx): New.
12344 (gcc::context::gt_pch_nx): New.
12345 (gcc::context::gt_pch_nx): New.
12346 (g): Add GTY marking.
12347 (gt_ggc_mx (gcc::context *)): New.
12348 (gt_pch_nx (gcc::context *)): New.
12349 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
12350 void *cookie)): New.
12351 * gengtype.c (open_base_files) <ifiles>: Add context.h.
12353 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
12355 * dce.c (fini_dce): Call df_analyze again just in case
12356 delete_unmarked_insns removed anything.
12358 2013-08-20 Teresa Johnson <tejohnson@google.com>
12360 PR rtl-optimizations/57451
12361 * final.c (reemit_insn_block_notes): Prevent lexical blocks
12362 from crossing split section boundaries.
12364 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12366 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
12367 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
12368 with MULTLIB_DEFAULTS.
12370 2013-08-20 Nick Clifton <nickc@redhat.com>
12372 * target.def (narrow_volatile_bitfield): Note that the default
12373 value is false, not !TARGET_STRICT_ALIGN.
12374 * doc/tm.texi: Regenerate.
12376 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
12378 Fix LIB_SPEC for systems without libpthread.
12380 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
12381 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
12383 * config/i386/linux-common.h: Likewise.
12384 * config/mips/linux-common.h: Likewise.
12386 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12388 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
12391 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12393 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
12394 instances can own GC refs.
12396 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
12397 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
12398 (gcc::context::gt_pch_nx): Likewise.
12399 (gcc::context::gt_pch_nx): Likewise.
12400 * ggc.h (gt_ggc_mx <T>): New.
12401 (gt_pch_nx_with_op <T>): New.
12402 (gt_pch_nx <T>): New.
12403 * passes.c (opt_pass::gt_ggc_mx): New.
12404 (opt_pass::gt_pch_nx): New.
12405 (opt_pass::gt_pch_nx_with_op): New.
12406 (pass_manager::gt_ggc_mx): New.
12407 (pass_manager::gt_pch_nx): New.
12408 (pass_manager::gt_pch_nx_with_op): New.
12409 (pass_manager::operator new): Use
12410 ggc_internal_cleared_alloc_stat rather than xcalloc.
12411 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
12412 (pass_manager::gt_ggc_mx): New.
12413 (pass_manager::gt_pch_nx): New.
12414 (pass_manager::gt_pch_nx_with_op): New.
12415 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
12416 (opt_pass::operator new): New.
12417 (opt_pass::gt_ggc_mx): New.
12418 (opt_pass::gt_pch_nx): New.
12419 (opt_pass::gt_pch_nx_with_op): New.
12421 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12423 * Makefile.in (GTFILES): Add context.h.
12424 * context.c (gcc::context::operator new): New.
12425 (gcc::context::gt_ggc_mx): New.
12426 (gcc::context::gt_pch_nx): New.
12427 (gcc::context::gt_pch_nx): New.
12428 * context.h (gcc::context): Add GTY((user)) marking.
12429 (gcc::context::operator new): New.
12430 (gcc::context::gt_ggc_mx): New.
12431 (gcc::context::gt_pch_nx): New.
12432 (gcc::context::gt_pch_nx): New.
12433 (g): Add GTY marking.
12434 (gt_ggc_mx (gcc::context *)): New.
12435 (gt_pch_nx (gcc::context *)): New.
12436 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
12437 void *cookie)): New.
12438 * gengtype.c (open_base_files) <ifiles>: Add context.h.
12440 2013-08-20 Alan Modra <amodra@gmail.com>
12443 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
12444 (rs6000_emit_epilogue): Likewise.
12446 2013-08-19 Dehao Chen <dehao@google.com>
12448 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
12450 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
12451 Jakub Jelinek <jakub@redhat.com>
12453 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
12454 (BUILT_IN_FABSD64): Likewise.
12455 (BUILT_IN_FABSD128): Likewise.
12456 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
12457 (fold_builtin_1): Likewise.
12458 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
12459 destination and source operands.
12460 (*abstd2_fpr): Likewise.
12461 (*nabstd2_fpr): Likewise.
12463 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
12465 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
12468 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
12470 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
12473 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
12475 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
12476 * builtins.c (default_libc_has_function): New.
12477 (gnu_libc_has_function): Ditto.
12478 (no_c99_libc_has_function): Ditto.
12479 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
12480 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
12481 (fold_builtin_sincos): Likewise.
12482 (fold_builtin_cexp): Likewise.
12483 * builtins.def (DEF_C94_BUILTIN): Likewise.
12484 (DEF_C99_BUILTIN): Likewise.
12485 (DEF_C99_C90RES_BUILTIN): Likewise.
12486 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
12487 definitions to using this define.
12488 * config/darwin-protos.h (darwin_libc_has_function): New.
12489 * config/darwin.c (darwin_libc_has_function): Ditto.
12490 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
12491 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
12492 * config/darwin.h: Ditto.
12493 * config/elfos.h: Ditto.
12494 * config/freebsd.h: Ditto.
12495 * config/i386/cygming.h: Ditto.
12496 * config/i386/djgpp.h: Ditto.
12497 * config/i386/i386-interix.h: Ditto.
12498 * config/microblaze/microblaze.h: Ditto.
12499 * config/mmix/mmix.h: Ditto.
12500 * config/gnu-user.h: Ditto.
12501 * config/ia64/hpux.h: Ditto.
12502 * config/pa/pa-hpux.h: Ditto.
12503 * config/pdp11/pdp11.h: Ditto.
12504 * config/picochip/picochip.h: Ditto.
12505 * config/linux.h: Ditto.
12506 * config/netbsd.h: Ditto.
12507 * config/openbsd.h: Ditto.
12508 * config/rs6000/aix43.h: Ditto.
12509 * config/rs6000/aix51.h: Ditto.
12510 * config/rs6000/aix52.h: Ditto.
12511 * config/rs6000/aix53.h: Ditto.
12512 * config/rs6000/aix61.h: Ditto.
12513 * config/rs6000/darwin.h: Ditto.
12514 * config/rs6000/linux.h: Ditto.
12515 * config/rs6000/linux64.h: Ditto.
12516 * config/s390/tpf.h: Ditto.
12517 * config/sol2-10.h: Ditto.
12518 * config/sol2.h: Ditto.
12519 * config/vms/vms.h: Ditto.
12520 * config/vxworks.h: Ditto.
12521 * config/linux-android.c (linux_android_libc_has_function):
12522 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
12523 * config/linux-protos.h (linux_android_libc_has_function):
12525 * config/i386/i386.c (ix86_libc_has_function): New.
12526 * config/i386/i386-protos.h
12527 (ix86_libc_has_function): New declaration.
12528 * config/i386/i386.md
12529 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
12530 ("isinf<mode>2): Likewise.
12531 * convert.c (convert_to_integer): Using new target hook
12532 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
12533 TARGET_C99_FUNCTIONS.
12534 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
12535 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
12536 * coretypes.h (function_class): New enum for different
12537 classes of functions.
12538 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
12539 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
12540 (TARGET_HAS_SINCOS): Likewise.
12541 (TARGET_LIBC_HAS_FUNCTION): New.
12542 * doc/tm.texi: Regenerated.
12543 * targhooks.h (default_libc_has_function): New declaration.
12544 (no_c99_libc_has_function): Ditto.
12545 (gnu_libc_has_function): Ditto.
12546 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
12547 and TARGET_HAS_SINCOS.
12549 2013-08-18 Jan Hubicka <jh@suse.cz>
12551 * Makeifle-in (ipa-devirt.o): New.
12552 (GTFILES): Add ipa-utils.h and ipa-devirt.c
12553 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
12554 (analyze_functions): Look into possible targets of polymorphic call.
12555 * dumpfile.c (dump_files): Add type-inheritance dump.
12556 * dumpfile.h (TDI_inheritance): New.
12557 * ipa-devirt.c: New file.
12558 * ipa-utils.h (odr_type_d): Forward declare.
12559 (odr_type): New type.
12560 (build_type_inheritance_graph): Declare.
12561 (possible_polymorphic_call_targets): Declare and introduce inline
12562 variant when only edge is pased.
12563 (dump_possible_polymorphic_call_targets): Likewise.
12564 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
12565 * tree.c (type_in_anonymous_namespace_p): Break out from ...
12566 (types_same_for_odr): ... here.
12567 * tree.h (type_in_anonymous_namespace_p): Declare.
12569 2013-08-18 Jakub Jelinek <jakub@redhat.com>
12571 PR tree-optimization/58006
12572 * tree-parloops.c (take_address_of): Don't ICE if get_name
12574 (eliminate_local_variables_stmt): Remove clobber stmts.
12576 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
12578 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
12579 error message is issued for an alias to undefined symbol.
12581 2013-08-18 Jan Hubicka <jh@suse.cz>
12583 * cgraph.c (cgraph_create_indirect_edge): Discover
12584 polymorphic calls and record basic info into indirect_info.
12585 * gimple-fold.c (gimple_fold_call): When doing BINFO based
12586 devirtualization, ignore objc function calls.
12587 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
12588 call with no parm index info.
12589 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
12590 * tree.c (virtual_method_call_p): New function.
12591 * tree.h (virtual_method_call_p): Declare.
12593 2013-08-16 Jan Hubicka <jh@suse.cz>
12595 PR middle-end/58179
12596 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
12598 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
12600 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
12603 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12605 * gengtype.c (type_for_name): Add special-case support for
12606 locating types within the "gcc::" namespace.
12607 (open_base_files): Emit a "using namespace gcc" directive.
12609 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12612 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
12613 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
12615 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
12616 array instead of each individual operand as a separate argument.
12617 (emit_fusion_gpr_load): Likewise.
12618 (expand_fusion_gpr_load): Add new function declaration.
12620 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
12621 signature to have the operands passed as an array, instead of as
12622 separate arguments. Allow ZERO_EXTEND to be in the memory
12623 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
12624 depend on the register live/dead flags when peepholes are run.
12625 (expand_fusion_gpr_load): New function to be called from the
12626 peephole2 pass, to change the register that addis sets to be the
12628 (emit_fusion_gpr_load): Change the calling signature to have the
12629 operands passed as an array, instead of as separate arguments.
12630 Allow ZERO_EXTEND to be in the memory address, and also
12631 SIGN_EXTEND if -mpower8-fusion-sign.
12633 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
12634 unspec enumeration.
12635 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
12636 adjust the register addis loads up in the peephole2 pass. Do not
12637 depend on the register live/dead state when the peephole pass is done.
12639 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12641 * gengtype.c (create_user_defined_type): Ensure that the kind
12642 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
12643 declaration is seen before the GTY((user)) marking.
12645 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
12648 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
12650 2013-08-16 Jan Hubicka <jh@suse.cz>
12652 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
12653 arugment expected_type.
12654 (gimple_fold_call): Use it.
12655 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
12656 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
12657 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
12658 (try_make_edge_direct_virtual_call): Likewise.
12659 * tree.c (obj_type_ref_class): New.
12660 * tree.h (obj_type_ref_class): Use it.
12662 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
12664 * sched-vis.c (rtl_slim_pp_initialized): Remove.
12665 (rtl_slim_pp): Likewise.
12666 (init_rtl_slim_pretty_print): Likewise.
12667 (dump_value_slim): Don't call it. Use local pretty printer.
12668 (dump_insn_slim): Likewise.
12669 (dump_rtl_slim): Likewise.
12670 (str_pattern_slim): Likewise.
12671 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
12674 2013-08-16 Jakub Jelinek <jakub@redhat.com>
12676 PR tree-optimization/58164
12677 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
12678 walk gimple_goto_dest of GIMPLE_GOTO.
12680 PR tree-optimization/58165
12681 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
12682 bi_call must be the last stmt in a bb, don't split_block, instead
12683 use fallthru edge from it and give up if there is none.
12684 Release conds vector when returning early.
12686 2013-08-14 Xinliang David Li <davidxl@google.com>
12688 * config/i386/i386.c (ix86_option_override_internal):
12689 Remove unused variable and field.
12691 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12694 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
12695 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
12696 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
12697 and Linux, correct BLKmode alignment when 128-bit alignment is
12698 required and compatibility flag is not set.
12699 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
12700 for zero-size arguments when compatibility flag is not set.
12702 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12704 PR tree-optimization/58145
12705 * tree-sra.c (build_ref_for_offset): If prev_base has
12706 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
12708 2013-08-14 Xinliang David Li <davidxl@google.com>
12710 * config/i386/i386.c (ix86_option_override_internal):
12711 Fix uninitialized variable error.
12713 2013-08-14 Xinliang David Li <davidxl@google.com>
12715 * config/i386/i386.opt: Define two new options.
12716 * config/i386/x86-tune.def: Add arch selector field in macros.
12717 * config/i386/i386.h: Adjust macro definition.
12718 * config/i386/i386.c (ix86_option_override_internal):
12720 (parse_mtune_ctrl_str): New function.
12721 (set_ix86_tune_features): New function.
12722 (ix86_function_specific_restore): Call the new helper function.
12724 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
12726 PR rtl-optimization/57662
12727 * sel-sched.c (code_motion_process_successors): When the current insn
12728 is removed after the recursive traversal, break from the loop.
12729 Add comments and debug printouts.
12731 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12732 Alexandre Oliva <aoliva@redhat.com>
12735 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
12736 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
12737 there also UNSPEC_PLTOFF.
12739 2013-08-14 Marek Polacek <polacek@redhat.com>
12741 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
12742 NUM_CONDITIONS bit positions.
12744 2013-08-13 Cary Coutant <ccoutant@google.com>
12746 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
12747 (attr_checksum): Hash vector contents instead of pointer.
12748 (attr_checksum_ordered): Likewise.
12750 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
12752 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
12753 when Pmode != word_mode. Add length_address attribute.
12754 (sse3_monitor_<mode>): Merge from sse3_monitor and
12755 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
12756 Pmode != word_mode. Update insn length attribute.
12757 * config/i386/i386.c (ix86_option_override_internal): Update
12758 ix86_gen_monitor selection for merged sse3_monitor insn.
12760 2013-08-13 Julian Brown <julian@codesourcery.com>
12762 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
12763 perform invalid legitimization on greater-than-word-size modes for
12764 TARGET_E500_DOUBLE.
12766 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
12768 * ira.c (setup_class_translate_array): Use aclass instead of cl
12769 for classes not fully covered by allocno classes.
12771 2013-08-13 Jakub Jelinek <jakub@redhat.com>
12773 PR tree-optimization/57661
12774 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
12775 * tree-inline.c (tree_function_versioning): Initialize it.
12776 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
12777 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
12778 that is not being copied.
12781 * asan.c (instrument_strlen_call): Fix typo in comment.
12782 Use char * type even for the lhs of POINTER_PLUS_EXPR.
12784 2013-08-13 Steve Ellcey <sellcey@mips.com>
12786 * config/mips/mips.md (prefetch): Use lw instead of ld on
12787 loongson in 32bit mode.
12789 2013-08-13 Nick Clifton <nickc@redhat.com>
12791 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
12793 2013-08-13 Jan Hubicka <jh@suse.cz>
12795 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
12796 introduced edge; fix typo in sanity check.
12797 (cgraph_resolve_speculation): Export; improve diagnostic.
12798 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
12799 speculation at type mismatch.
12800 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
12801 (cgraph_resolve_speculation): Declare.
12802 (symtab_can_be_discarded): New function.
12803 * value-prof.c (gimple_ic_transform): Remove actual transform code.
12804 * ipa-inline-transform.c (speculation_removed): New global var.
12805 (clone_inlined_nodes): See if speculation can be removed.
12806 (inline_call): If speculations was removed, we growths may not match.
12807 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
12808 (speculation_useful_p): New function.
12809 (resolve_noninline_speculation): New function.
12810 (inline_small_functions): Resolve useless speculations.
12811 * ipa-inline.h (speculation_useful_p): Declare
12812 * ipa.c (can_replace_by_local_alias): Simplify.
12813 (ipa_profile): Produce speculative calls in non-lto, too;
12814 add simple cost model; produce local aliases.
12816 2013-08-13 David Malcolm <dmalcolm@redhat.com>
12818 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
12821 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
12823 * config/i386/i386.c (ix86_function_versions): Use error + inform.
12825 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
12827 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
12828 iterator instead of X87MODEF.
12830 2013-08-12 Perez Read <netfirewall@gmail.com>
12833 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
12834 operand 0 for intel asm alternative.
12835 (*movabs<mode>_2): Ditto for operand 1.
12837 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
12839 * config/aarch64/arm_none.h
12840 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
12842 2013-08-12 Nick Clifton <nickc@redhat.com>
12844 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
12846 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12848 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
12849 expand for QI/HImode operand to produce more effictive code for
12850 unsigned char(short) --> float(double) conversion.
12852 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
12854 * doc/invoke.texi: Mention that -ftls-model does not force the final
12857 2013-08-12 Marek Polacek <polacek@redhat.com>
12858 Marc Glisse <marc.glisse@inria.fr>
12860 PR tree-optimization/57980
12861 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12862 when creating -1 constant.
12864 2013-08-10 Jan Hubicka <jh@suse.cz>
12866 Workaround binutils PR14342.
12867 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12868 (gimple_init_edge_profiler): Likewise.
12869 (gimple_gen_ic_func_profiler): Likewise.
12871 2013-08-09 Jan Hubicka <jh@suse.cz>
12873 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12875 2013-08-09 Xinliang David Li <davidxl@google.com>
12877 * config/i386/stringop.def: New file.
12878 * config/i386/stringop.opt: New file.
12879 * config/i386/i386-opts.h: Include stringopt.def.
12880 * config/i386/i386.opt: Include stringopt.opt.
12881 * config/i386/i386.c (ix86_option_override_internal):
12882 Override default size based stringop inline strategies with options.
12883 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12886 2013-08-09 Jan Hubicka <jh@suse.cz>
12888 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12890 2013-08-09 Jan Hubicka <jh@suse.cz>
12892 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12894 (dump_cgraph_node): Dump profile-id.
12895 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12896 and common_target_probability.
12897 * lto-cgraph.c (lto_output_edge): Stream common targets.
12898 (lto_output_node): Stream profile ids.
12899 (input_node): Stream profile ids.
12900 (input_edge): Stream common targets.
12901 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12902 * ipa.c: Include value-prof.h
12903 (ipa_profile_generate_summary): Turn indirect call statement histograms
12904 into common targets.
12905 (ipa_profile): Turn common targets into speculative edges.
12907 2013-08-09 Jan Hubicka <jh@suse.cz>
12909 * cgraph.h (cgraph_node): Add profile_id.
12910 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12911 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12912 (del_node_map): Update.
12913 (find_func_by_funcdef_no): Replace by ...
12914 (find_func_by_profile_id): ... this one.
12915 (gimple_ic_transform): Do not remove useful histograms when
12916 speculation is not done; dump info when indirect call removal
12918 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12919 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12920 (__gcov_indirect_call_profiler_v2): .. this one.
12921 * profile.h (init_node_map): Update.
12922 * coverage.c (coverage_compute_profile_id): New function.
12923 * coverage.h (coverage_compute_profile_id): Declare.
12924 * tree-profile.c (init_ic_make_global_vars): Make
12925 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12926 (gimple_init_edge_profiler): Update prototype of
12927 __gcov_indirect_call_profiler.
12928 (gimple_gen_ic_func_profiler): Simplify.
12929 (tree_profiling): Use init_node_map
12931 2013-08-09 Jan Hubicka <jh@suse.cz>
12933 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12934 non-speculative refs.
12935 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12936 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12937 (cgraph_set_call_stmt): Likewise.
12938 (cgraph_create_edge_1): Fix release checking compilatoin;
12939 clear lto_stmt_uid.
12940 (cgraph_free_edge): Free indirect info.
12941 (cgraph_turn_edge_to_speculative): New function.
12942 (cgraph_speculative_call_info): New function.
12943 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12944 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12945 (dump_cgraph_node): Dump speculation.
12946 (verify_edge_count_and_frequency): Accept speculative edges.
12947 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12948 (verify_cgraph_node): Handle speculation.
12949 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12950 (cgraph_set_call_stmt): Update prototype.
12951 (cgraph_make_edge_direct): Update prototype.
12952 (cgraph_speculative_call_info): Declare.
12953 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12954 to change; update call of ipa_find_references.
12955 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12956 lto_stmt_uid and speculative flags.
12957 (ipa_dump_references): Dump speculation.
12958 (ipa_clone_references): Clone speculative flag.
12959 (ipa_clone_referring): Likewise.
12960 (ipa_clone_ref): New function.
12961 (ipa_find_reference): Look into lto_stmt_uids
12962 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12963 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12964 (ipa_find_reference): Update declaration.
12965 (ipa_clone_ref): Declare.
12966 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12967 stream speculative flag.
12968 (lto_output_ref): Stream statements uids and speculation.
12969 (input_ref): Likewise.
12970 (input_edge): Stream speuclation.
12971 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12972 (cgraph_set_call_stmt_including_clones): Handle speculation.
12973 * ipa-inline.c (heap_edge_removal_hook): New function.
12974 (inline_small_functions): Register it.
12975 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12976 also initialize refs.
12977 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12979 (try_make_edge_direct_simple_call): Likewise.
12980 (try_make_edge_direct_simple_call): Likewise.
12981 (update_indirect_edges_after_inlining): Likewise.
12982 (remove_described_reference): Look proper lto_stmt_uid.
12983 (propagate_controlled_uses): Likewise.
12984 (propagate_controlled_uses): Liekwise.
12985 * tree-inline.c (copy_bb): Copy speculative edges.
12986 (redirect_all_calls): New function.
12987 (copy_cfg_body): Do redirection after loop info is updated.
12988 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12990 2013-08-09 Jan Hubicka <jh@suse.cz>
12992 * lto-streamer-out.c (output_function): Renumber PHIs.
12993 * lto-streamer-in.c (input_function): Likewise.
12995 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12997 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12998 (get_lane_unsigned): Likewise.
12999 (dup_lane_scalar): Likewise.
13000 (get_lane): enable for VALL.
13001 * config/aarch64/aarch64-simd.md
13002 (aarch64_dup_lane_scalar<mode>): Remove.
13003 (aarch64_get_lane_signed<mode>): Likewise.
13004 (aarch64_get_lane_unsigned<mode>): Likewise.
13005 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
13006 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
13007 (aarch64_get_lane<mode>): Enable for all vector modes.
13008 (aarch64_get_lanedi): Remove misleading constraints.
13009 * config/aarch64/arm_neon.h
13010 (__aarch64_vget_lane_any): Define.
13011 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
13012 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
13013 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
13014 * config/aarch64/iterators.md (VDQQH): New.
13015 (VDQQHS): Likewise.
13016 (vwcore): Likewise.
13018 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
13020 * configure.ac: Add GAS check for LEON instructions on SPARC.
13021 * configure: Regenerate.
13022 * config.in: Likewise.
13023 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
13025 * config/sparc/sparc.opt (LEON, LEON3): New masks.
13026 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
13028 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
13029 (AS_LEON_FLAG): New macro.
13030 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
13031 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
13032 Deal with LEON and LEON3 for the memory model.
13033 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
13034 (atomic_compare_and_swap<mode>_1): Likewise.
13035 (*atomic_compare_and_swap<mode>_1): Likewise.
13037 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13039 * config/arm/neon.md (vcond): Fix floating-point vector
13040 comparisons against 0.
13042 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
13044 * lra-constraints.c (emit_spill_move): Remove assert.
13045 (process_alt_operands): Add more debugging
13046 output. Increase reject for spilling into memory. Decrease
13047 reject for reloading scratch.
13048 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
13050 2013-08-08 Steve Ellcey <sellcey@mips.com>
13052 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
13053 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
13054 micromips incompatible. Add nan2008.
13055 (MULTILIB_DIRNAMES): Add nan2008.
13056 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
13057 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
13058 and micromips incompatible. Add nan2008.
13059 (MULTILIB_DIRNAMES): Add nan2008.
13060 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
13062 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
13064 PR rtl-optimization/58079
13065 * combine.c (combine_simplify_rtx): Avoid using SUBST if
13066 simplify_comparison has widened a comparison with an integer.
13068 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13070 * config/arm/neon.md (movmisalign<mode>): Disable when we
13071 don't allow unaligned accesses.
13072 (*movmisalign<mode>_neon_store): Likewise.
13073 (*movmisalign<mode>_neon_load): Likewise.
13074 (*movmisalign<mode>_neon_store): Likewise.
13075 (*movmisalign<mode>_neon_load): Likewise.
13077 2013-08-08 Jan Hubicka <jh@suse.cz>
13079 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
13080 (make_pass_rebuild_cgraph_edges): Also clear references.
13081 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
13082 * ipa-inline-transform.c (inline_transform): Remove all references
13084 * cgraphunit.c (expand_function): Remove all references after
13086 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
13087 (ipa_find_reference): Rewrite to iterator.
13088 (remove_stmt_references): Likewise.
13089 (ipa_clear_stmts_in_references): New function.
13090 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
13091 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
13093 * ipa-split.c (split_function): Remove references in split function.
13095 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
13098 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
13099 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
13101 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
13104 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
13105 suggested mode for the assignment isn't compatible with the
13106 registers required.
13108 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
13111 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
13113 2013-08-07 Xinliang David Li <davidxl@google.com>
13115 * config/i386/i386.opt: New option -mtune-ctrl=.
13116 * config/i386/x86-tune.def: New file.
13117 * config/i386/i386.h: include x86-tune.def.
13118 * config/i386/i386.c (ix86_option_override_internal):
13119 Parsing -mtune-ctrl= option and set tune features.
13121 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
13124 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
13125 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
13128 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
13130 * rtl.h (update_alignments): Declare.
13131 * final.c (grow_label_align): New function extracted from...
13132 (shorten_branches): ...here. Call it.
13133 (update_alignments): New function.
13134 * reorg.c (sibling_labels): New variable.
13135 (get_label_before): Add SIBLING parameter. If it is non-zero, push
13136 the new label along with it onto the sibling_labels vector.
13137 (fill_simple_delay_slots): Adjust call to get_label_before.
13138 (fill_slots_from_thread): Likewise.
13139 (relax_delay_slots): Likewise.
13140 (make_return_insns): Likewise.
13141 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
13143 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
13145 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
13146 document its semantics.
13147 (diagnostic_report_diagnostic): Adjust accordingly.
13149 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13151 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
13152 (sparc_option_override): ...and port to new C++ pass API.
13153 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
13155 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
13157 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
13159 2013-08-06 Caroline Tice <cmtice@google.com>
13161 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
13162 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
13163 * tree-pass.h: Add pass_vtable_verify.
13164 * varasm.c (assemble_variable): Add code to properly set the comdat
13165 section and name for the .vtable_map_vars section.
13166 (assemble_vtyv_preinit_initializer): New function.
13167 (default_sectin_type_flags): Make sure .vtable_map_vars section has
13169 * output.h: Add function decl for assemble_vtv_preinit_initializer.
13170 * vtable-verify.c: New file.
13171 * vtable-verify.h: New file.
13172 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
13173 initialiation levels.
13174 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
13175 * passes.def: Insert pass_vtable_verify.
13176 * aclocal.m4: Reorder includes.
13177 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
13178 -fvtv-counts options.
13179 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
13180 as appropriate, if -fvtable-verify=... is used.
13181 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
13182 -fvtable-verify=... is used.
13183 * Makefile.in (OBJS): Add vtable-verify.o to list.
13184 (vtable-verify.o): Add new build rule.
13185 (GTFILES): Add vtable-verify.c to list.
13186 * common.opt (fvtable-verify=): New flag.
13187 (vtv_priority): Values for fvtable-verify= flag.
13188 (fvtv-counts): New flag.
13189 (fvtv-debug): New flag.
13190 * tree.h (save_vtable_map_decl): New extern function decl.
13192 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13194 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
13195 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
13196 (pass_data_rl78_devirt): ...new pass_data instance and...
13197 (make_pass_rl78_devirt): ...new function.
13198 (rl78_asm_file_start): Port pass registration to new C++ API.
13200 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13202 * coretypes.h (rtl_opt_pass): Add.
13203 (gcc::context): Add.
13204 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
13205 (epiphany_init): Port to new C++ pass API.
13206 (epiphany_optimize_mode_switching): Likewise.
13207 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
13208 (pass_manager::get_pass_mode_switching): New.
13209 (pass_manager::get_pass_peephole2): New.
13210 * mode-switching.c (pass_mode_switching): Add clone method.
13211 * recog.c (pass_peephole2): Add clone method.
13212 (pass_split_all_insns): Add clone method.
13214 2013-08-06 David Malcolm <dmalcolm@redhat.com>
13216 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
13217 (mips_option_override): ...here, porting to new C++ API for passes.
13219 2013-08-06 Jan Hubicka <jh@suse.cz>
13221 * cgraph.c (cgraph_get_body): New function based on lto.c
13223 * cgraph.h (cgraph_get_body): Declare.
13224 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
13226 * cgraphunit.c (expand_function): Get body prior expanding.
13227 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
13228 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
13230 * passes.c (do_per_function_toporder): Get body.
13231 * tree-inline.c (expand_call_inline): Get body prior inlining it.
13232 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
13234 2013-08-06 Martin Jambor <mjambor@suse.cz>
13237 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
13238 is not re-finalized. Rename second parameter to no_collect.
13240 2013-08-06 Martin Jambor <mjambor@suse.cz>
13242 PR middle-end/58041
13243 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
13244 MEM_REF has proper alignment information.
13246 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
13249 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
13251 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
13252 rtx (*) (rtx, ...) with insn_gen_fn.
13253 * genoutput.c (output_insn_data): Cast gen_? function pointers to
13254 insn_gen_fn::stored_funcptr. Add initializer braces.
13256 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13258 Rewrite how instances of passes are cloned to remove assumptions
13259 about their sizes (thus allowing pass subclasses to have
13260 additional data fields, albeit non-GC-managed ones at this point).
13262 * passes.c (make_pass_instance): Now that passes have clone
13263 methods, rewrite this function to eliminate XNEW and memcpy
13264 calls that used hardcoded sizes. Since this function no longer
13265 creates pass instances, rename it to...
13266 (add_pass_instance): ...this. Document the old way that passes were
13267 numbered and flagged, and rework this function to continue using it.
13268 (next_pass_1): Add an initial_pass argument for use by
13270 (position_pass): When adding multiple instances of a pass, use
13271 the pass's clone method, rather than relying on the XNEW/memcpy
13272 within the former make_pass_instance (now add_pass_instance).
13273 (pass_manager::pass_manager): When invoking next_pass_1, also supply
13274 the initial instance of the current pass within the pass manager.
13276 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13278 This is the automated part of the conversion of passes from C
13279 structs to C++ classes.
13281 Patch autogenerated by refactor_passes.py from
13282 https://github.com/davidmalcolm/gcc-refactoring-scripts
13283 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
13285 * asan.c (pass_asan): Convert from a global struct to a subclass of
13286 gimple_opt_pass along with...
13287 (pass_data_asan): ...new pass_data instance and...
13288 (make_pass_asan): ...new function.
13289 (pass_asan_O0): Convert from a global struct to a subclass of
13290 gimple_opt_pass along with...
13291 (pass_data_asan_O0): ...new pass_data instance and...
13292 (make_pass_asan_O0): ...new function.
13293 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
13294 subclass of rtl_opt_pass along with...
13295 (pass_data_inc_dec): ...new pass_data instance and...
13296 (make_pass_inc_dec): ...new function.
13297 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
13298 a subclass of rtl_opt_pass along with...
13299 (pass_data_reorder_blocks): ...new pass_data instance and...
13300 (make_pass_reorder_blocks): ...new function.
13301 (pass_duplicate_computed_gotos): Convert from a global struct to a
13302 subclass of rtl_opt_pass along with...
13303 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
13304 (make_pass_duplicate_computed_gotos): ...new function.
13305 (pass_partition_blocks): Convert from a global struct to a subclass of
13306 rtl_opt_pass along with...
13307 (pass_data_partition_blocks): ...new pass_data instance and...
13308 (make_pass_partition_blocks): ...new function.
13309 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
13310 struct to a subclass of rtl_opt_pass along with...
13311 (pass_data_branch_target_load_optimize1): ...new pass_data instance
13313 (make_pass_branch_target_load_optimize1): ...new function.
13314 (pass_branch_target_load_optimize2): Convert from a global struct to a
13315 subclass of rtl_opt_pass along with...
13316 (pass_data_branch_target_load_optimize2): ...new pass_data instance
13318 (make_pass_branch_target_load_optimize2): ...new function.
13319 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
13320 of rtl_opt_pass along with...
13321 (pass_data_jump): ...new pass_data instance and...
13322 (make_pass_jump): ...new function.
13323 (pass_jump2): Convert from a global struct to a subclass of
13324 rtl_opt_pass along with...
13325 (pass_data_jump2): ...new pass_data instance and...
13326 (make_pass_jump2): ...new function.
13327 * cfgexpand.c (pass_expand): Convert from a global struct to a
13328 subclass of rtl_opt_pass along with...
13329 (pass_data_expand): ...new pass_data instance and...
13330 (make_pass_expand): ...new function.
13331 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
13332 of rtl_opt_pass along with...
13333 (pass_data_free_cfg): ...new pass_data instance and...
13334 (make_pass_free_cfg): ...new function.
13335 (pass_into_cfg_layout_mode): Convert from a global struct to a
13336 subclass of rtl_opt_pass along with...
13337 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
13338 (make_pass_into_cfg_layout_mode): ...new function.
13339 (pass_outof_cfg_layout_mode): Convert from a global struct to a
13340 subclass of rtl_opt_pass along with...
13341 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
13342 (make_pass_outof_cfg_layout_mode): ...new function.
13343 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
13344 struct to a subclass of gimple_opt_pass along with...
13345 (pass_data_build_cgraph_edges): ...new pass_data instance and...
13346 (make_pass_build_cgraph_edges): ...new function.
13347 (pass_rebuild_cgraph_edges): Convert from a global struct to a
13348 subclass of gimple_opt_pass along with...
13349 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
13350 (make_pass_rebuild_cgraph_edges): ...new function.
13351 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
13352 subclass of gimple_opt_pass along with...
13353 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
13355 (make_pass_remove_cgraph_callee_edges): ...new function.
13356 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
13357 struct to a subclass of rtl_opt_pass along with...
13358 (pass_data_stack_adjustments): ...new pass_data instance and...
13359 (make_pass_stack_adjustments): ...new function.
13360 * combine.c (pass_combine): Convert from a global struct to a subclass
13361 of rtl_opt_pass along with...
13362 (pass_data_combine): ...new pass_data instance and...
13363 (make_pass_combine): ...new function.
13364 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
13365 global struct to a subclass of rtl_opt_pass along with...
13366 (pass_data_compare_elim_after_reload): ...new pass_data instance
13368 (make_pass_compare_elim_after_reload): ...new function.
13369 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
13370 of rtl_opt_pass along with...
13371 (pass_data_rtl_cprop): ...new pass_data instance and...
13372 (make_pass_rtl_cprop): ...new function.
13373 * cse.c (pass_cse): Convert from a global struct to a subclass of
13374 rtl_opt_pass along with...
13375 (pass_data_cse): ...new pass_data instance and...
13376 (make_pass_cse): ...new function.
13377 (pass_cse2): Convert from a global struct to a subclass of
13378 rtl_opt_pass along with...
13379 (pass_data_cse2): ...new pass_data instance and...
13380 (make_pass_cse2): ...new function.
13381 (pass_cse_after_global_opts): Convert from a global struct to a
13382 subclass of rtl_opt_pass along with...
13383 (pass_data_cse_after_global_opts): ...new pass_data instance and...
13384 (make_pass_cse_after_global_opts): ...new function.
13385 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
13386 of rtl_opt_pass along with...
13387 (pass_data_ud_rtl_dce): ...new pass_data instance and...
13388 (make_pass_ud_rtl_dce): ...new function.
13389 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
13390 rtl_opt_pass along with...
13391 (pass_data_fast_rtl_dce): ...new pass_data instance and...
13392 (make_pass_fast_rtl_dce): ...new function.
13393 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
13394 a subclass of rtl_opt_pass along with...
13395 (pass_data_df_initialize_opt): ...new pass_data instance and...
13396 (make_pass_df_initialize_opt): ...new function.
13397 (pass_df_initialize_no_opt): Convert from a global struct to a
13398 subclass of rtl_opt_pass along with...
13399 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
13400 (make_pass_df_initialize_no_opt): ...new function.
13401 (pass_df_finish): Convert from a global struct to a subclass of
13402 rtl_opt_pass along with...
13403 (pass_data_df_finish): ...new pass_data instance and...
13404 (make_pass_df_finish): ...new function.
13405 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
13406 rtl_opt_pass along with...
13407 (pass_data_rtl_dse1): ...new pass_data instance and...
13408 (make_pass_rtl_dse1): ...new function.
13409 (pass_rtl_dse2): Convert from a global struct to a subclass of
13410 rtl_opt_pass along with...
13411 (pass_data_rtl_dse2): ...new pass_data instance and...
13412 (make_pass_rtl_dse2): ...new function.
13413 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
13414 subclass of rtl_opt_pass along with...
13415 (pass_data_dwarf2_frame): ...new pass_data instance and...
13416 (make_pass_dwarf2_frame): ...new function.
13417 * except.c (pass_set_nothrow_function_flags): Convert from a global
13418 struct to a subclass of rtl_opt_pass along with...
13419 (pass_data_set_nothrow_function_flags): ...new pass_data instance
13421 (make_pass_set_nothrow_function_flags): ...new function.
13422 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
13423 subclass of rtl_opt_pass along with...
13424 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
13426 (make_pass_convert_to_eh_region_ranges): ...new function.
13427 * final.c (pass_compute_alignments): Convert from a global struct to a
13428 subclass of rtl_opt_pass along with...
13429 (pass_data_compute_alignments): ...new pass_data instance and...
13430 (make_pass_compute_alignments): ...new function.
13431 (pass_final): Convert from a global struct to a subclass of
13432 rtl_opt_pass along with...
13433 (pass_data_final): ...new pass_data instance and...
13434 (make_pass_final): ...new function.
13435 (pass_shorten_branches): Convert from a global struct to a subclass of
13436 rtl_opt_pass along with...
13437 (pass_data_shorten_branches): ...new pass_data instance and...
13438 (make_pass_shorten_branches): ...new function.
13439 (pass_clean_state): Convert from a global struct to a subclass of
13440 rtl_opt_pass along with...
13441 (pass_data_clean_state): ...new pass_data instance and...
13442 (make_pass_clean_state): ...new function.
13443 * function.c (pass_instantiate_virtual_regs): Convert from a global
13444 struct to a subclass of rtl_opt_pass along with...
13445 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
13446 (make_pass_instantiate_virtual_regs): ...new function.
13447 (pass_leaf_regs): Convert from a global struct to a subclass of
13448 rtl_opt_pass along with...
13449 (pass_data_leaf_regs): ...new pass_data instance and...
13450 (make_pass_leaf_regs): ...new function.
13451 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
13452 subclass of rtl_opt_pass along with...
13453 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
13455 (make_pass_thread_prologue_and_epilogue): ...new function.
13456 (pass_match_asm_constraints): Convert from a global struct to a
13457 subclass of rtl_opt_pass along with...
13458 (pass_data_match_asm_constraints): ...new pass_data instance and...
13459 (make_pass_match_asm_constraints): ...new function.
13460 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
13461 subclass of rtl_opt_pass along with...
13462 (pass_data_rtl_fwprop): ...new pass_data instance and...
13463 (make_pass_rtl_fwprop): ...new function.
13464 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
13465 rtl_opt_pass along with...
13466 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
13467 (make_pass_rtl_fwprop_addr): ...new function.
13468 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
13469 rtl_opt_pass along with...
13470 (pass_data_rtl_pre): ...new pass_data instance and...
13471 (make_pass_rtl_pre): ...new function.
13472 (pass_rtl_hoist): Convert from a global struct to a subclass of
13473 rtl_opt_pass along with...
13474 (pass_data_rtl_hoist): ...new pass_data instance and...
13475 (make_pass_rtl_hoist): ...new function.
13476 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
13477 subclass of gimple_opt_pass along with...
13478 (pass_data_lower_cf): ...new pass_data instance and...
13479 (make_pass_lower_cf): ...new function.
13480 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
13481 from a global struct to a subclass of gimple_opt_pass along with...
13482 (pass_data_strength_reduction): ...new pass_data instance and...
13483 (make_pass_strength_reduction): ...new function.
13484 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
13485 of rtl_opt_pass along with...
13486 (pass_data_rtl_ifcvt): ...new pass_data instance and...
13487 (make_pass_rtl_ifcvt): ...new function.
13488 (pass_if_after_combine): Convert from a global struct to a subclass of
13489 rtl_opt_pass along with...
13490 (pass_data_if_after_combine): ...new pass_data instance and...
13491 (make_pass_if_after_combine): ...new function.
13492 (pass_if_after_reload): Convert from a global struct to a subclass of
13493 rtl_opt_pass along with...
13494 (pass_data_if_after_reload): ...new pass_data instance and...
13495 (make_pass_if_after_reload): ...new function.
13496 * init-regs.c (pass_initialize_regs): Convert from a global struct to
13497 a subclass of rtl_opt_pass along with...
13498 (pass_data_initialize_regs): ...new pass_data instance and...
13499 (make_pass_initialize_regs): ...new function.
13500 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
13501 of ipa_opt_pass_d along with...
13502 (pass_data_ipa_cp): ...new pass_data instance and...
13503 (make_pass_ipa_cp): ...new function.
13504 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
13505 global struct to a subclass of gimple_opt_pass along with...
13506 (pass_data_inline_parameters): ...new pass_data instance and...
13507 (make_pass_inline_parameters): ...new function.
13508 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
13509 subclass of gimple_opt_pass along with...
13510 (pass_data_early_inline): ...new pass_data instance and...
13511 (make_pass_early_inline): ...new function.
13512 (pass_ipa_inline): Convert from a global struct to a subclass of
13513 ipa_opt_pass_d along with...
13514 (pass_data_ipa_inline): ...new pass_data instance and...
13515 (make_pass_ipa_inline): ...new function.
13516 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
13517 struct to a subclass of gimple_opt_pass along with...
13518 (pass_data_local_pure_const): ...new pass_data instance and...
13519 (make_pass_local_pure_const): ...new function.
13520 (pass_ipa_pure_const): Convert from a global struct to a subclass of
13521 ipa_opt_pass_d along with...
13522 (pass_data_ipa_pure_const): ...new pass_data instance and...
13523 (make_pass_ipa_pure_const): ...new function.
13524 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
13525 to a subclass of ipa_opt_pass_d along with...
13526 (pass_data_ipa_reference): ...new pass_data instance and...
13527 (make_pass_ipa_reference): ...new function.
13528 * ipa-split.c (pass_split_functions): Convert from a global struct to
13529 a subclass of gimple_opt_pass along with...
13530 (pass_data_split_functions): ...new pass_data instance and...
13531 (make_pass_split_functions): ...new function.
13532 (pass_feedback_split_functions): Convert from a global struct to a
13533 subclass of gimple_opt_pass along with...
13534 (pass_data_feedback_split_functions): ...new pass_data instance and...
13535 (make_pass_feedback_split_functions): ...new function.
13536 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
13537 global struct to a subclass of simple_ipa_opt_pass along with...
13538 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
13540 (make_pass_ipa_function_and_variable_visibility): ...new function.
13541 (pass_ipa_free_inline_summary): Convert from a global struct to a
13542 subclass of simple_ipa_opt_pass along with...
13543 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
13544 (make_pass_ipa_free_inline_summary): ...new function.
13545 (pass_ipa_whole_program_visibility): Convert from a global struct to a
13546 subclass of ipa_opt_pass_d along with...
13547 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
13549 (make_pass_ipa_whole_program_visibility): ...new function.
13550 (pass_ipa_profile): Convert from a global struct to a subclass of
13551 ipa_opt_pass_d along with...
13552 (pass_data_ipa_profile): ...new pass_data instance and...
13553 (make_pass_ipa_profile): ...new function.
13554 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
13555 ipa_opt_pass_d along with...
13556 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
13557 (make_pass_ipa_cdtor_merge): ...new function.
13558 * ira.c (pass_ira): Convert from a global struct to a subclass of
13559 rtl_opt_pass along with...
13560 (pass_data_ira): ...new pass_data instance and...
13561 (make_pass_ira): ...new function.
13562 (pass_reload): Convert from a global struct to a subclass of
13563 rtl_opt_pass along with...
13564 (pass_data_reload): ...new pass_data instance and...
13565 (make_pass_reload): ...new function.
13566 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
13567 subclass of rtl_opt_pass along with...
13568 (pass_data_cleanup_barriers): ...new pass_data instance and...
13569 (make_pass_cleanup_barriers): ...new function.
13570 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
13571 of rtl_opt_pass along with...
13572 (pass_data_loop2): ...new pass_data instance and...
13573 (make_pass_loop2): ...new function.
13574 (pass_rtl_loop_init): Convert from a global struct to a subclass of
13575 rtl_opt_pass along with...
13576 (pass_data_rtl_loop_init): ...new pass_data instance and...
13577 (make_pass_rtl_loop_init): ...new function.
13578 (pass_rtl_loop_done): Convert from a global struct to a subclass of
13579 rtl_opt_pass along with...
13580 (pass_data_rtl_loop_done): ...new pass_data instance and...
13581 (make_pass_rtl_loop_done): ...new function.
13582 (pass_rtl_move_loop_invariants): Convert from a global struct to a
13583 subclass of rtl_opt_pass along with...
13584 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
13585 (make_pass_rtl_move_loop_invariants): ...new function.
13586 (pass_rtl_unswitch): Convert from a global struct to a subclass of
13587 rtl_opt_pass along with...
13588 (pass_data_rtl_unswitch): ...new pass_data instance and...
13589 (make_pass_rtl_unswitch): ...new function.
13590 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
13591 subclass of rtl_opt_pass along with...
13592 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
13594 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13595 (pass_rtl_doloop): Convert from a global struct to a subclass of
13596 rtl_opt_pass along with...
13597 (pass_data_rtl_doloop): ...new pass_data instance and...
13598 (make_pass_rtl_doloop): ...new function.
13599 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
13600 a subclass of rtl_opt_pass along with...
13601 (pass_data_lower_subreg): ...new pass_data instance and...
13602 (make_pass_lower_subreg): ...new function.
13603 (pass_lower_subreg2): Convert from a global struct to a subclass of
13604 rtl_opt_pass along with...
13605 (pass_data_lower_subreg2): ...new pass_data instance and...
13606 (make_pass_lower_subreg2): ...new function.
13607 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
13608 struct to a subclass of ipa_opt_pass_d along with...
13609 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
13610 (make_pass_ipa_lto_gimple_out): ...new function.
13611 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
13612 of ipa_opt_pass_d along with...
13613 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
13614 (make_pass_ipa_lto_finish_out): ...new function.
13615 * mode-switching.c (pass_mode_switching): Convert from a global struct
13616 to a subclass of rtl_opt_pass along with...
13617 (pass_data_mode_switching): ...new pass_data instance and...
13618 (make_pass_mode_switching): ...new function.
13619 * modulo-sched.c (pass_sms): Convert from a global struct to a
13620 subclass of rtl_opt_pass along with...
13621 (pass_data_sms): ...new pass_data instance and...
13622 (make_pass_sms): ...new function.
13623 * omp-low.c (pass_expand_omp): Convert from a global struct to a
13624 subclass of gimple_opt_pass along with...
13625 (pass_data_expand_omp): ...new pass_data instance and...
13626 (make_pass_expand_omp): ...new function.
13627 (pass_lower_omp): Convert from a global struct to a subclass of
13628 gimple_opt_pass along with...
13629 (pass_data_lower_omp): ...new pass_data instance and...
13630 (make_pass_lower_omp): ...new function.
13631 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
13632 of gimple_opt_pass along with...
13633 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
13634 (make_pass_diagnose_omp_blocks): ...new function.
13635 * passes.c (pass_early_local_passes): Convert from a global struct to
13636 a subclass of simple_ipa_opt_pass along with...
13637 (pass_data_early_local_passes): ...new pass_data instance and...
13638 (make_pass_early_local_passes): ...new function.
13639 (pass_all_early_optimizations): Convert from a global struct to a
13640 subclass of gimple_opt_pass along with...
13641 (pass_data_all_early_optimizations): ...new pass_data instance and...
13642 (make_pass_all_early_optimizations): ...new function.
13643 (pass_all_optimizations): Convert from a global struct to a subclass
13644 of gimple_opt_pass along with...
13645 (pass_data_all_optimizations): ...new pass_data instance and...
13646 (make_pass_all_optimizations): ...new function.
13647 (pass_all_optimizations_g): Convert from a global struct to a subclass
13648 of gimple_opt_pass along with...
13649 (pass_data_all_optimizations_g): ...new pass_data instance and...
13650 (make_pass_all_optimizations_g): ...new function.
13651 (pass_rest_of_compilation): Convert from a global struct to a subclass
13652 of rtl_opt_pass along with...
13653 (pass_data_rest_of_compilation): ...new pass_data instance and...
13654 (make_pass_rest_of_compilation): ...new function.
13655 (pass_postreload): Convert from a global struct to a subclass of
13656 rtl_opt_pass along with...
13657 (pass_data_postreload): ...new pass_data instance and...
13658 (make_pass_postreload): ...new function.
13659 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
13660 subclass of rtl_opt_pass along with...
13661 (pass_data_gcse2): ...new pass_data instance and...
13662 (make_pass_gcse2): ...new function.
13663 * postreload.c (pass_postreload_cse): Convert from a global struct to
13664 a subclass of rtl_opt_pass along with...
13665 (pass_data_postreload_cse): ...new pass_data instance and...
13666 (make_pass_postreload_cse): ...new function.
13667 * predict.c (pass_profile): Convert from a global struct to a subclass
13668 of gimple_opt_pass along with...
13669 (pass_data_profile): ...new pass_data instance and...
13670 (make_pass_profile): ...new function.
13671 (pass_strip_predict_hints): Convert from a global struct to a subclass
13672 of gimple_opt_pass along with...
13673 (pass_data_strip_predict_hints): ...new pass_data instance and...
13674 (make_pass_strip_predict_hints): ...new function.
13675 * recog.c (pass_peephole2): Convert from a global struct to a subclass
13676 of rtl_opt_pass along with...
13677 (pass_data_peephole2): ...new pass_data instance and...
13678 (make_pass_peephole2): ...new function.
13679 (pass_split_all_insns): Convert from a global struct to a subclass of
13680 rtl_opt_pass along with...
13681 (pass_data_split_all_insns): ...new pass_data instance and...
13682 (make_pass_split_all_insns): ...new function.
13683 (pass_split_after_reload): Convert from a global struct to a subclass
13684 of rtl_opt_pass along with...
13685 (pass_data_split_after_reload): ...new pass_data instance and...
13686 (make_pass_split_after_reload): ...new function.
13687 (pass_split_before_regstack): Convert from a global struct to a
13688 subclass of rtl_opt_pass along with...
13689 (pass_data_split_before_regstack): ...new pass_data instance and...
13690 (make_pass_split_before_regstack): ...new function.
13691 (pass_split_before_sched2): Convert from a global struct to a subclass
13692 of rtl_opt_pass along with...
13693 (pass_data_split_before_sched2): ...new pass_data instance and...
13694 (make_pass_split_before_sched2): ...new function.
13695 (pass_split_for_shorten_branches): Convert from a global struct to a
13696 subclass of rtl_opt_pass along with...
13697 (pass_data_split_for_shorten_branches): ...new pass_data instance
13699 (make_pass_split_for_shorten_branches): ...new function.
13700 * ree.c (pass_ree): Convert from a global struct to a subclass of
13701 rtl_opt_pass along with...
13702 (pass_data_ree): ...new pass_data instance and...
13703 (make_pass_ree): ...new function.
13704 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
13705 subclass of rtl_opt_pass along with...
13706 (pass_data_stack_regs): ...new pass_data instance and...
13707 (make_pass_stack_regs): ...new function.
13708 (pass_stack_regs_run): Convert from a global struct to a subclass of
13709 rtl_opt_pass along with...
13710 (pass_data_stack_regs_run): ...new pass_data instance and...
13711 (make_pass_stack_regs_run): ...new function.
13712 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
13713 subclass of rtl_opt_pass along with...
13714 (pass_data_cprop_hardreg): ...new pass_data instance and...
13715 (make_pass_cprop_hardreg): ...new function.
13716 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
13717 subclass of rtl_opt_pass along with...
13718 (pass_data_reginfo_init): ...new pass_data instance and...
13719 (make_pass_reginfo_init): ...new function.
13720 * regmove.c (pass_regmove): Convert from a global struct to a subclass
13721 of rtl_opt_pass along with...
13722 (pass_data_regmove): ...new pass_data instance and...
13723 (make_pass_regmove): ...new function.
13724 * regrename.c (pass_regrename): Convert from a global struct to a
13725 subclass of rtl_opt_pass along with...
13726 (pass_data_regrename): ...new pass_data instance and...
13727 (make_pass_regrename): ...new function.
13728 * reorg.c (pass_delay_slots): Convert from a global struct to a
13729 subclass of rtl_opt_pass along with...
13730 (pass_data_delay_slots): ...new pass_data instance and...
13731 (make_pass_delay_slots): ...new function.
13732 (pass_machine_reorg): Convert from a global struct to a subclass of
13733 rtl_opt_pass along with...
13734 (pass_data_machine_reorg): ...new pass_data instance and...
13735 (make_pass_machine_reorg): ...new function.
13736 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
13737 of rtl_opt_pass along with...
13738 (pass_data_sched): ...new pass_data instance and...
13739 (make_pass_sched): ...new function.
13740 (pass_sched2): Convert from a global struct to a subclass of
13741 rtl_opt_pass along with...
13742 (pass_data_sched2): ...new pass_data instance and...
13743 (make_pass_sched2): ...new function.
13744 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
13745 to a subclass of rtl_opt_pass along with...
13746 (pass_data_stack_ptr_mod): ...new pass_data instance and...
13747 (make_pass_stack_ptr_mod): ...new function.
13748 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
13749 to a subclass of rtl_opt_pass along with...
13750 (pass_data_rtl_store_motion): ...new pass_data instance and...
13751 (make_pass_rtl_store_motion): ...new function.
13752 * tracer.c (pass_tracer): Convert from a global struct to a subclass
13753 of gimple_opt_pass along with...
13754 (pass_data_tracer): ...new pass_data instance and...
13755 (make_pass_tracer): ...new function.
13756 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
13757 to a subclass of gimple_opt_pass along with...
13758 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
13759 (make_pass_diagnose_tm_blocks): ...new function.
13760 (pass_lower_tm): Convert from a global struct to a subclass of
13761 gimple_opt_pass along with...
13762 (pass_data_lower_tm): ...new pass_data instance and...
13763 (make_pass_lower_tm): ...new function.
13764 (pass_tm_init): Convert from a global struct to a subclass of
13765 gimple_opt_pass along with...
13766 (pass_data_tm_init): ...new pass_data instance and...
13767 (make_pass_tm_init): ...new function.
13768 (pass_tm_mark): Convert from a global struct to a subclass of
13769 gimple_opt_pass along with...
13770 (pass_data_tm_mark): ...new pass_data instance and...
13771 (make_pass_tm_mark): ...new function.
13772 (pass_tm_edges): Convert from a global struct to a subclass of
13773 gimple_opt_pass along with...
13774 (pass_data_tm_edges): ...new pass_data instance and...
13775 (make_pass_tm_edges): ...new function.
13776 (pass_tm_memopt): Convert from a global struct to a subclass of
13777 gimple_opt_pass along with...
13778 (pass_data_tm_memopt): ...new pass_data instance and...
13779 (make_pass_tm_memopt): ...new function.
13780 (pass_ipa_tm): Convert from a global struct to a subclass of
13781 simple_ipa_opt_pass along with...
13782 (pass_data_ipa_tm): ...new pass_data instance and...
13783 (make_pass_ipa_tm): ...new function.
13784 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
13785 subclass of gimple_opt_pass along with...
13786 (pass_data_call_cdce): ...new pass_data instance and...
13787 (make_pass_call_cdce): ...new function.
13788 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
13789 subclass of gimple_opt_pass along with...
13790 (pass_data_build_cfg): ...new pass_data instance and...
13791 (make_pass_build_cfg): ...new function.
13792 (pass_split_crit_edges): Convert from a global struct to a subclass of
13793 gimple_opt_pass along with...
13794 (pass_data_split_crit_edges): ...new pass_data instance and...
13795 (make_pass_split_crit_edges): ...new function.
13796 (pass_warn_function_return): Convert from a global struct to a
13797 subclass of gimple_opt_pass along with...
13798 (pass_data_warn_function_return): ...new pass_data instance and...
13799 (make_pass_warn_function_return): ...new function.
13800 (pass_warn_function_noreturn): Convert from a global struct to a
13801 subclass of gimple_opt_pass along with...
13802 (pass_data_warn_function_noreturn): ...new pass_data instance and...
13803 (make_pass_warn_function_noreturn): ...new function.
13804 (pass_warn_unused_result): Convert from a global struct to a subclass
13805 of gimple_opt_pass along with...
13806 (pass_data_warn_unused_result): ...new pass_data instance and...
13807 (make_pass_warn_unused_result): ...new function.
13808 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
13809 a subclass of gimple_opt_pass along with...
13810 (pass_data_merge_phi): ...new pass_data instance and...
13811 (make_pass_merge_phi): ...new function.
13812 * tree-complex.c (pass_lower_complex): Convert from a global struct to
13813 a subclass of gimple_opt_pass along with...
13814 (pass_data_lower_complex): ...new pass_data instance and...
13815 (make_pass_lower_complex): ...new function.
13816 (pass_lower_complex_O0): Convert from a global struct to a subclass of
13817 gimple_opt_pass along with...
13818 (pass_data_lower_complex_O0): ...new pass_data instance and...
13819 (make_pass_lower_complex_O0): ...new function.
13820 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
13821 subclass of gimple_opt_pass along with...
13822 (pass_data_lower_eh): ...new pass_data instance and...
13823 (make_pass_lower_eh): ...new function.
13824 (pass_refactor_eh): Convert from a global struct to a subclass of
13825 gimple_opt_pass along with...
13826 (pass_data_refactor_eh): ...new pass_data instance and...
13827 (make_pass_refactor_eh): ...new function.
13828 (pass_lower_resx): Convert from a global struct to a subclass of
13829 gimple_opt_pass along with...
13830 (pass_data_lower_resx): ...new pass_data instance and...
13831 (make_pass_lower_resx): ...new function.
13832 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
13833 of gimple_opt_pass along with...
13834 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
13835 (make_pass_lower_eh_dispatch): ...new function.
13836 (pass_cleanup_eh): Convert from a global struct to a subclass of
13837 gimple_opt_pass along with...
13838 (pass_data_cleanup_eh): ...new pass_data instance and...
13839 (make_pass_cleanup_eh): ...new function.
13840 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
13841 to a subclass of simple_ipa_opt_pass along with...
13842 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
13843 (make_pass_ipa_lower_emutls): ...new function.
13844 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
13845 a subclass of gimple_opt_pass along with...
13846 (pass_data_if_conversion): ...new pass_data instance and...
13847 (make_pass_if_conversion): ...new function.
13848 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
13849 subclass of gimple_opt_pass along with...
13850 (pass_data_build_ssa): ...new pass_data instance and...
13851 (make_pass_build_ssa): ...new function.
13852 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
13853 global struct to a subclass of gimple_opt_pass along with...
13854 (pass_data_loop_distribution): ...new pass_data instance and...
13855 (make_pass_loop_distribution): ...new function.
13856 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13857 subclass of gimple_opt_pass along with...
13858 (pass_data_mudflap_1): ...new pass_data instance and...
13859 (make_pass_mudflap_1): ...new function.
13860 (pass_mudflap_2): Convert from a global struct to a subclass of
13861 gimple_opt_pass along with...
13862 (pass_data_mudflap_2): ...new pass_data instance and...
13863 (make_pass_mudflap_2): ...new function.
13864 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13865 subclass of gimple_opt_pass along with...
13866 (pass_data_mudflap_1): ...new pass_data instance and...
13867 (make_pass_mudflap_1): ...new function.
13868 (pass_mudflap_2): Convert from a global struct to a subclass of
13869 gimple_opt_pass along with...
13870 (pass_data_mudflap_2): ...new pass_data instance and...
13871 (make_pass_mudflap_2): ...new function.
13872 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13873 gimple_opt_pass along with...
13874 (pass_data_nrv): ...new pass_data instance and...
13875 (make_pass_nrv): ...new function.
13876 (pass_return_slot): Convert from a global struct to a subclass of
13877 gimple_opt_pass along with...
13878 (pass_data_return_slot): ...new pass_data instance and...
13879 (make_pass_return_slot): ...new function.
13880 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13881 to a subclass of gimple_opt_pass along with...
13882 (pass_data_object_sizes): ...new pass_data instance and...
13883 (make_pass_object_sizes): ...new function.
13884 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13885 global struct to a subclass of gimple_opt_pass along with...
13886 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13888 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13889 (pass_fixup_cfg): Convert from a global struct to a subclass of
13890 gimple_opt_pass along with...
13891 (pass_data_fixup_cfg): ...new pass_data instance and...
13892 (make_pass_fixup_cfg): ...new function.
13893 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13894 (make_pass_mudflap_1): ...new function.
13895 (pass_mudflap_2): Replace declaration with that of...
13896 (make_pass_mudflap_2): ...new function.
13897 (pass_asan): Replace declaration with that of...
13898 (make_pass_asan): ...new function.
13899 (pass_asan_O0): Replace declaration with that of...
13900 (make_pass_asan_O0): ...new function.
13901 (pass_tsan): Replace declaration with that of...
13902 (make_pass_tsan): ...new function.
13903 (pass_tsan_O0): Replace declaration with that of...
13904 (make_pass_tsan_O0): ...new function.
13905 (pass_lower_cf): Replace declaration with that of...
13906 (make_pass_lower_cf): ...new function.
13907 (pass_refactor_eh): Replace declaration with that of...
13908 (make_pass_refactor_eh): ...new function.
13909 (pass_lower_eh): Replace declaration with that of...
13910 (make_pass_lower_eh): ...new function.
13911 (pass_lower_eh_dispatch): Replace declaration with that of...
13912 (make_pass_lower_eh_dispatch): ...new function.
13913 (pass_lower_resx): Replace declaration with that of...
13914 (make_pass_lower_resx): ...new function.
13915 (pass_build_cfg): Replace declaration with that of...
13916 (make_pass_build_cfg): ...new function.
13917 (pass_early_tree_profile): Replace declaration with that of...
13918 (make_pass_early_tree_profile): ...new function.
13919 (pass_cleanup_eh): Replace declaration with that of...
13920 (make_pass_cleanup_eh): ...new function.
13921 (pass_sra): Replace declaration with that of...
13922 (make_pass_sra): ...new function.
13923 (pass_sra_early): Replace declaration with that of...
13924 (make_pass_sra_early): ...new function.
13925 (pass_early_ipa_sra): Replace declaration with that of...
13926 (make_pass_early_ipa_sra): ...new function.
13927 (pass_tail_recursion): Replace declaration with that of...
13928 (make_pass_tail_recursion): ...new function.
13929 (pass_tail_calls): Replace declaration with that of...
13930 (make_pass_tail_calls): ...new function.
13931 (pass_tree_loop): Replace declaration with that of...
13932 (make_pass_tree_loop): ...new function.
13933 (pass_tree_loop_init): Replace declaration with that of...
13934 (make_pass_tree_loop_init): ...new function.
13935 (pass_lim): Replace declaration with that of...
13936 (make_pass_lim): ...new function.
13937 (pass_tree_unswitch): Replace declaration with that of...
13938 (make_pass_tree_unswitch): ...new function.
13939 (pass_predcom): Replace declaration with that of...
13940 (make_pass_predcom): ...new function.
13941 (pass_iv_canon): Replace declaration with that of...
13942 (make_pass_iv_canon): ...new function.
13943 (pass_scev_cprop): Replace declaration with that of...
13944 (make_pass_scev_cprop): ...new function.
13945 (pass_empty_loop): Replace declaration with that of...
13946 (make_pass_empty_loop): ...new function.
13947 (pass_record_bounds): Replace declaration with that of...
13948 (make_pass_record_bounds): ...new function.
13949 (pass_graphite): Replace declaration with that of...
13950 (make_pass_graphite): ...new function.
13951 (pass_graphite_transforms): Replace declaration with that of...
13952 (make_pass_graphite_transforms): ...new function.
13953 (pass_if_conversion): Replace declaration with that of...
13954 (make_pass_if_conversion): ...new function.
13955 (pass_loop_distribution): Replace declaration with that of...
13956 (make_pass_loop_distribution): ...new function.
13957 (pass_vectorize): Replace declaration with that of...
13958 (make_pass_vectorize): ...new function.
13959 (pass_slp_vectorize): Replace declaration with that of...
13960 (make_pass_slp_vectorize): ...new function.
13961 (pass_complete_unroll): Replace declaration with that of...
13962 (make_pass_complete_unroll): ...new function.
13963 (pass_complete_unrolli): Replace declaration with that of...
13964 (make_pass_complete_unrolli): ...new function.
13965 (pass_parallelize_loops): Replace declaration with that of...
13966 (make_pass_parallelize_loops): ...new function.
13967 (pass_loop_prefetch): Replace declaration with that of...
13968 (make_pass_loop_prefetch): ...new function.
13969 (pass_iv_optimize): Replace declaration with that of...
13970 (make_pass_iv_optimize): ...new function.
13971 (pass_tree_loop_done): Replace declaration with that of...
13972 (make_pass_tree_loop_done): ...new function.
13973 (pass_ch): Replace declaration with that of...
13974 (make_pass_ch): ...new function.
13975 (pass_ccp): Replace declaration with that of...
13976 (make_pass_ccp): ...new function.
13977 (pass_phi_only_cprop): Replace declaration with that of...
13978 (make_pass_phi_only_cprop): ...new function.
13979 (pass_build_ssa): Replace declaration with that of...
13980 (make_pass_build_ssa): ...new function.
13981 (pass_build_alias): Replace declaration with that of...
13982 (make_pass_build_alias): ...new function.
13983 (pass_build_ealias): Replace declaration with that of...
13984 (make_pass_build_ealias): ...new function.
13985 (pass_dominator): Replace declaration with that of...
13986 (make_pass_dominator): ...new function.
13987 (pass_dce): Replace declaration with that of...
13988 (make_pass_dce): ...new function.
13989 (pass_dce_loop): Replace declaration with that of...
13990 (make_pass_dce_loop): ...new function.
13991 (pass_cd_dce): Replace declaration with that of...
13992 (make_pass_cd_dce): ...new function.
13993 (pass_call_cdce): Replace declaration with that of...
13994 (make_pass_call_cdce): ...new function.
13995 (pass_merge_phi): Replace declaration with that of...
13996 (make_pass_merge_phi): ...new function.
13997 (pass_split_crit_edges): Replace declaration with that of...
13998 (make_pass_split_crit_edges): ...new function.
13999 (pass_pre): Replace declaration with that of...
14000 (make_pass_pre): ...new function.
14001 (pass_profile): Replace declaration with that of...
14002 (make_pass_profile): ...new function.
14003 (pass_strip_predict_hints): Replace declaration with that of...
14004 (make_pass_strip_predict_hints): ...new function.
14005 (pass_lower_complex_O0): Replace declaration with that of...
14006 (make_pass_lower_complex_O0): ...new function.
14007 (pass_lower_complex): Replace declaration with that of...
14008 (make_pass_lower_complex): ...new function.
14009 (pass_lower_vector): Replace declaration with that of...
14010 (make_pass_lower_vector): ...new function.
14011 (pass_lower_vector_ssa): Replace declaration with that of...
14012 (make_pass_lower_vector_ssa): ...new function.
14013 (pass_lower_omp): Replace declaration with that of...
14014 (make_pass_lower_omp): ...new function.
14015 (pass_diagnose_omp_blocks): Replace declaration with that of...
14016 (make_pass_diagnose_omp_blocks): ...new function.
14017 (pass_expand_omp): Replace declaration with that of...
14018 (make_pass_expand_omp): ...new function.
14019 (pass_expand_omp_ssa): Replace declaration with that of...
14020 (make_pass_expand_omp_ssa): ...new function.
14021 (pass_object_sizes): Replace declaration with that of...
14022 (make_pass_object_sizes): ...new function.
14023 (pass_strlen): Replace declaration with that of...
14024 (make_pass_strlen): ...new function.
14025 (pass_fold_builtins): Replace declaration with that of...
14026 (make_pass_fold_builtins): ...new function.
14027 (pass_stdarg): Replace declaration with that of...
14028 (make_pass_stdarg): ...new function.
14029 (pass_early_warn_uninitialized): Replace declaration with that of...
14030 (make_pass_early_warn_uninitialized): ...new function.
14031 (pass_late_warn_uninitialized): Replace declaration with that of...
14032 (make_pass_late_warn_uninitialized): ...new function.
14033 (pass_cse_reciprocals): Replace declaration with that of...
14034 (make_pass_cse_reciprocals): ...new function.
14035 (pass_cse_sincos): Replace declaration with that of...
14036 (make_pass_cse_sincos): ...new function.
14037 (pass_optimize_bswap): Replace declaration with that of...
14038 (make_pass_optimize_bswap): ...new function.
14039 (pass_optimize_widening_mul): Replace declaration with that of...
14040 (make_pass_optimize_widening_mul): ...new function.
14041 (pass_warn_function_return): Replace declaration with that of...
14042 (make_pass_warn_function_return): ...new function.
14043 (pass_warn_function_noreturn): Replace declaration with that of...
14044 (make_pass_warn_function_noreturn): ...new function.
14045 (pass_cselim): Replace declaration with that of...
14046 (make_pass_cselim): ...new function.
14047 (pass_phiopt): Replace declaration with that of...
14048 (make_pass_phiopt): ...new function.
14049 (pass_forwprop): Replace declaration with that of...
14050 (make_pass_forwprop): ...new function.
14051 (pass_phiprop): Replace declaration with that of...
14052 (make_pass_phiprop): ...new function.
14053 (pass_tree_ifcombine): Replace declaration with that of...
14054 (make_pass_tree_ifcombine): ...new function.
14055 (pass_dse): Replace declaration with that of...
14056 (make_pass_dse): ...new function.
14057 (pass_nrv): Replace declaration with that of...
14058 (make_pass_nrv): ...new function.
14059 (pass_rename_ssa_copies): Replace declaration with that of...
14060 (make_pass_rename_ssa_copies): ...new function.
14061 (pass_sink_code): Replace declaration with that of...
14062 (make_pass_sink_code): ...new function.
14063 (pass_fre): Replace declaration with that of...
14064 (make_pass_fre): ...new function.
14065 (pass_check_data_deps): Replace declaration with that of...
14066 (make_pass_check_data_deps): ...new function.
14067 (pass_copy_prop): Replace declaration with that of...
14068 (make_pass_copy_prop): ...new function.
14069 (pass_vrp): Replace declaration with that of...
14070 (make_pass_vrp): ...new function.
14071 (pass_uncprop): Replace declaration with that of...
14072 (make_pass_uncprop): ...new function.
14073 (pass_return_slot): Replace declaration with that of...
14074 (make_pass_return_slot): ...new function.
14075 (pass_reassoc): Replace declaration with that of...
14076 (make_pass_reassoc): ...new function.
14077 (pass_rebuild_cgraph_edges): Replace declaration with that of...
14078 (make_pass_rebuild_cgraph_edges): ...new function.
14079 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
14080 (make_pass_remove_cgraph_callee_edges): ...new function.
14081 (pass_build_cgraph_edges): Replace declaration with that of...
14082 (make_pass_build_cgraph_edges): ...new function.
14083 (pass_local_pure_const): Replace declaration with that of...
14084 (make_pass_local_pure_const): ...new function.
14085 (pass_tracer): Replace declaration with that of...
14086 (make_pass_tracer): ...new function.
14087 (pass_warn_unused_result): Replace declaration with that of...
14088 (make_pass_warn_unused_result): ...new function.
14089 (pass_diagnose_tm_blocks): Replace declaration with that of...
14090 (make_pass_diagnose_tm_blocks): ...new function.
14091 (pass_lower_tm): Replace declaration with that of...
14092 (make_pass_lower_tm): ...new function.
14093 (pass_tm_init): Replace declaration with that of...
14094 (make_pass_tm_init): ...new function.
14095 (pass_tm_mark): Replace declaration with that of...
14096 (make_pass_tm_mark): ...new function.
14097 (pass_tm_memopt): Replace declaration with that of...
14098 (make_pass_tm_memopt): ...new function.
14099 (pass_tm_edges): Replace declaration with that of...
14100 (make_pass_tm_edges): ...new function.
14101 (pass_split_functions): Replace declaration with that of...
14102 (make_pass_split_functions): ...new function.
14103 (pass_feedback_split_functions): Replace declaration with that of...
14104 (make_pass_feedback_split_functions): ...new function.
14105 (pass_strength_reduction): Replace declaration with that of...
14106 (make_pass_strength_reduction): ...new function.
14107 (pass_ipa_lower_emutls): Replace declaration with that of...
14108 (make_pass_ipa_lower_emutls): ...new function.
14109 (pass_ipa_function_and_variable_visibility): Replace declaration with
14111 (make_pass_ipa_function_and_variable_visibility): ...new function.
14112 (pass_ipa_tree_profile): Replace declaration with that of...
14113 (make_pass_ipa_tree_profile): ...new function.
14114 (pass_early_local_passes): Replace declaration with that of...
14115 (make_pass_early_local_passes): ...new function.
14116 (pass_ipa_whole_program_visibility): Replace declaration with that
14118 (make_pass_ipa_whole_program_visibility): ...new function.
14119 (pass_ipa_lto_gimple_out): Replace declaration with that of...
14120 (make_pass_ipa_lto_gimple_out): ...new function.
14121 (pass_ipa_increase_alignment): Replace declaration with that of...
14122 (make_pass_ipa_increase_alignment): ...new function.
14123 (pass_ipa_inline): Replace declaration with that of...
14124 (make_pass_ipa_inline): ...new function.
14125 (pass_ipa_free_lang_data): Replace declaration with that of...
14126 (make_pass_ipa_free_lang_data): ...new function.
14127 (pass_ipa_free_inline_summary): Replace declaration with that of...
14128 (make_pass_ipa_free_inline_summary): ...new function.
14129 (pass_ipa_cp): Replace declaration with that of...
14130 (make_pass_ipa_cp): ...new function.
14131 (pass_ipa_reference): Replace declaration with that of...
14132 (make_pass_ipa_reference): ...new function.
14133 (pass_ipa_pure_const): Replace declaration with that of...
14134 (make_pass_ipa_pure_const): ...new function.
14135 (pass_ipa_pta): Replace declaration with that of...
14136 (make_pass_ipa_pta): ...new function.
14137 (pass_ipa_lto_finish_out): Replace declaration with that of...
14138 (make_pass_ipa_lto_finish_out): ...new function.
14139 (pass_ipa_tm): Replace declaration with that of...
14140 (make_pass_ipa_tm): ...new function.
14141 (pass_ipa_profile): Replace declaration with that of...
14142 (make_pass_ipa_profile): ...new function.
14143 (pass_ipa_cdtor_merge): Replace declaration with that of...
14144 (make_pass_ipa_cdtor_merge): ...new function.
14145 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
14147 (make_pass_cleanup_cfg_post_optimizing): ...new function.
14148 (pass_init_datastructures): Replace declaration with that of...
14149 (make_pass_init_datastructures): ...new function.
14150 (pass_fixup_cfg): Replace declaration with that of...
14151 (make_pass_fixup_cfg): ...new function.
14152 (pass_expand): Replace declaration with that of...
14153 (make_pass_expand): ...new function.
14154 (pass_instantiate_virtual_regs): Replace declaration with that of...
14155 (make_pass_instantiate_virtual_regs): ...new function.
14156 (pass_rtl_fwprop): Replace declaration with that of...
14157 (make_pass_rtl_fwprop): ...new function.
14158 (pass_rtl_fwprop_addr): Replace declaration with that of...
14159 (make_pass_rtl_fwprop_addr): ...new function.
14160 (pass_jump): Replace declaration with that of...
14161 (make_pass_jump): ...new function.
14162 (pass_jump2): Replace declaration with that of...
14163 (make_pass_jump2): ...new function.
14164 (pass_lower_subreg): Replace declaration with that of...
14165 (make_pass_lower_subreg): ...new function.
14166 (pass_cse): Replace declaration with that of...
14167 (make_pass_cse): ...new function.
14168 (pass_fast_rtl_dce): Replace declaration with that of...
14169 (make_pass_fast_rtl_dce): ...new function.
14170 (pass_ud_rtl_dce): Replace declaration with that of...
14171 (make_pass_ud_rtl_dce): ...new function.
14172 (pass_rtl_dce): Replace declaration with that of...
14173 (make_pass_rtl_dce): ...new function.
14174 (pass_rtl_dse1): Replace declaration with that of...
14175 (make_pass_rtl_dse1): ...new function.
14176 (pass_rtl_dse2): Replace declaration with that of...
14177 (make_pass_rtl_dse2): ...new function.
14178 (pass_rtl_dse3): Replace declaration with that of...
14179 (make_pass_rtl_dse3): ...new function.
14180 (pass_rtl_cprop): Replace declaration with that of...
14181 (make_pass_rtl_cprop): ...new function.
14182 (pass_rtl_pre): Replace declaration with that of...
14183 (make_pass_rtl_pre): ...new function.
14184 (pass_rtl_hoist): Replace declaration with that of...
14185 (make_pass_rtl_hoist): ...new function.
14186 (pass_rtl_store_motion): Replace declaration with that of...
14187 (make_pass_rtl_store_motion): ...new function.
14188 (pass_cse_after_global_opts): Replace declaration with that of...
14189 (make_pass_cse_after_global_opts): ...new function.
14190 (pass_rtl_ifcvt): Replace declaration with that of...
14191 (make_pass_rtl_ifcvt): ...new function.
14192 (pass_into_cfg_layout_mode): Replace declaration with that of...
14193 (make_pass_into_cfg_layout_mode): ...new function.
14194 (pass_outof_cfg_layout_mode): Replace declaration with that of...
14195 (make_pass_outof_cfg_layout_mode): ...new function.
14196 (pass_loop2): Replace declaration with that of...
14197 (make_pass_loop2): ...new function.
14198 (pass_rtl_loop_init): Replace declaration with that of...
14199 (make_pass_rtl_loop_init): ...new function.
14200 (pass_rtl_move_loop_invariants): Replace declaration with that of...
14201 (make_pass_rtl_move_loop_invariants): ...new function.
14202 (pass_rtl_unswitch): Replace declaration with that of...
14203 (make_pass_rtl_unswitch): ...new function.
14204 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
14205 (make_pass_rtl_unroll_and_peel_loops): ...new function.
14206 (pass_rtl_doloop): Replace declaration with that of...
14207 (make_pass_rtl_doloop): ...new function.
14208 (pass_rtl_loop_done): Replace declaration with that of...
14209 (make_pass_rtl_loop_done): ...new function.
14210 (pass_web): Replace declaration with that of...
14211 (make_pass_web): ...new function.
14212 (pass_cse2): Replace declaration with that of...
14213 (make_pass_cse2): ...new function.
14214 (pass_df_initialize_opt): Replace declaration with that of...
14215 (make_pass_df_initialize_opt): ...new function.
14216 (pass_df_initialize_no_opt): Replace declaration with that of...
14217 (make_pass_df_initialize_no_opt): ...new function.
14218 (pass_reginfo_init): Replace declaration with that of...
14219 (make_pass_reginfo_init): ...new function.
14220 (pass_inc_dec): Replace declaration with that of...
14221 (make_pass_inc_dec): ...new function.
14222 (pass_stack_ptr_mod): Replace declaration with that of...
14223 (make_pass_stack_ptr_mod): ...new function.
14224 (pass_initialize_regs): Replace declaration with that of...
14225 (make_pass_initialize_regs): ...new function.
14226 (pass_combine): Replace declaration with that of...
14227 (make_pass_combine): ...new function.
14228 (pass_if_after_combine): Replace declaration with that of...
14229 (make_pass_if_after_combine): ...new function.
14230 (pass_ree): Replace declaration with that of...
14231 (make_pass_ree): ...new function.
14232 (pass_partition_blocks): Replace declaration with that of...
14233 (make_pass_partition_blocks): ...new function.
14234 (pass_match_asm_constraints): Replace declaration with that of...
14235 (make_pass_match_asm_constraints): ...new function.
14236 (pass_regmove): Replace declaration with that of...
14237 (make_pass_regmove): ...new function.
14238 (pass_split_all_insns): Replace declaration with that of...
14239 (make_pass_split_all_insns): ...new function.
14240 (pass_fast_rtl_byte_dce): Replace declaration with that of...
14241 (make_pass_fast_rtl_byte_dce): ...new function.
14242 (pass_lower_subreg2): Replace declaration with that of...
14243 (make_pass_lower_subreg2): ...new function.
14244 (pass_mode_switching): Replace declaration with that of...
14245 (make_pass_mode_switching): ...new function.
14246 (pass_sms): Replace declaration with that of...
14247 (make_pass_sms): ...new function.
14248 (pass_sched): Replace declaration with that of...
14249 (make_pass_sched): ...new function.
14250 (pass_ira): Replace declaration with that of...
14251 (make_pass_ira): ...new function.
14252 (pass_reload): Replace declaration with that of...
14253 (make_pass_reload): ...new function.
14254 (pass_clean_state): Replace declaration with that of...
14255 (make_pass_clean_state): ...new function.
14256 (pass_branch_prob): Replace declaration with that of...
14257 (make_pass_branch_prob): ...new function.
14258 (pass_value_profile_transformations): Replace declaration with that
14260 (make_pass_value_profile_transformations): ...new function.
14261 (pass_postreload_cse): Replace declaration with that of...
14262 (make_pass_postreload_cse): ...new function.
14263 (pass_gcse2): Replace declaration with that of...
14264 (make_pass_gcse2): ...new function.
14265 (pass_split_after_reload): Replace declaration with that of...
14266 (make_pass_split_after_reload): ...new function.
14267 (pass_branch_target_load_optimize1): Replace declaration with that
14269 (make_pass_branch_target_load_optimize1): ...new function.
14270 (pass_thread_prologue_and_epilogue): Replace declaration with that
14272 (make_pass_thread_prologue_and_epilogue): ...new function.
14273 (pass_stack_adjustments): Replace declaration with that of...
14274 (make_pass_stack_adjustments): ...new function.
14275 (pass_peephole2): Replace declaration with that of...
14276 (make_pass_peephole2): ...new function.
14277 (pass_if_after_reload): Replace declaration with that of...
14278 (make_pass_if_after_reload): ...new function.
14279 (pass_regrename): Replace declaration with that of...
14280 (make_pass_regrename): ...new function.
14281 (pass_cprop_hardreg): Replace declaration with that of...
14282 (make_pass_cprop_hardreg): ...new function.
14283 (pass_reorder_blocks): Replace declaration with that of...
14284 (make_pass_reorder_blocks): ...new function.
14285 (pass_branch_target_load_optimize2): Replace declaration with that
14287 (make_pass_branch_target_load_optimize2): ...new function.
14288 (pass_leaf_regs): Replace declaration with that of...
14289 (make_pass_leaf_regs): ...new function.
14290 (pass_split_before_sched2): Replace declaration with that of...
14291 (make_pass_split_before_sched2): ...new function.
14292 (pass_compare_elim_after_reload): Replace declaration with that of...
14293 (make_pass_compare_elim_after_reload): ...new function.
14294 (pass_sched2): Replace declaration with that of...
14295 (make_pass_sched2): ...new function.
14296 (pass_stack_regs): Replace declaration with that of...
14297 (make_pass_stack_regs): ...new function.
14298 (pass_stack_regs_run): Replace declaration with that of...
14299 (make_pass_stack_regs_run): ...new function.
14300 (pass_df_finish): Replace declaration with that of...
14301 (make_pass_df_finish): ...new function.
14302 (pass_compute_alignments): Replace declaration with that of...
14303 (make_pass_compute_alignments): ...new function.
14304 (pass_duplicate_computed_gotos): Replace declaration with that of...
14305 (make_pass_duplicate_computed_gotos): ...new function.
14306 (pass_variable_tracking): Replace declaration with that of...
14307 (make_pass_variable_tracking): ...new function.
14308 (pass_free_cfg): Replace declaration with that of...
14309 (make_pass_free_cfg): ...new function.
14310 (pass_machine_reorg): Replace declaration with that of...
14311 (make_pass_machine_reorg): ...new function.
14312 (pass_cleanup_barriers): Replace declaration with that of...
14313 (make_pass_cleanup_barriers): ...new function.
14314 (pass_delay_slots): Replace declaration with that of...
14315 (make_pass_delay_slots): ...new function.
14316 (pass_split_for_shorten_branches): Replace declaration with that of...
14317 (make_pass_split_for_shorten_branches): ...new function.
14318 (pass_split_before_regstack): Replace declaration with that of...
14319 (make_pass_split_before_regstack): ...new function.
14320 (pass_convert_to_eh_region_ranges): Replace declaration with that
14322 (make_pass_convert_to_eh_region_ranges): ...new function.
14323 (pass_shorten_branches): Replace declaration with that of...
14324 (make_pass_shorten_branches): ...new function.
14325 (pass_set_nothrow_function_flags): Replace declaration with that of...
14326 (make_pass_set_nothrow_function_flags): ...new function.
14327 (pass_dwarf2_frame): Replace declaration with that of...
14328 (make_pass_dwarf2_frame): ...new function.
14329 (pass_final): Replace declaration with that of...
14330 (make_pass_final): ...new function.
14331 (pass_rtl_seqabstr): Replace declaration with that of...
14332 (make_pass_rtl_seqabstr): ...new function.
14333 (pass_release_ssa_names): Replace declaration with that of...
14334 (make_pass_release_ssa_names): ...new function.
14335 (pass_early_inline): Replace declaration with that of...
14336 (make_pass_early_inline): ...new function.
14337 (pass_inline_parameters): Replace declaration with that of...
14338 (make_pass_inline_parameters): ...new function.
14339 (pass_update_address_taken): Replace declaration with that of...
14340 (make_pass_update_address_taken): ...new function.
14341 (pass_convert_switch): Replace declaration with that of...
14342 (make_pass_convert_switch): ...new function.
14343 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
14344 to a subclass of simple_ipa_opt_pass along with...
14345 (pass_data_ipa_tree_profile): ...new pass_data instance and...
14346 (make_pass_ipa_tree_profile): ...new function.
14347 * tree-sra.c (pass_sra_early): Convert from a global struct to a
14348 subclass of gimple_opt_pass along with...
14349 (pass_data_sra_early): ...new pass_data instance and...
14350 (make_pass_sra_early): ...new function.
14351 (pass_sra): Convert from a global struct to a subclass of
14352 gimple_opt_pass along with...
14353 (pass_data_sra): ...new pass_data instance and...
14354 (make_pass_sra): ...new function.
14355 (pass_early_ipa_sra): Convert from a global struct to a subclass of
14356 gimple_opt_pass along with...
14357 (pass_data_early_ipa_sra): ...new pass_data instance and...
14358 (make_pass_early_ipa_sra): ...new function.
14359 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
14360 subclass of gimple_opt_pass along with...
14361 (pass_data_ccp): ...new pass_data instance and...
14362 (make_pass_ccp): ...new function.
14363 (pass_fold_builtins): Convert from a global struct to a subclass of
14364 gimple_opt_pass along with...
14365 (pass_data_fold_builtins): ...new pass_data instance and...
14366 (make_pass_fold_builtins): ...new function.
14367 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
14368 subclass of gimple_opt_pass along with...
14369 (pass_data_copy_prop): ...new pass_data instance and...
14370 (make_pass_copy_prop): ...new function.
14371 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
14372 global struct to a subclass of gimple_opt_pass along with...
14373 (pass_data_rename_ssa_copies): ...new pass_data instance and...
14374 (make_pass_rename_ssa_copies): ...new function.
14375 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
14376 subclass of gimple_opt_pass along with...
14377 (pass_data_dce): ...new pass_data instance and...
14378 (make_pass_dce): ...new function.
14379 (pass_dce_loop): Convert from a global struct to a subclass of
14380 gimple_opt_pass along with...
14381 (pass_data_dce_loop): ...new pass_data instance and...
14382 (make_pass_dce_loop): ...new function.
14383 (pass_cd_dce): Convert from a global struct to a subclass of
14384 gimple_opt_pass along with...
14385 (pass_data_cd_dce): ...new pass_data instance and...
14386 (make_pass_cd_dce): ...new function.
14387 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
14388 subclass of gimple_opt_pass along with...
14389 (pass_data_dominator): ...new pass_data instance and...
14390 (make_pass_dominator): ...new function.
14391 (pass_phi_only_cprop): Convert from a global struct to a subclass of
14392 gimple_opt_pass along with...
14393 (pass_data_phi_only_cprop): ...new pass_data instance and...
14394 (make_pass_phi_only_cprop): ...new function.
14395 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
14396 subclass of gimple_opt_pass along with...
14397 (pass_data_dse): ...new pass_data instance and...
14398 (make_pass_dse): ...new function.
14399 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
14400 a subclass of gimple_opt_pass along with...
14401 (pass_data_forwprop): ...new pass_data instance and...
14402 (make_pass_forwprop): ...new function.
14403 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
14404 struct to a subclass of gimple_opt_pass along with...
14405 (pass_data_tree_ifcombine): ...new pass_data instance and...
14406 (make_pass_tree_ifcombine): ...new function.
14407 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
14408 subclass of gimple_opt_pass along with...
14409 (pass_data_ch): ...new pass_data instance and...
14410 (make_pass_ch): ...new function.
14411 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
14412 subclass of gimple_opt_pass along with...
14413 (pass_data_tree_loop): ...new pass_data instance and...
14414 (make_pass_tree_loop): ...new function.
14415 (pass_tree_loop_init): Convert from a global struct to a subclass of
14416 gimple_opt_pass along with...
14417 (pass_data_tree_loop_init): ...new pass_data instance and...
14418 (make_pass_tree_loop_init): ...new function.
14419 (pass_lim): Convert from a global struct to a subclass of
14420 gimple_opt_pass along with...
14421 (pass_data_lim): ...new pass_data instance and...
14422 (make_pass_lim): ...new function.
14423 (pass_tree_unswitch): Convert from a global struct to a subclass of
14424 gimple_opt_pass along with...
14425 (pass_data_tree_unswitch): ...new pass_data instance and...
14426 (make_pass_tree_unswitch): ...new function.
14427 (pass_predcom): Convert from a global struct to a subclass of
14428 gimple_opt_pass along with...
14429 (pass_data_predcom): ...new pass_data instance and...
14430 (make_pass_predcom): ...new function.
14431 (pass_vectorize): Convert from a global struct to a subclass of
14432 gimple_opt_pass along with...
14433 (pass_data_vectorize): ...new pass_data instance and...
14434 (make_pass_vectorize): ...new function.
14435 (pass_graphite): Convert from a global struct to a subclass of
14436 gimple_opt_pass along with...
14437 (pass_data_graphite): ...new pass_data instance and...
14438 (make_pass_graphite): ...new function.
14439 (pass_graphite_transforms): Convert from a global struct to a subclass
14440 of gimple_opt_pass along with...
14441 (pass_data_graphite_transforms): ...new pass_data instance and...
14442 (make_pass_graphite_transforms): ...new function.
14443 (pass_check_data_deps): Convert from a global struct to a subclass of
14444 gimple_opt_pass along with...
14445 (pass_data_check_data_deps): ...new pass_data instance and...
14446 (make_pass_check_data_deps): ...new function.
14447 (pass_iv_canon): Convert from a global struct to a subclass of
14448 gimple_opt_pass along with...
14449 (pass_data_iv_canon): ...new pass_data instance and...
14450 (make_pass_iv_canon): ...new function.
14451 (pass_scev_cprop): Convert from a global struct to a subclass of
14452 gimple_opt_pass along with...
14453 (pass_data_scev_cprop): ...new pass_data instance and...
14454 (make_pass_scev_cprop): ...new function.
14455 (pass_record_bounds): Convert from a global struct to a subclass of
14456 gimple_opt_pass along with...
14457 (pass_data_record_bounds): ...new pass_data instance and...
14458 (make_pass_record_bounds): ...new function.
14459 (pass_complete_unroll): Convert from a global struct to a subclass of
14460 gimple_opt_pass along with...
14461 (pass_data_complete_unroll): ...new pass_data instance and...
14462 (make_pass_complete_unroll): ...new function.
14463 (pass_complete_unrolli): Convert from a global struct to a subclass of
14464 gimple_opt_pass along with...
14465 (pass_data_complete_unrolli): ...new pass_data instance and...
14466 (make_pass_complete_unrolli): ...new function.
14467 (pass_parallelize_loops): Convert from a global struct to a subclass
14468 of gimple_opt_pass along with...
14469 (pass_data_parallelize_loops): ...new pass_data instance and...
14470 (make_pass_parallelize_loops): ...new function.
14471 (pass_loop_prefetch): Convert from a global struct to a subclass of
14472 gimple_opt_pass along with...
14473 (pass_data_loop_prefetch): ...new pass_data instance and...
14474 (make_pass_loop_prefetch): ...new function.
14475 (pass_iv_optimize): Convert from a global struct to a subclass of
14476 gimple_opt_pass along with...
14477 (pass_data_iv_optimize): ...new pass_data instance and...
14478 (make_pass_iv_optimize): ...new function.
14479 (pass_tree_loop_done): Convert from a global struct to a subclass of
14480 gimple_opt_pass along with...
14481 (pass_data_tree_loop_done): ...new pass_data instance and...
14482 (make_pass_tree_loop_done): ...new function.
14483 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
14484 struct to a subclass of gimple_opt_pass along with...
14485 (pass_data_cse_reciprocals): ...new pass_data instance and...
14486 (make_pass_cse_reciprocals): ...new function.
14487 (pass_cse_sincos): Convert from a global struct to a subclass of
14488 gimple_opt_pass along with...
14489 (pass_data_cse_sincos): ...new pass_data instance and...
14490 (make_pass_cse_sincos): ...new function.
14491 (pass_optimize_bswap): Convert from a global struct to a subclass of
14492 gimple_opt_pass along with...
14493 (pass_data_optimize_bswap): ...new pass_data instance and...
14494 (make_pass_optimize_bswap): ...new function.
14495 (pass_optimize_widening_mul): Convert from a global struct to a
14496 subclass of gimple_opt_pass along with...
14497 (pass_data_optimize_widening_mul): ...new pass_data instance and...
14498 (make_pass_optimize_widening_mul): ...new function.
14499 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
14500 subclass of gimple_opt_pass along with...
14501 (pass_data_phiopt): ...new pass_data instance and...
14502 (make_pass_phiopt): ...new function.
14503 (pass_cselim): Convert from a global struct to a subclass of
14504 gimple_opt_pass along with...
14505 (pass_data_cselim): ...new pass_data instance and...
14506 (make_pass_cselim): ...new function.
14507 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
14508 subclass of gimple_opt_pass along with...
14509 (pass_data_phiprop): ...new pass_data instance and...
14510 (make_pass_phiprop): ...new function.
14511 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
14512 subclass of gimple_opt_pass along with...
14513 (pass_data_pre): ...new pass_data instance and...
14514 (make_pass_pre): ...new function.
14515 (pass_fre): Convert from a global struct to a subclass of
14516 gimple_opt_pass along with...
14517 (pass_data_fre): ...new pass_data instance and...
14518 (make_pass_fre): ...new function.
14519 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
14520 subclass of gimple_opt_pass along with...
14521 (pass_data_reassoc): ...new pass_data instance and...
14522 (make_pass_reassoc): ...new function.
14523 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
14524 subclass of gimple_opt_pass along with...
14525 (pass_data_sink_code): ...new pass_data instance and...
14526 (make_pass_sink_code): ...new function.
14527 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
14528 subclass of gimple_opt_pass along with...
14529 (pass_data_strlen): ...new pass_data instance and...
14530 (make_pass_strlen): ...new function.
14531 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
14532 struct to a subclass of gimple_opt_pass along with...
14533 (pass_data_build_alias): ...new pass_data instance and...
14534 (make_pass_build_alias): ...new function.
14535 (pass_build_ealias): Convert from a global struct to a subclass of
14536 gimple_opt_pass along with...
14537 (pass_data_build_ealias): ...new pass_data instance and...
14538 (make_pass_build_ealias): ...new function.
14539 (pass_ipa_pta): Convert from a global struct to a subclass of
14540 simple_ipa_opt_pass along with...
14541 (pass_data_ipa_pta): ...new pass_data instance and...
14542 (make_pass_ipa_pta): ...new function.
14543 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
14544 subclass of gimple_opt_pass along with...
14545 (pass_data_uncprop): ...new pass_data instance and...
14546 (make_pass_uncprop): ...new function.
14547 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
14548 global struct to a subclass of gimple_opt_pass along with...
14549 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
14550 (make_pass_late_warn_uninitialized): ...new function.
14551 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
14552 to a subclass of gimple_opt_pass along with...
14553 (pass_data_init_datastructures): ...new pass_data instance and...
14554 (make_pass_init_datastructures): ...new function.
14555 (pass_early_warn_uninitialized): Convert from a global struct to a
14556 subclass of gimple_opt_pass along with...
14557 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
14558 (make_pass_early_warn_uninitialized): ...new function.
14559 (pass_update_address_taken): Convert from a global struct to a
14560 subclass of gimple_opt_pass along with...
14561 (pass_data_update_address_taken): ...new pass_data instance and...
14562 (make_pass_update_address_taken): ...new function.
14563 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
14564 struct to a subclass of gimple_opt_pass along with...
14565 (pass_data_release_ssa_names): ...new pass_data instance and...
14566 (make_pass_release_ssa_names): ...new function.
14567 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
14568 subclass of gimple_opt_pass along with...
14569 (pass_data_stdarg): ...new pass_data instance and...
14570 (make_pass_stdarg): ...new function.
14571 * tree-switch-conversion.c (pass_convert_switch): Convert from a
14572 global struct to a subclass of gimple_opt_pass along with...
14573 (pass_data_convert_switch): ...new pass_data instance and...
14574 (make_pass_convert_switch): ...new function.
14575 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
14576 to a subclass of gimple_opt_pass along with...
14577 (pass_data_tail_recursion): ...new pass_data instance and...
14578 (make_pass_tail_recursion): ...new function.
14579 (pass_tail_calls): Convert from a global struct to a subclass of
14580 gimple_opt_pass along with...
14581 (pass_data_tail_calls): ...new pass_data instance and...
14582 (make_pass_tail_calls): ...new function.
14583 * tree-vect-generic.c (pass_lower_vector): Convert from a global
14584 struct to a subclass of gimple_opt_pass along with...
14585 (pass_data_lower_vector): ...new pass_data instance and...
14586 (make_pass_lower_vector): ...new function.
14587 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
14588 gimple_opt_pass along with...
14589 (pass_data_lower_vector_ssa): ...new pass_data instance and...
14590 (make_pass_lower_vector_ssa): ...new function.
14591 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
14592 to a subclass of gimple_opt_pass along with...
14593 (pass_data_slp_vectorize): ...new pass_data instance and...
14594 (make_pass_slp_vectorize): ...new function.
14595 (pass_ipa_increase_alignment): Convert from a global struct to a
14596 subclass of simple_ipa_opt_pass along with...
14597 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
14598 (make_pass_ipa_increase_alignment): ...new function.
14599 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
14600 gimple_opt_pass along with...
14601 (pass_data_vrp): ...new pass_data instance and...
14602 (make_pass_vrp): ...new function.
14603 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
14604 subclass of simple_ipa_opt_pass along with...
14605 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
14606 (make_pass_ipa_free_lang_data): ...new function.
14607 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
14608 gimple_opt_pass along with...
14609 (pass_data_tsan): ...new pass_data instance and...
14610 (make_pass_tsan): ...new function.
14611 (pass_tsan_O0): Convert from a global struct to a subclass of
14612 gimple_opt_pass along with...
14613 (pass_data_tsan_O0): ...new pass_data instance and...
14614 (make_pass_tsan_O0): ...new function.
14615 * var-tracking.c (pass_variable_tracking): Convert from a global
14616 struct to a subclass of rtl_opt_pass along with...
14617 (pass_data_variable_tracking): ...new pass_data instance and...
14618 (make_pass_variable_tracking): ...new function.
14619 * web.c (pass_web): Convert from a global struct to a subclass of
14620 rtl_opt_pass along with...
14621 (pass_data_web): ...new pass_data instance and...
14622 (make_pass_web): ...new function.
14623 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
14624 declaration with that of...
14625 (make_pass_mode_switch_use): ...new function.
14626 (pass_resolve_sw_modes): Replace declaration with that of...
14627 (make_pass_resolve_sw_modes): ...new function.
14628 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
14629 from a global struct to a subclass of rtl_opt_pass along with...
14630 (pass_data_mode_switch_use): ...new pass_data instance and...
14631 (make_pass_mode_switch_use): ...new function.
14632 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
14633 from a global struct to a subclass of rtl_opt_pass along with...
14634 (pass_data_resolve_sw_modes): ...new pass_data instance and...
14635 (make_pass_resolve_sw_modes): ...new function.
14636 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
14637 struct to a subclass of rtl_opt_pass along with...
14638 (pass_data_insert_vzeroupper): ...new pass_data instance and...
14639 (make_pass_insert_vzeroupper): ...new function.
14640 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
14641 global struct to a subclass of rtl_opt_pass along with...
14642 (pass_data_work_around_errata): ...new pass_data instance and...
14643 (make_pass_work_around_errata): ...new function.
14644 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
14645 struct to a subclass of rtl_opt_pass along with...
14646 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
14647 (make_pass_mips_machine_reorg2): ...new function.
14649 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14651 * passes.c (pass_manager::operator new): New.
14653 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14655 Handwritten part of conversion of passes to C++ classes.
14657 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
14658 (toplev.o): Add dep on PASS_MANAGER_H.
14659 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
14660 of early local pases to reflect this moving from a global to a
14661 member of gcc::pass_manager.
14662 (cgraph_add_new_function): Likewise.
14663 * lto-cgraph.c (lto_output_node): Update for conversion of
14664 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
14665 * passes.c (opt_pass::clone): New.
14666 (opt_pass::gate): New.
14667 (opt_pass::execute): New.
14668 (opt_pass::opt_pass): New.
14669 (pass_manager::execute_early_local_passes): New.
14670 (pass_manager::execute_pass_mode_switching): new.
14671 (finish_optimization_passes): Convert to...
14672 (pass_manager::finish_optimization_passes): ...this.
14673 (finish_optimization_passes): Update for conversion of passes to
14675 (register_dump_files_1): Use has_gate since we cannot portably
14676 check a vtable entry against NULL.
14677 (dump_one_pass): Likewise.
14678 (ipa_write_summaries_2): Likewise.
14679 (ipa_write_optimization_summaries_1): Likewise.
14680 (ipa_read_summaries_1): Likewise.
14681 (ipa_read_optimization_summaries_1): Likewise.
14682 (execute_ipa_stmt_fixups): Likewise.
14683 (pass_manager::pass_manager): Rewrite pass-creation, invoking
14684 pass-creation functions rather than wiring up globals, and
14685 storing the results in fields of pass_manager generated using
14686 pass-instances.def.
14687 (pass_manager::dump_profile_report): Update for conversion of
14688 passes to C++ classes.
14689 (pass_manager::execute_ipa_summary_passes): Likewise.
14690 (execute_one_ipa_transform_pass): Likewise.
14691 (execute_one_pass): Use has_gate and has_execute since we cannot
14692 portably check a vtable entry against NULL.
14693 * pass_manager.h (pass_manager::finish_optimization_passes): New.
14694 (pass_manager): Use pass-instances.def to add fields for the
14695 various pass instances.
14696 * toplev.c (finalize): Update for move of
14697 finish_optimization_passes to a method of gcc::pass_manager.
14698 * toplev.h (finish_optimization_passes): Move to method of class
14700 * tree-pass.h (struct pass_data): New.
14701 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
14702 (opt_pass::gate): Convert to virtual function.
14703 (opt_pass::~opt_pass): New.
14704 (opt_pass::clone): New.
14705 (opt_pass::execute): Convert to virtual function.
14706 (opt_pass::opt_pass): New.
14707 (opt_pass::ctxt_): new.
14708 (gimple_opt_pass): Convert to subclass of opt_pass.
14709 (gimple_opt_pass::gimple_opt_pass): New.
14710 (rtl_opt_pass): Convert to subclass of opt_pass.
14711 (rtl_opt_pass::rtl_opt_pass): New.
14712 (ipa_opt_pass_d): Convert to subclass of opt_pass.
14713 (ipa_opt_pass_d::ipa_opt_pass_d): New.
14714 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
14715 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
14716 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
14717 invocation of pass_mode_switching to reflect this moving from a
14718 global to a member of gcc::pass_manager.
14719 (ix86_option_override): Rework how pass_insert_vzeroupper is
14720 added to the pass_manager to reflect autogenerated changes.
14721 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
14723 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
14725 PR rtl-optimization/57708
14726 * recog.c (peep2_find_free_register): Validate all regs in a
14729 2013-08-05 Jan Hubicka <jh@suse.cz>
14732 * cgraph.c (verify_cgraph_node): Accept local flags from other
14734 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
14735 (function_and_variable_visibility): Likewise.
14736 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
14738 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14740 * graph.c (init_graph_slim_pretty_print): Remove.
14741 (print_graph_cfg): Do not call it. Use local pretty printer.
14742 (start_graph_dump): Likewise.
14744 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14746 * gimple-pretty-print.c (buffer): Remove.
14747 (initialized): Likewise.
14748 (maybe_init_pretty_print): Likewise.
14749 (print_gimple_stmt): Do not call it. Use non-static local
14750 pretty_printer variable.
14751 (print_gimple_expr): Likewise.
14752 (print_gimple_seq): Likewise.
14753 (gimple_dump_bb): Likewise.
14755 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14757 * asan.c (asan_pp): Remove.
14758 (asan_pp_initialized): Likewise.
14759 (asan_pp_initialize): Likewise.
14760 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
14761 (asan_emit_stack_protection): Tidy. Use local pretty printer.
14762 (asan_add_global): Likewise.
14764 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
14766 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
14767 * diagnostic.h (diagnostic_flush_buffer): Adjust.
14768 * pretty-print.c (pp_formatted_text_data): Likewise.
14769 (pp_indent): Rename from pp_base_indent.
14770 (pp_format): Rename from pp_base_format.
14771 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
14772 (pp_format_verbatim): Rename from pp_base_format_verbatim.
14773 (pp_flush): Rename from pp_base_flush.
14774 (pp_set_line_maximum_length): Rename from
14775 pp_base_set_line_maximum_length.
14776 (pp_clear_output_area): Rename from pp_base_clear_output_area.
14777 (pp_set_prefix): Rename from pp_base_set_prefix.
14778 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
14779 (pp_emit_prefix): Rename from pp_base_emit_prefix.
14780 (pp_append_text): Rename from pp_base_append_text.
14781 (pp_formatted_text): Rename from pp_base_formatted_text.
14782 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
14783 (pp_remaining_character_count_for_line): Rename from
14784 pp_base_remaining_character_count_for_line.
14785 (pp_newline): Rename from pp_base_newline.
14786 (pp_character): Rename from pp_base_character.
14787 (pp_string): Rename from pp_base_string.
14788 (pp_maybe_space): Rename from pp_base_maybe_space.
14789 * asan.c (asan_pp_string): Adjust.
14790 (asan_emit_stack_protection): Likewise.
14791 (asan_add_global): Likewise.
14792 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
14793 * tree-mudflap.c (mf_varname_tree): Likewise.
14794 * tree-pretty-print.c (pp_tree_identifier): Rename from
14795 pp_base_tree_identifier.
14796 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
14797 Declare as function.
14799 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14801 * pretty-print.h (pp_bar_bar): New.
14802 (pp_ampersand_ampersand): Likewise.
14803 (pp_less_equal): Likewise.
14804 (pp_greater_equal): Likewise.
14805 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
14806 printer functions instead of pp_string or operators and punctuators.
14807 (dump_gimple_call): Likewise.
14808 (dump_gimple_omp_for): Likewise.
14809 (dump_gimple_transaction): Likewise.
14810 (dump_gimple_phi): Likewise.
14811 (pp_gimple_stmt_1): Likewise.
14812 * sched-vis.c (print_insn): Likewise.
14813 * tree-mudflap.c (mf_varname_tree): Likewise.
14814 * tree-pretty-print.c (dump_block_node): Likewise.
14815 (dump_generic_node): Likewise.
14817 2013-08-02 Jan Hubicka <jh@suse.cz>
14819 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
14821 * lto-streamer-out.c (tree_is_indexable): Results decls and
14822 parm decls are not indexable.
14823 (DFS_write_tree_body): Do not follow args and results.
14824 (hash_tree): Likewise.
14825 (output_functions): Rearrange so struct function is needed
14826 only when real body is output; be able to also ouptut abstract
14827 functions; output DECL_ARGUMENTS and DECL_RESULT.
14828 (lto_output): When not in WPA, ale store abstract functions.
14829 (write_symbol): Do not care about RESULT_DECL.
14830 (output_symbol_p): Handle correctly sbtract decls.
14831 * lto-streamer-in.c (input_function): Rearrange so struct
14832 function can be NULL at entry; allow streaming of
14833 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
14834 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
14835 sanity check during LTO.
14836 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
14837 RESULT_DECl and DECL_ARGUMENTS.
14838 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
14841 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14843 * pretty-print.h (pp_underscore): New.
14845 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
14846 printer functions instead of pp_character.
14847 (dump_binary_rhs): Likewise.
14848 (dump_ternary_rhs): Likewise.
14849 (dump_gimple_call_args): Likewise.
14850 (pp_points_to_solution): Likewise.
14851 (dump_gimple_call): Likewise.
14852 (dump_gimple_switch): Likewise.
14853 (dump_gimple_cond): Likewise.
14854 (dump_gimple_bind): Likewise.
14855 (dump_gimple_try): Likewise.
14856 (dump_gimple_omp_for): Likewise.
14857 (dump_gimple_omp_continue): Likewise.
14858 (dump_gimple_omp_single): Likewise.
14859 (dump_gimple_omp_sections): Likewise.
14860 (dump_gimple_omp_block): Likewise.
14861 (dump_gimple_omp_critical): Likewise.
14862 (dump_gimple_transaction): Likewise.
14863 (dump_gimple_asm): Likewise.
14864 (dump_gimple_phi): Likewise.
14865 (dump_gimple_omp_parallel): Likewise.
14866 (dump_gimple_omp_task): Likewise.
14867 (dump_gimple_omp_atomic_load): Likewise.
14868 (dump_gimple_omp_atomic_store): Likewise.
14869 (dump_gimple_mem_ops): Likewise.
14870 (pp_gimple_stmt_1): Likewise.
14871 (pp_cfg_jump): Likewise.
14872 (dump_implicit_edges): Likewise.
14873 (gimple_dump_bb_for_graph): Likewise.
14874 * graph.c (draw_cfg_node): Likewise.
14875 * langhooks.c (lhd_print_error_function): Likewise.
14876 * sched-vis.c (print_exp): Likewise.
14877 (print_value): Likewise.
14878 (print_pattern): Likewise.
14879 (print_insn): Likewise.
14880 (rtl_dump_bb_for_graph): Likewise.
14881 * tree-pretty-print.c (dump_function_declaration): Likewise.
14882 (dump_array_domain): Likewise.
14883 (dump_omp_clause): Likewise.
14884 (dump_location): Likewise.
14885 (dump_generic_node): Likewise.
14886 (print_struct_decl): Likewise.
14887 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14889 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14891 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14892 candidate table when replacing a candidate statement.
14893 (replace_rhs_if_not_dup): Likewise.
14894 (replace_one_candidate): Likewise.
14896 2013-08-02 Jan Hubicka <jh@suse.cz>
14897 Martin Liska <marxin.liska@gmail.com>
14899 * cgraphunit.c (add_new_function): Fix logic when adding from
14901 (assemble_thunk): Rename to ...
14902 (expand_thunk); .. this one; export; get it working with
14903 general functions; make produced gimple valid.
14904 * cgraph.h (expand_thunk): Declare.
14906 2013-08-02 Jan Hubicka <jh@suse.cz>
14908 * ipa-cp.c (gather_context_independent_values): Use
14909 ipa_get_param_move_cost.
14910 (get_replacement_map): Remove PARAM; move parameter folding
14912 (create_specialized_node): Update.
14913 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14914 assert that we have gimple body; update move_cost.
14915 (count_formal_params): Assert that we have gimple body.
14916 (ipa_dump_param): New function.
14917 (ipa_alloc_node_params): Break out from ...
14918 (ipa_initialize_node_params): ... here.
14919 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14920 (ipa_write_node_info): Stream move costs.
14921 (ipa_read_node_info): Read move costs.
14922 (ipa_update_after_lto_read): Do not recompute node params.
14923 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14924 (ipa_get_param): Check we are not in WPA.
14925 (ipa_get_param_move_cost): New.
14926 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14927 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14928 parm numbers to be present.
14930 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14932 PR rtl-optimization/58048
14933 * lra-constraints.c (process_alt_operands): Don't check asm
14934 operand on register.
14936 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14938 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14939 the implied StoreLoad barrier for atomic operations if before.
14941 2013-08-02 Jan Hubicka <jh@suse.cz>
14942 Martin Liska <marxin.liska@gmail.com>
14944 * cgraph.c (cgraph_function_body_availability): Do not check
14946 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14947 symtab_node_availability): Declare.
14948 * ipa.c (can_replace_by_local_alias): New.
14949 (function_and_variable_visibility): Use it.
14950 * symtab.c (symtab_for_node_and_aliases,
14951 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14953 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14955 PR rtl-optimization/57963
14956 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14957 (lra_constraints): Use them.
14959 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14961 * config/arm/types.md (define_attr "type"): Add "load_acq"
14963 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14965 (cortex_a53_store1): Likewise.
14967 2013-08-01 Jan Hubicka <jh@suse.cz>
14969 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14970 partitions are not needed.
14972 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14974 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14975 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14976 MAYBE_NON_Q_CLASS_P where appropriate.
14978 2013-08-01 Jan Hubicka <jh@suse.cz>
14980 * cgraph.h (release_function_body): Declare.
14981 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14982 of unused delcarations.
14984 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14986 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14987 RTL form when subtracting a constant.
14989 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14991 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14992 Generate canonical plus rtx with negated immediate instead of minus
14994 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14996 2013-08-01 Jan Hubicka <jh@suse.cz>
14998 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14999 (cgraph_release_function_body): Likewise.
15000 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
15001 * cgraph.h (cgrpah_node): Rename abstract_and_needed
15002 to used_as_abstract_origin.
15003 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
15004 symbols used as abstract origins.
15005 * cgraphunit.c (analyze_functions): Update.
15006 * ipa.c (symtab_remove_unreachable_nodes): Recompute
15007 used_as_abstract_origin.
15008 * tree-inline.c (tree_function_versioning): Update
15009 used_as_abstract_origin; be ready for DECL_RESULT and
15010 DECL_ARGUMENTS to be NULL.
15012 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
15013 for abstract functions.
15014 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
15017 2013-08-01 Jan Hubicka <jh@suse.cz>
15019 * profile.c (compute_value_histograms): Fix thinko.
15021 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
15023 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
15024 aarch-common-protos.h to extra_headers.
15025 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
15026 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
15027 * config/aarch64/t-aarch64 (aarch-common.o): Define.
15029 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
15031 * config/aarch64/aarch64.md (define_attr "type"): Delete.
15032 Include "../arm/types.md". Define "type" attribute for all patterns.
15033 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
15036 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
15038 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
15039 to support power8 load fusion.
15040 (fusion_gpr_mem_load): Likewise.
15042 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
15044 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
15045 declarations for power8 load fusion.
15046 (emit_fusion_gpr_load): Likewise.
15048 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
15049 tuning for power8, turn on fusion mode by default. Turn on sign
15050 extending fusion mode if normal fusion mode is on, and we are at
15052 (fusion_gpr_load_p): New function, return true if we can fuse an
15053 addis instruction with a dependent load to a GPR.
15054 (emit_fusion_gpr_load): Emit the instructions for power8 load
15057 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
15058 (VSX load fusion peepholes): New peepholes to fuse together an
15059 addi instruction with a VSX load instruction.
15061 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
15062 peepholes to fuse an addis instruction with a load to a GPR base
15063 register. If we are supporting sign extending fusions, convert
15064 sign extending loads to zero extending loads and add an explicit
15067 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
15069 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
15070 aarch-common-protos.h to extra_headers.
15071 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
15072 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
15073 (arm_early_store_addr_dep): Likewise.
15074 (arm_no_early_alu_shift_dep): Likewise.
15075 (arm_no_early_alu_shift_value_dep): Likewise.
15076 (arm_no_early_mul_dep): Likewise.
15077 (arm_no_early_store_addr_dep): Likewise.
15078 (arm_mac_accumulator_is_mul_result): Likewise.
15079 (arm_mac_accumulator_is_result): Likewise.
15080 * config/arm/aarch-common.c: ... here. New file.
15081 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
15083 (arm_early_store_addr_dep): Likewise.
15084 (arm_no_early_alu_shift_dep): Likewise.
15085 (arm_no_early_alu_shift_value_dep): Likewise.
15086 (arm_no_early_mul_dep): Likewise.
15087 (arm_no_early_store_addr_dep): Likewise.
15088 (arm_mac_accumulator_is_mul_result): Likewise.
15089 (arm_mac_accumulator_is_result): Likewise.
15090 * config/arm/aarch-common-protos.h: ... here. New file.
15091 * config/arm/t-arm (aarch-common.o): Define.
15093 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
15095 * config/arm/arm.md: Include new file "types.md".
15096 (define_attr "type"): Move from here to ...
15097 (define_attr "mul32"): Likewise.
15098 (define_attr "mul64"): Likewise.
15099 * config/arm/types.md: ... here. New file.
15101 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
15103 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
15104 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
15106 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15108 * gen-pass-instances.awk: Fix offset of substr().
15110 2013-07-31 David Malcolm <dmalcolm@redhat.com>
15112 * Makefile.in (pass-instances.def): New.
15113 (passes.o): Replace dependency on passes.def with one on
15116 * gen-pass-instances.awk: New.
15118 * passes.c (pass_manager::pass_manager): Use pass-instances.def
15119 rather than passes.def, updating local definition of NEXT_PASS
15120 macro to add an extra NUM parameter (currently unused).
15122 2013-07-30 David Malcolm <dmalcolm@redhat.com>
15124 * Makefile.in (PASS_MANAGER_H): New.
15125 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
15126 (passes.o): Likewise.
15127 (statistics.o): Likewise.
15128 (cgraphunit.o): Likewise.
15129 (context.o): Depend on PASS_MANAGER_H.
15131 * pass_manager.h: New.
15133 * cgraphunit.c (cgraph_add_new_function): Update for moves
15134 of globals to fields of pass_manager.
15135 (analyze_function): Likewise.
15136 (expand_function): Likewise.
15137 (ipa_passes): Likewise.
15138 (compile): Likewise.
15140 * context.c (context::context): New.
15141 * context.h (context::context): New.
15142 (context::get_passes): New.
15143 (context::passes_): New.
15145 * lto-cgraph.c (input_node): Update for moves of globals to
15146 fields of pass_manager.
15148 * passes.c (all_passes): Remove, in favor of a field of the
15149 same name within the new class pass_manager.
15150 (all_small_ipa_passes): Likewise.
15151 (all_lowering_passes): Likewise.
15152 (all_regular_ipa_passes): Likewise.
15153 (all_late_ipa_passes): Likewise.
15154 (all_lto_gen_passes): Likewise.
15155 (passes_by_id): Likewise.
15156 (passes_by_id_size): Likewise.
15157 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
15158 the new class pass_manager.
15159 (set_pass_for_id): Convert to...
15160 (pass_manager::set_pass_for_id): ...method.
15161 (get_pass_for_id): Convert to...
15162 (pass_manager::get_pass_for_id): ...method.
15163 (register_one_dump_file): Move body of implementation into...
15164 (pass_manager::register_one_dump_file): ...here.
15165 (register_dump_files_1): Convert to...
15166 (pass_manager::register_dump_files_1): ...method.
15167 (register_dump_files): Convert to...
15168 (pass_manager::register_dump_files): ...method.
15169 (create_pass_tab): Update for moves of globals to fields of
15171 (dump_passes): Move body of implementation into...
15172 (pass_manager::dump_passes): ...here.
15173 (register_pass): Move body of implementation into...
15174 (pass_manager::register_pass): ...here.
15175 (init_optimization_passes): Convert into...
15176 (pass_manager::pass_manager): ...constructor for new
15177 pass_manager class, and initialize the pass_lists array.
15178 (check_profile_consistency): Update for moves of globals to
15179 fields of pass_manager.
15180 (dump_profile_report): Move body of implementation into...
15181 (pass_manager::dump_profile_report): ...here.
15182 (ipa_write_summaries_1): Update for moves of pass lists from
15183 being globals to fields of pass_manager.
15184 (ipa_write_optimization_summaries): Likewise.
15185 (ipa_read_summaries): Likewise.
15186 (ipa_read_optimization_summaries): Likewise.
15187 (execute_all_ipa_stmt_fixups): Likewise.
15189 * statistics.c (statistics_fini): Update for moves of globals to
15190 fields of pass_manager.
15192 * toplev.c (general_init): Replace call to
15193 init_optimization_passes with construction of the pass_manager
15196 * tree-pass.h (all_passes): Remove, in favor of a field of the
15197 same name within the new class pass_manager.
15198 (all_small_ipa_passes): Likewise.
15199 (all_lowering_passes): Likewise.
15200 (all_regular_ipa_passes): Likewise.
15201 (all_lto_gen_passes): Likewise.
15202 (all_late_ipa_passes): Likewise.
15203 (passes_by_id): Likewise.
15204 (passes_by_id_size): Likewise.
15205 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
15206 the new class pass_manager.
15207 (get_pass_for_id): Remove.
15209 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
15211 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
15214 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
15216 * arm.md (mulhi3): New expand pattern.
15218 2013-07-30 Jan Hubicka <jh@suse.cz>
15219 Martin Liska <marxin.liska@gmail.com>
15221 * profile.c (compute_value_histograms): Do not ICE when
15222 there is mismatch only on some counters.
15224 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15226 PR rtl-optimization/57637
15227 * function.c (move_insn_for_shrink_wrap): Also check the
15228 GEN set of the LIVE problem for the liveness analysis
15229 if it exists, otherwise give up.
15231 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15233 PR tree-optimization/57993
15234 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
15235 replaced statement in the candidate table.
15236 (phi_add_costs): Return infinite cost when the hidden basis does
15237 not dominate all phis on which the candidate is dependent.
15238 (replace_one_candidate): Record replaced statement in the
15241 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
15243 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
15244 (ashlv2si3): New expander.
15245 (*ashlv2si3_i): New define_insn_and_split.
15246 * predicates.md (float_operation): Allow patterns with three
15247 basic sub-patterns.
15249 PR rtl-optimization/58021
15250 * mode-switching.c (create_pre_exit): Always split off preceding
15251 insns if we are not at the basic block head.
15253 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
15255 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
15256 (UCLIBC_DYNAMIC_LINKER): New macro.
15257 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
15259 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
15260 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
15262 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
15263 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
15264 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
15265 for SF and DF modes. Use ieee_quad_format for TF mode.
15266 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
15267 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
15268 (mips_option_override): Handle `-mnan=legacy'.
15269 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
15270 `-mabs=2008' and `-mnan=2008'.
15271 (OPTION_DEFAULT_SPECS): Add "nan" default.
15272 (ASM_SPEC): Handle `-mnan='.
15273 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
15274 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
15275 comment accordingly.
15276 (neg<mode>2): Likewise.
15277 * config/mips/mips.opt (mabs, mnan): New options.
15278 * doc/install.texi (Configuration): Document `--with-nan=' option.
15279 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
15281 (MIPS Options): Document them.
15282 * config.gcc <mips*-*-*>: Handle `--with-nan='.
15283 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
15284 * configure: Regenerate.
15285 * config.in: Regenerate.
15287 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
15289 * config/i386/i386.md (float post-reload splitters): Do not check
15290 for subregs of SSE registers.
15292 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
15293 H.J. Lu <hongjiu.lu@intel.com>
15297 * config/i386/i386.md (post-reload splitter
15298 to avoid partial SSE reg dependency stalls): New pattern.
15300 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
15302 * config/s390/s390.md ("movcc"): Swap load and store instructions.
15304 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
15306 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
15307 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
15309 2013-07-26 Cary Coutant <ccoutant@google.com>
15311 * dwarf2out.c (die_checksum_ordered): Don't include template
15312 instantiations in signature.
15313 (is_template_parameter): New function.
15314 (is_template_instantiation): New function.
15315 (generate_skeleton_bottom_up): Don't include template instantiations
15317 (generate_skeleton): Likewise.
15318 (break_out_comdat_types): Move recursive call to break out nested
15320 (prune_unused_types_mark_generic_parms_dies): Call
15321 is_template_parameter.
15323 2013-07-26 Ian Bolton <ian.bolton@arm.com>
15325 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
15326 uses vector registers.
15327 * config/aarch64/iterators.md: Add attributes rtn and vas.
15329 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15330 Richard Earnshaw <richard.earnshaw@arm.com>
15332 * combine.c (simplify_comparison): Re-canonicalize operands
15334 * config/arm/arm.md (movcond_addsi): New splitter.
15336 2013-07-25 Sterling Augustine <saugustine@google.com>
15338 * dwarf2out.c (size_of_pubnames): Move code to...
15339 (include_pubname_in_output): ...here. New.
15340 (want_pubnames): Rearrange.
15341 (output_pubnames): Call include_pubname_in_output. Move assertion.
15343 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
15345 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
15347 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
15350 * doc/extend.texi: Remove obsolete builtins. Fix
15351 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
15353 2013-07-25 Jan Hubicka <jh@suse.cz>
15355 * cgraph.c (release_function_body): Break out from ...
15356 (cgraph_release_function_body): ... this one; also release DECL_RESULT
15357 and DECL_ARGUMENTS.
15358 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
15359 old_tree in the map.
15360 (create_specialized_node): Update.
15361 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
15363 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
15364 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
15365 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
15366 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
15369 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15371 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
15372 addsi3_carryin_alt2_<optab>): Correct output template.
15374 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15376 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
15377 Adjust for arm_restrict_it.
15378 Remove trailing whitespace.
15380 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
15382 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
15383 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
15385 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
15387 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
15389 PR rtl-optimization/57960
15390 * lra-constraints.c (process_alt_operands): Use the right mode
15391 when checking strict_low.
15393 2013-07-25 Jan Hubicka <jh@suse.cz>
15395 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
15396 * cgraph.c (cgraph_remove_node): Do not release function body
15397 when in cgraph streaming.
15398 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
15399 in other partitions are not considered reachable; fix handling of
15402 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15404 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
15406 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15411 * config/arm/arm.md ("*sibcall_insn): Replace use of
15412 Ss with US. Adjust output for v5 and v4t.
15413 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
15415 * config/arm/constraints.md ("Ss"): Rename to US.
15417 2013-07-25 Terry Guo <terry.guo@arm.com>
15419 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
15420 shift_add/shift_sub0/shift_sub1 RTXs.
15422 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
15423 Anton Blanchard <anton@au1.ibm.com>
15425 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
15426 (altivec_vpks<VI_char>ss): Likewise.
15427 (altivec_vpks<VI_char>us): Likewise.
15428 (altivec_vpku<VI_char>us): Likewise.
15429 (altivec_vpku<VI_char>um): Likewise.
15431 2013-07-24 David Malcolm <dmalcolm@redhat.com>
15433 Introduce context class.
15435 * Makefile.in (CONTEXT_H): New.
15436 (OBJS): Add context.o.
15437 (toplev.o): Add CONTEXT_H to dependencies.
15440 * toplev.c (general_init): Create the singleton gcc::context instance.
15446 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
15448 PR rtl-optimization/57968
15449 * mode-switching.c (create_pre_exit): Allow instructions that
15450 don't set a return register to need a non-exit mode.
15452 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15453 Anton Blanchard <anton@au1.ibm.com>
15455 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
15456 operands to vperm for little endian.
15457 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
15458 of lvsl to create the control mask for a vperm for little endian.
15460 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15461 Anton Blanchard <anton@au1.ibm.com>
15463 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
15464 two operands for little-endian.
15466 2013-07-23 Steve Ellcey <sellcey@mips.com>
15468 * config/mips/mips.c (mips_case_values_threshold): New.
15469 (TARGET_CASE_VALUES_THRESHOLD): Define.
15471 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15472 Anton Blanchard <anton@au1.ibm.com>
15474 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
15475 selection of field for vector splat in little endian mode.
15477 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
15479 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
15480 expanders to rs6000.md.
15481 (ior<mode>3): Likewise.
15482 (and<mode>3): Likewise.
15483 (one_cmpl<mode>2): Likewise.
15484 (nor<mode>3): Likewise.
15485 (andc<mode>3): Likewise.
15486 (eqv<mode>3): Likewise.
15487 (nand<mode>3): Likewise.
15488 (orc<mode>3): Likewise.
15490 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
15493 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
15494 to split multi-word logical operations.
15495 (rs6000_split_logical_di): Likewise.
15496 (rs6000_split_logical): Likewise.
15498 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
15499 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
15500 and allow TImode operations in 32-bit.
15501 (vsx_and<mode>3_64bit): Likewise.
15502 (vsx_ior<mode>3_32bit): Likewise.
15503 (vsx_ior<mode>3_64bit): Likewise.
15504 (vsx_xor<mode>3_32bit): Likewise.
15505 (vsx_xor<mode>3_64bit): Likewise.
15506 (vsx_one_cmpl<mode>2_32bit): Likewise.
15507 (vsx_one_cmpl<mode>2_64bit): Likewise.
15508 (vsx_nor<mode>3_32bit): Likewise.
15509 (vsx_nor<mode>3_64bit): Likewise.
15510 (vsx_andc<mode>3_32bit): Likewise.
15511 (vsx_andc<mode>3_64bit): Likewise.
15512 (vsx_eqv<mode>3_32bit): Likewise.
15513 (vsx_eqv<mode>3_64bit): Likewise.
15514 (vsx_nand<mode>3_32bit): Likewise.
15515 (vsx_nand<mode>3_64bit): Likewise.
15516 (vsx_orc<mode>3_32bit): Likewise.
15517 (vsx_orc<mode>3_64bit): Likewise.
15519 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
15520 logical types in GPRs.
15522 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
15523 logical insns to rs6000.md, and allow TImode operations in
15525 (altivec_ior<mode>3): Likewise.
15526 (altivec_xor<mode>3): Likewise.
15527 (altivec_one_cmpl<mode>2): Likewise.
15528 (altivec_nor<mode>3): Likewise.
15529 (altivec_andc<mode>3): Likewise.
15531 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
15532 attributes for moving the 128-bit logical operations into
15534 (BOOL_REGS_OUTPUT): Likewise.
15535 (BOOL_REGS_OP1): Likewise.
15536 (BOOL_REGS_OP2): Likewise.
15537 (BOOL_REGS_UNARY): Likewise.
15538 (BOOL_REGS_AND_CR0): Likewise.
15539 (one_cmpl<mode>2): Add support for DI logical operations on
15540 32-bit, splitting the operations to 32-bit.
15541 (anddi3): Likewise.
15542 (iordi3): Likewise.
15543 (xordi3): Likewise.
15544 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
15545 changes to combine the 32/64-bit code, allow logical operations on
15546 TI mode in 32-bit, and to use similar match_operator patterns like
15547 scalar mode uses. Combine the Altivec and VSX code for logical
15548 operations, and move it here.
15549 (ior<mode>3, 128-bit types): Likewise.
15550 (xor<mode>3, 128-bit types): Likewise.
15551 (one_cmpl<mode>3, 128-bit types): Likewise.
15552 (nor<mode>3, 128-bit types): Likewise.
15553 (andc<mode>3, 128-bit types): Likewise.
15554 (eqv<mode>3, 128-bit types): Likewise.
15555 (nand<mode>3, 128-bit types): Likewise.
15556 (orc<mode>3, 128-bit types): Likewise.
15557 (and<mode>3_internal): Likewise.
15558 (bool<mode>3_internal): Likewise.
15559 (boolc<mode>3_internal1): Likewise.
15560 (boolc<mode>3_internal2): Likewise.
15561 (boolcc<mode>3_internal1): Likewise.
15562 (boolcc<mode>3_internal2): Likewise.
15563 (eqv<mode>3_internal1): Likewise.
15564 (eqv<mode>3_internal2): Likewise.
15565 (one_cmpl1<mode>3_internal): Likewise.
15567 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15569 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15570 Rename flag_stack_usage to flag_stack_usage_info.
15572 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15574 * config/microblaze/sync.md: New file.
15575 * config/microblaze/microblaze.md: Include sync.md
15576 * config/microblaze/microblaze.c: Add print_operand 'y'.
15577 * config/microblaze/constraints.md: Add memory_contraint
15578 'Q' which is a single register.
15580 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
15582 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
15584 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15586 * reload.c (find_reloads): Exit loop once we find this operand
15587 cannot be reloaded somehow for this alternative.
15589 * reload.c (find_reloads): Exit loop once we find a hard register.
15591 * rtlanal.c (computed_jump_p): Exit loop once we find label
15594 * i386.c (ix86_pad_returns): Exit loop after setting replace.
15596 * cfgloopmanip.c (remove_path): Exit loop after setting
15599 * gensupport.c (subst_dup): Avoid loop if code is not
15600 MATCH_DUP nor MATCH_OP_DUP.
15602 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
15604 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
15606 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15608 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
15609 true for SP_REGNUM if mode == ptr_mode.
15610 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
15611 with value R0_REGNUM + 31.
15613 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15615 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
15616 pad pointer-typed argument downward.
15618 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15620 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
15621 and __ILP32__ when the ILP32 model is in use.
15623 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15625 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
15626 (aarch64_load_symref_appropriately): In the case of
15627 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
15628 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
15629 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
15630 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
15631 if 'mode' doesn't equal to 'ptr_mode'.
15632 (aarch64_output_mi_thunk): Add an assertion on the alignment of
15633 'vcall_offset'; change to call aarch64_emit_move differently depending
15634 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
15635 to calculate the upper bound of 'vcall_offset'.
15636 (aarch64_cannot_force_const_mem): Change to also return true if
15638 (aarch64_legitimize_reload_address): In the case of large
15639 displacements, add new local variable 'xmode' and an assertion
15640 based on it; change to use 'xmode' to generate the new rtx and
15642 (aarch64_asm_trampoline_template): Change to generate the template
15643 differently depending on TARGET_ILP32 or not; change to use
15644 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
15645 (aarch64_trampoline_size): Removed.
15646 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
15647 and replace immediate literals with it. Change to use 'ptr_mode'
15648 instead of 'DImode' and call convert_memory_address if the mode
15649 of 'fnaddr' doesn't equal to 'ptr_mode'.
15650 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
15652 (aarch64_elf_asm_destructor): Likewise.
15653 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
15654 on TARGET_ILP32 instead of aarch64_trampoline_size.
15655 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
15656 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
15657 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
15658 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
15659 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
15660 (storewb_pair<GPI:mode>_<P:mode>): ... this.
15661 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
15662 depending on the value of 'mode'.
15663 (add_losym_<mode>): New.
15664 (ldr_got_small_<mode>): New, based on ldr_got_small.
15665 (ldr_got_small): Remove.
15666 (ldr_got_small_sidi): New.
15667 * config/aarch64/iterators.md (P): New.
15668 (PTR): Change to 'ptr_mode' in the condition.
15670 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15672 * config.gcc (aarch64*-*-*): Support --with-abi.
15673 (aarch64*-*-elf): Support --with-multilib-list.
15674 (aarch64*-*-linux*): Likewise.
15675 (supported_defaults): Add abi to aarch64*-*-*.
15676 * configure.ac: Mention AArch64 for --with-multilib-list.
15677 * configure: Re-generated.
15678 * config/aarch64/biarchilp32.h: New file.
15679 * config/aarch64/biarchlp64.h: New file.
15680 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
15682 (MULTILIB_DEFAULTS): Ditto.
15683 (DRIVER_SELF_SPECS): Ditto.
15684 (ASM_SPEC): Update to also substitute -mabi.
15685 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
15686 file whose name depends on -mabi= and -mbig-endian.
15687 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
15689 (POINTER_SIZE): New define.
15690 (POINTERS_EXTEND_UNSIGNED): Ditto.
15691 (enum aarch64_abi_type): New enumeration tag.
15692 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
15693 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
15694 (TARGET_ILP32): New define.
15695 * config/aarch64/aarch64.opt (mabi): New.
15696 (aarch64_abi): New.
15697 (ilp32, lp64): New values for -mabi.
15698 * config/aarch64/t-aarch64 (comma): New define.
15699 (MULTILIB_OPTIONS): Ditto.
15700 (MULTILIB_DIRNAMES): Ditto.
15701 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
15702 * doc/invoke.texi: Document -mabi for AArch64.
15704 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
15706 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
15708 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15709 Anton Blanchard <anton@au1.ibm.com>
15711 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
15712 endianness when selecting field to splat.
15714 2013-07-22 Eric Christopher <echristo@gmail.com>
15716 * dwarf2out.c (die_odr_checksum): New function to use
15717 CHECKSUM_ macros and ULEB128 for DIE tag.
15718 (generate_type_signature): Use.
15720 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
15722 * config.gcc (sparc*-*-*): Accept leon3 processor.
15723 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
15724 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
15725 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
15726 * config/sparc/sparc.opt (enum processor_type): Add leon3.
15727 (mfix-ut699): Adjust comment.
15728 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
15729 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
15730 (CPP_CPU_SPEC): Likewise.
15731 (ASM_CPU_SPEC): Likewise.
15732 * config/sparc/sparc.c (leon3_cost): New constant.
15733 (sparc_option_override): Add leon3 support.
15734 (mem_ref): New function.
15735 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
15736 (sparc_do_work_around_errata): Look into the instruction in the delay
15737 slot and adjust accordingly. Add fix for the data cache nullify issues
15738 of the UT699. Change insertion position for the NOP.
15739 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
15740 (leon3_load): New reservation.
15741 (leon_store): Bump latency to 2.
15742 (grfpu): New automaton.
15743 (grfpu_alu): New unit.
15744 (grfpu_ds): Likewise.
15745 (leon_fp_alu): Adjust.
15746 (leon_fp_mult): Delete.
15747 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
15748 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
15749 * config/sparc/sparc.md (cpu): Add leon3.
15750 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
15751 (swapsi): Likewise.
15752 (atomic_test_and_set): Likewise.
15753 (ldstub): Likewise.
15755 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
15757 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
15758 default for R5900 targets.
15759 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
15760 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
15761 * config/mips/mips.c (mips_option_override): Report an error for
15762 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
15763 for -march=r5900 -mhard-float.
15765 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15767 * df-problems.c (can_move_insns_across): Exit loop once we
15768 find a non-fixed, non-global register.
15770 * ipa-pure-const.c (propagate_nothrow): Exit loop after
15773 * omega.c (omega_eliminate_red): Break after setting red_found.
15774 (omega_problem_has_red_equations): Similarly after setting found.
15775 (omega_query_variable): Similarly after setting coupled.
15777 2013-07-22 Marek Polacek <polacek@redhat.com>
15779 * gimplify.c: Don't include gimple.h twice.
15781 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15783 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
15784 instead of TARGET_THUMB1.
15785 (Pz): New constraint.
15786 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
15788 (compare_negsi_si): Likewise.
15789 (compare_addsi2_op0): Likewise.
15790 (compare_addsi2_op1): Likewise.
15791 (addsi3_carryin_<optab>): Likewise.
15792 (addsi3_carryin_alt2_<optab>): Likewise.
15793 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
15794 for arm_restrict_it.
15795 (subsi3_carryin): Likewise.
15796 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
15797 (minmax_arithsi): Disable for arm_restrict_it.
15798 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
15799 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
15800 (satsi_<SAT:code>_shift): Likewise.
15801 (arm_shiftsi3): Add alternative for 16-bit encoding.
15802 (arm32_movhf): Disable for arm_restrict_it.
15803 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
15804 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
15806 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
15808 * config/arm/arm.md (attribute "insn"): Delete.
15809 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
15810 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
15811 (not_shiftsi): Update for attribute change.
15812 (not_shiftsi_compare0): Likewise.
15813 (not_shiftsi_compare0_scratch): Likewise.
15814 (arm_one_cmplsi2): Likewise.
15815 (thumb1_one_cmplsi2): Likewise.
15816 (notsi_compare0): Likewise.
15817 (notsi_compare0_scratch): Likewise.
15818 (thumb1_movdi_insn): Likewise.
15819 (arm_movsi_insn): Likewise.
15820 (movhi_insn_arch4): Likewise.
15821 (movhi_bytes): Likewise.
15822 (arm_movqi_insn): Likewise.
15823 (thumb1_movqi_insn): Likewise.
15824 (arm32_movhf): Likewise.
15825 (thumb1_movhf): Likewise.
15826 (arm_movsf_soft_insn): Likewise.
15827 (thumb1_movsf_insn): Likewise.
15828 (thumb_movdf_insn): Likewise.
15829 (movsicc_insn): Likewise.
15830 (movsfcc_soft_insn): Likewise.
15831 (and_scc): Likewise.
15832 (cond_move): Likewise.
15833 (if_move_not): Likewise.
15834 (if_not_move): Likewise.
15835 (if_shift_move): Likewise.
15836 (if_move_shift): Likewise.
15837 (if_shift_shift): Likewise.
15838 (if_not_arith): Likewise.
15839 (if_arith_not): Likewise.
15840 (cond_move_not): Likewise.
15841 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
15842 (neon_mov<mode>): Likewise.
15843 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
15844 (thumb2_movsi_vfp): Likewise.
15845 (movsf_vfp): Likewise.
15846 (thumb2_movsf_vfp): Likewise.
15847 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15849 (cortexa7_older_only): Likewise.
15850 (cortexa7_younger): Likewise.
15851 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
15852 (1020alu_shift_op): Likewise.
15853 (1020alu_shift_reg_op): Likewise.
15854 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15855 (alu_shift_op): Likewise.
15856 (alu_shift_reg_op): Likewise.
15857 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15858 (11_alu_shift_op): Likewise.
15859 (11_alu_shift_reg_op): Likewise.
15860 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15861 (9_alu_shift_reg_op): Likewise.
15862 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15864 (cortex_a15_alu_shift): Likewise.
15865 (cortex_a15_alu_shift_reg): Likewise.
15866 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15867 (cortex_a5_alu_shift): Likewise.
15868 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15870 (cortex_a53_alu_shift): Likewise.
15871 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15873 (cortex_a7_alu_reg): Likewise.
15874 (cortex_a7_alu_shift): Likewise.
15875 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15876 (cortex_a8_alu_shift): Likewise.
15877 (cortex_a8_alu_shift_reg): Likewise.
15878 (cortex_a8_mov): Likewise.
15879 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15880 (cortex_a9_dp_shift): Likewise.
15881 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15882 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15883 (cortex_r4_mov): Likewise.
15884 (cortex_r4_alu_shift): Likewise.
15885 (cortex_r4_alu_shift_reg): Likewise.
15886 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15887 (526_alu_shift_op): Likewise.
15888 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15889 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15890 (626te_alu_shift_op): Likewise.
15891 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15892 (726te_alu_op): Likewise.
15893 (726te_alu_shift_op): Likewise.
15894 (726te_alu_shift_reg_op): Likewise.
15895 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15896 (mp626_alu_shift_op): Likewise.
15897 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15898 (pj4_alu_e1_conds): Likewise.
15899 (pj4_alu): Likewise.
15900 (pj4_alu_conds): Likewise.
15901 (pj4_shift): Likewise.
15902 (pj4_shift_conds): Likewise.
15903 (pj4_alu_shift): Likewise.
15904 (pj4_alu_shift_conds): Likewise.
15906 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15908 * config/arm/predicates.md (shiftable_operator_strict_it):
15910 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15911 Disable cond_exec version for arm_restrict_it.
15912 (thumb2_smaxsi3): Convert to generate cond_exec.
15913 (thumb2_sminsi3): Likewise.
15914 (thumb32_umaxsi3): Likewise.
15915 (thumb2_uminsi3): Likewise.
15916 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15917 (thumb2_neg_abssi2): Likewise.
15918 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15919 (thumb2_movsicc_insn): Adjust alternatives.
15920 (thumb2_mov_negscc): Disable for arm_restrict_it.
15921 (thumb2_mov_negscc_strict_it): New pattern.
15922 (thumb2_mov_notscc_strict_it): New pattern.
15923 (thumb2_mov_notscc): Disable for arm_restrict_it.
15924 (thumb2_ior_scc): Likewise.
15925 (thumb2_ior_scc_strict_it): New pattern.
15926 (thumb2_cond_move): Adjust for arm_restrict_it.
15927 (thumb2_cond_arith): Disable for arm_restrict_it.
15928 (thumb2_cond_arith_strict_it): New pattern.
15929 (thumb2_cond_sub): Adjust for arm_restrict_it.
15930 (thumb2_movcond): Likewise.
15931 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15932 (thumb2_zero_extendhisi2_v6): Likewise.
15933 (thumb2_zero_extendqisi2_v6): Likewise.
15934 (orsi_notsi_si): Likewise.
15935 (orsi_not_shiftsi_si): Likewise.
15937 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15939 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15940 instruction sequence is 1 byte shorter.
15942 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15944 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15945 it is not needed after split.
15947 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15950 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15951 second label for nonlocal goto receivers. Don't output pic base labels
15952 unless we're producing PIC; mark that action unreachable().
15953 (ix86_save_reg): If the function contains a nonlocal label, save the
15955 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15956 * config/darwin.c (emitted_pic_label_num): New GTY.
15957 (update_pic_label_number_if_needed): New.
15958 (machopic_output_function_base_name): Adjust for nonlocal receiver
15960 (machopic_should_output_picbase_label): New.
15961 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15962 (nonlocal_goto_receiver): New insn and split.
15964 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15966 * config/aarch64/aarch64-builtins.c
15967 (aarch64_fold_builtin): Fold abs in all modes.
15968 * config/aarch64/aarch64-simd-builtins.def
15969 (abs): Enable for all modes.
15970 * config/aarch64/arm_neon.h
15971 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15972 (vabs_f64): Add missing intrinsic.
15974 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15976 * config/aarch64/arm_neon.h (vabs_s64): New function
15978 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15981 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15982 * config/avr/avr.md (adjust_len): Add `round'.
15983 * config/avr/avr-protos.h (avr_out_round): New prototype.
15984 (avr_out_plus): Add `out_label' argument.
15985 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15986 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15987 Handle the case where `insn' is just a pattern.
15988 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15989 (avr_out_round): New function.
15990 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15992 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15994 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15995 Add check for flag_stack_usage to handle -fstack-usage support
15997 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15999 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
16000 interaction for new Power8 flags and VSX.
16002 2013-07-18 Sriraman Tallam <tmsriram@google.com>
16004 PR middle-end/57698
16005 * tree-inline.c (expand_call_inline): Emit errors during
16006 early_inlining only if optimization is not turned on.
16008 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16012 * passes.c (init_optimization_passes): Move the construction of
16013 the pass hierarchy into a new passes.def file.
16015 * Makefile.in (passes.o): Add dependency on passes.def.
16017 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16019 * passes.c (init_optimization_passes): Introduce macros for
16020 constructing the tree of passes (INSERT_PASSES_AFTER,
16021 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
16022 TERMINATE_PASS_LIST).
16024 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
16025 Wei Mi <wmi@google.com>
16027 PR rtl-optimization/57878
16028 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
16030 (reload_pseudo_compare_func): Check nregs first for reload
16033 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16035 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
16037 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
16039 * read-rtl.c (validate_const_int): Once an invalid character is
16040 seen, quit the loop.
16042 * gengtype.c (write_roots): Similarly once we find the "deletable"
16043 or "if_marked" option.
16045 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
16047 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
16048 "xtab" and "sat". Move value "clz" from here to ...
16049 (attriubte "type"): ... here.
16050 (satsi_<SAT:code>): Delete "insn" attribute.
16051 (satsi_<SAT:code>_shift): Likewise.
16052 (arm_zero_extendqisi2addsi): Likewise.
16053 (arm_extendqisi2addsi): Likewise.
16054 (clzsi2): Update for attribute changes.
16055 (rbitsi2): Likewise.
16056 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
16058 (arm_usatsihi): Likewise.
16059 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
16061 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
16063 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
16064 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
16065 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
16066 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
16067 in alphabetical order.
16068 (attribute "core_cycles"): Update for attribute changes.
16069 (arm_addsi3): Likewise.
16070 (addsi3_compare0): Likewise.
16071 (addsi3_compare0_scratch): Likewise.
16072 (addsi3_compare_op1): Likewise.
16073 (addsi3_compare_op2): Likewise.
16074 (compare_addsi2_op0): Likewise.
16075 (compare_addsi2_op1): Likewise.
16076 (addsi3_carryin_shift_<optab>): Likewise.
16077 (subsi3_carryin_shift): Likewise.
16078 (rsbsi3_carryin_shift): Likewise.
16079 (arm_subsi3_insn): Likewise.
16080 (subsi3_compare0): Likewise.
16081 (subsi3_compare): Likewise.
16082 (arm_andsi3_insn): Likewise.
16083 (thumb1_andsi3_insn): Likewise.
16084 (andsi3_compare0): Likewise.
16085 (andsi3_compare0_scratch): Likewise.
16086 (zeroextractsi_compare0_scratch
16087 (andsi_not_shiftsi_si): Likewise.
16088 (iorsi3_insn): Likewise.
16089 (iorsi3_compare0): Likewise.
16090 (iorsi3_compare0_scratch): Likewise.
16091 (arm_xorsi3): Likewise.
16092 (thumb1_xorsi3_insn): Likewise.
16093 (xorsi3_compare0): Likewise.
16094 (xorsi3_compare0_scratch): Likewise.
16095 (satsi_<SAT:code>_shift): Likewise.
16097 (arm_shiftsi3): Likewise.
16098 (shiftsi3_compare0): Likewise.
16099 (not_shiftsi): Likewise.
16100 (not_shiftsi_compare0): Likewise.
16101 (not_shiftsi_compare0_scratch): Likewise.
16102 (arm_one_cmplsi2): Likewise.
16103 (thumb_one_complsi2): Likewise.
16104 (notsi_compare0): Likewise.
16105 (notsi_compare0_scratch): Likewise.
16106 (thumb1_zero_extendhisi2): Likewise.
16107 (arm_zero_extendhisi2): Likewise.
16108 (arm_zero_extendhisi2_v6): Likewise.
16109 (arm_zero_extendhisi2addsi): Likewise.
16110 (thumb1_zero_extendqisi2): Likewise.
16111 (thumb1_zero_extendqisi2_v6): Likewise.
16112 (arm_zero_extendqisi2): Likewise.
16113 (arm_zero_extendqisi2_v6): Likewise.
16114 (arm_zero_extendqisi2addsi): Likewise.
16115 (thumb1_extendhisi2): Likewise.
16116 (arm_extendhisi2): Likewise.
16117 (arm_extendhisi2_v6): Likewise.
16118 (arm_extendqisi): Likewise.
16119 (arm_extendqisi_v6): Likewise.
16120 (arm_extendqisi2addsi): Likewise.
16121 (thumb1_extendqisi2): Likewise.
16122 (thumb1_movdi_insn): Likewise.
16123 (arm_movsi_insn): Likewise.
16124 (movsi_compare0): Likewise.
16125 (movhi_insn_arch4): Likewise.
16126 (movhi_bytes): Likewise.
16127 (arm_movqi_insn): Likewise.
16128 (thumb1_movqi_insn): Likewise.
16129 (arm32_movhf): Likewise.
16130 (thumb1_movhf): Likewise.
16131 (arm_movsf_soft_insn): Likewise.
16132 (thumb1_movsf_insn): Likewise.
16133 (movdf_soft_insn): Likewise.
16134 (thumb_movdf_insn): Likewise.
16135 (arm_cmpsi_insn): Likewise.
16136 (cmpsi_shiftsi): Likewise.
16137 (cmpsi_shiftsi_swp): Likewise.
16138 (arm_cmpsi_negshiftsi_si): Likewise.
16139 (movsicc_insn): Likewise.
16140 (movsfcc_soft_insn): Likewise.
16141 (arith_shiftsi): Likewise.
16142 (arith_shiftsi_compare0
16143 (arith_shiftsi_compare0_scratch
16144 (sub_shiftsi): Likewise.
16145 (sub_shiftsi_compare0
16146 (sub_shiftsi_compare0_scratch
16147 (and_scc): Likewise.
16148 (cond_move): Likewise.
16149 (if_plus_move): Likewise.
16150 (if_move_plus): Likewise.
16151 (if_move_not): Likewise.
16152 (if_not_move): Likewise.
16153 (if_shift_move): Likewise.
16154 (if_move_shift): Likewise.
16155 (if_shift_shift): Likewise.
16156 (if_not_arith): Likewise.
16157 (if_arith_not): Likewise.
16158 (cond_move_not): Likewise.
16159 (thumb1_ashlsi3): Set type attribute.
16160 (thumb1_ashrsi3): Likewise.
16161 (thumb1_lshrsi3): Likewise.
16162 (thumb1_rotrsi3): Likewise.
16163 (shiftsi3_compare0_scratch): Likewise.
16164 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
16165 (neon_mov<mode>): Likewise.
16166 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
16168 (thumb2_movsi_insn): Likewise.
16169 (thumb2_cmpsi_neg_shiftsi): Likewise.
16170 (thumb2_extendqisi_v6): Likewise.
16171 (thumb2_zero_extendhisi2_v6): Likewise.
16172 (thumb2_zero_extendqisi2_v6): Likewise.
16173 (thumb2_shiftsi3_short): Likewise.
16174 (thumb2_addsi3_compare0_scratch): Likewise.
16175 (orsi_not_shiftsi_si): Likewise.
16176 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
16177 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
16179 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
16180 (1020alu_shift_op): Likewise.
16181 (1020alu_shift_reg_op): Likewise.
16182 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
16183 (alu_shift_op): Likewise.
16184 (alu_shift_reg_op): Likewise.
16185 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
16186 (11_alu_shift_op): Likewise.
16187 (11_alu_shift_reg_op): Likewise.
16188 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
16189 (9_alu_shift_reg_op): Likewise.
16190 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
16192 (cortex_a15_alu_shift): Likewise.
16193 (cortex_a15_alu_shift_reg): Likewise.
16194 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
16196 (cortex_a5_alu_shift): Likewise.
16197 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
16199 (cortex_a53_alu_shift): Likewise.
16200 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
16202 (cortex_a7_alu_reg): Likewise.
16203 (cortex_a7_alu_shift): Likewise.
16204 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
16206 (cortex_a8_alu_shift): Likewise.
16207 (cortex_a8_alu_shift_reg): Likewise.
16208 (cortex_a8_mov): Likewise.
16209 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
16210 (cortex_a9_dp_shift): Likewise.
16211 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
16213 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
16215 (cortex_r4_mov): Likewise.
16216 (cortex_r4_alu_shift): Likewise.
16217 (cortex_r4_alu_shift_reg): Likewise.
16218 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
16219 (526_alu_shift_op): Likewise.
16220 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
16221 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
16222 (626te_alu_shift_op): Likewise.
16223 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
16224 (726te_alu_op): Likewise.
16225 (726te_alu_shift_op): Likewise.
16226 (726te_alu_shift_reg_op): Likewise.
16227 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
16228 (mp626_alu_shift_op): Likewise.
16229 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
16230 (pj4_alu_e1_conds): Likewise.
16231 (pj4_alu): Likewise.
16232 (pj4_alu_conds): Likewise.
16233 (pj4_shift): Likewise.
16234 (pj4_shift_conds): Likewise.
16235 (pj4_alu_shift): Likewise.
16236 (pj4_alu_shift_conds): Likewise.
16237 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
16239 (cortexa7_older_only): Likewise.
16240 (cortexa7_younger): Likewise.
16242 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16244 * ipa-pure-const.c (generate_summary): Rename to...
16245 (pure_const_generate_summary): ... this.
16247 2013-07-17 Iain Sandoe <iain@codesourcery.com>
16249 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
16251 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
16254 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
16257 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16259 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
16260 enabled without -march=zEC12.
16261 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
16264 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16266 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
16267 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
16268 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
16269 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
16270 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
16271 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
16272 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
16274 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
16275 and ISA_HAS_NMADD3_NMSUB3.
16276 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
16277 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
16278 (nmsub4<mode>, nmsub3<mode>): Likewise.
16279 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
16281 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16283 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
16284 TARGET_MIPS5400 checking.
16286 2013-07-16 Jakub Jelinek <jakub@redhat.com>
16287 Peter Bergner <bergner@vnet.ibm.com>
16289 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
16290 registers in the comment.
16291 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
16292 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
16293 rather than FIRST_PSEUDO_REGISTERS.
16295 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
16297 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
16298 enable extra ISA flags with TARGET_HTM.
16300 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16302 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
16305 2013-07-15 Cong Hou <congh@google.com>
16307 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
16308 in compare function for sorting.
16310 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
16312 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
16313 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
16314 * config/rs6000/rs6000.opt: Add -mhtm option.
16315 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
16316 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
16317 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16318 __HTM__ if the HTM instructions are available.
16319 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
16320 htm_spr_reg_operand): New define_predicates.
16321 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
16322 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
16324 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
16325 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
16326 HTM builtin functions.
16327 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
16328 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
16329 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
16330 (rs6000_builtin_mask_calculate): Likewise.
16331 (rs6000_option_override_internal): Likewise.
16332 (bdesc_htm): Add new HTM builtin support.
16333 (htm_spr_num): New function.
16334 (htm_spr_regno): Likewise.
16335 (rs6000_htm_spr_icode): Likewise.
16336 (htm_expand_builtin): Likewise.
16337 (htm_init_builtins): Likewise.
16338 (rs6000_expand_builtin): Add support for HTM builtin functions.
16339 (rs6000_init_builtins): Likewise.
16340 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
16342 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
16343 (TARGET_HTM, MASK_HTM): Define macros.
16344 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
16345 (FIXED_REGISTERS): Likewise.
16346 (CALL_USED_REGISTERS): Likewise.
16347 (CALL_REALLY_USED_REGISTERS): Likewise.
16348 (REG_ALLOC_ORDER): Likewise.
16349 (enum reg_class): Likewise.
16350 (REG_CLASS_NAMES): Likewise.
16351 (REG_CLASS_CONTENTS): Likewise.
16352 (REGISTER_NAMES): Likewise.
16353 (ADDITIONAL_REGISTER_NAMES): Likewise.
16354 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
16355 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
16356 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
16357 * config/rs6000/htm.md: New file.
16358 * config/rs6000/htmintrin.h: New file.
16359 * config/rs6000/htmxlintrin.h: New file.
16361 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
16363 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16364 Define SYMBOL_TINY_GOT, update comment.
16365 * config/aarch64/aarch64.c
16366 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
16367 (aarch64_expand_mov_immediate): Likewise.
16368 (aarch64_print_operand): Likewise.
16369 (aarch64_classify_symbol): Likewise.
16370 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
16371 (ldr_got_tiny): Define.
16373 2013-07-13 Tobias Grosser <tobias@grosser.es>
16375 PR tree-optimization/54094
16376 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
16377 scheduling dimension for the parallelism check from the polyhedral
16378 information in the AST.
16379 * graphite-dependences.c (carries_deps): Do not assume the schedule is
16382 2013-07-13 Jason Merrill <jason@redhat.com>
16384 * print-tree.c (debug_vec_tree): Use debug_raw.
16385 (debug_raw (vec<tree, va_gc> &)): New.
16386 (debug_raw (vec<tree, va_gc> *)): New.
16387 * tree.h: Declare them.
16389 2013-07-13 Bin Cheng <bin.cheng@arm.com>
16391 * ifcvt.c (ifcvt_after_combine): New static variable.
16392 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
16394 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
16395 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
16396 rest_of_handle_if_after_reload): Pass new argument for if_convert.
16398 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
16400 * config/mips/mips.c (mips_expand_call): Remove empty statement.
16402 2013-07-12 Michael Matz <matz@suse.de>
16404 PR middle-end/55771
16405 * convert.c (convert_to_real): Reject non-float inner types.
16407 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
16409 * config/aarch64/aarch64-protos.h
16410 (aarch64_simd_immediate_valid_for_move): Remove.
16411 * config/aarch64/aarch64.c (simd_immediate_info): New member.
16412 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
16414 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
16416 2013-07-11 Steve Ellcey <sellcey@mips.com>
16418 * config/mips/mips.c (mips_conditional_register_usage): Do not
16419 use t[0-7] registers in MIPS16 mode when optimizing for size.
16421 2013-07-11 Sriraman Tallam <tmsriram@google.com>
16423 * config/i386/i386.c (dispatch_function_versions): Fix array
16424 indexing of function_version_info to match actual_versions.
16426 2013-07-11 Teresa Johnson <tejohnson@google.com>
16428 * vec.h (struct va_gc): Move release out-of-line.
16429 (va_gc::release): Call ggc_free on released vec.
16431 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
16433 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
16434 Require GOT register as additional operand in UNSPEC.
16435 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
16436 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
16437 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
16438 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
16439 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
16440 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
16441 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
16443 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
16446 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
16447 name seen by assembler/linker rather if available.
16449 2013-07-11 Andreas Schwab <schwab@suse.de>
16451 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
16453 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
16455 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
16457 2013-07-10 Joseph Myers <joseph@codesourcery.com>
16459 * doc/tm.texi.in: Move hook documentation to ....
16460 * target.def: ... here.
16462 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
16463 text on @hook line.
16464 * doc/tm.texi: Regenerate.
16466 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
16469 * doc/invoke.texi: Document -Wconditionally-supported.
16471 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16474 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
16477 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16480 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
16481 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
16482 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
16483 Remove duplicate devices.
16484 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
16485 * config/avr/t-multilib: Regenerate.
16486 * config/avr/avr-tables.opt: Regenerate.
16487 * doc/avr-mmcu.texi: Regenerate.
16489 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16492 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
16494 2013-07-10 Graham Stott <graham.stott@btinternet.com>
16496 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
16497 the cost of MULT when optimizing for size.
16499 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16501 * config/cr16/cr16-protos.h: Don't include target.h.
16503 2013-07-09 Joseph Myers <joseph@codesourcery.com>
16505 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
16506 adjust register size for TDmode and TFmode for VSX registers.
16508 2013-07-08 Kai Tietz <ktietz@redhat.com>
16511 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
16512 hook_bool_const_tree_true.
16514 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16516 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
16517 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
16518 * config/s390/s390.md: Define FPR*_REGNUM constants.
16519 Fix FPR2_REGNUM constant (18 -> 17).
16520 ("*trunc<BFP:mode><DFP_ALL:mode>2")
16521 ("*trunc<DFP_ALL:mode><BFP:mode>2")
16522 ("trunc<BFP:mode><DFP_ALL:mode>2")
16523 ("trunc<DFP_ALL:mode><BFP:mode>2")
16524 ("*extend<BFP:mode><DFP_ALL:mode>2")
16525 ("*extend<DFP_ALL:mode><BFP:mode>2")
16526 ("extend<BFP:mode><DFP_ALL:mode>2")
16527 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
16530 2013-07-08 Graham Stott <graham.stott@btinternet.com>
16532 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
16534 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16536 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
16537 and cfun_fpr_bit_p to cfun_fpr_save_p.
16538 (s390_frame_area, s390_register_info, s390_frame_info)
16539 (s390_emit_prologue, s390_emit_epilogue)
16540 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
16542 * config/s390/s390.h: Define *_REGNUM macros for floating point
16545 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
16547 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
16549 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
16551 PR rtl-optimization/57786
16552 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
16553 and break out of the loop when it is set to false.
16555 2013-07-08 Jakub Jelinek <jakub@redhat.com>
16558 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
16559 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
16560 (const_int 63)) 0)).
16561 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
16562 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
16563 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
16565 PR rtl-optimization/57829
16566 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
16567 mask bits outside of mode are just sign-extension from mode to HWI.
16569 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
16571 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
16572 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
16573 adjust_address instead of change_address to keep info about alignment.
16574 (emit_strmov): Remove.
16575 (emit_memmov): New function.
16576 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
16577 (expand_movmem_epilogue): Likewise and return updated rtx for
16579 (expand_constant_movmem_prologue): Likewise and return updated rtx for
16580 destination and source.
16581 (decide_alignment): Refactor, handle vector_loop.
16582 (ix86_expand_movmem): Likewise.
16583 (ix86_expand_setmem): Likewise.
16584 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
16586 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
16588 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
16589 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
16591 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16593 * config/i386/sse.md (sse_movlhps): Change alternative 3
16594 of operand 2 to "m".
16596 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16599 * config/i386/sse.md (iptr): New mode attribute.
16600 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
16601 (<sse>_vm<plusminus_insn><mode>3): Ditto.
16602 (<sse>_vmmul<mode>3): Ditto.
16603 (<sse>_vmdiv<mode>3): Ditto.
16604 (sse_vmrcpv4sf2): Ditto.
16605 (<sse>_vmsqrt<mode>2): Ditto.
16606 (sse_vmrsqrtv4sf2): Ditto.
16607 (<sse>_vm<code><mode>3): Ditto.
16608 (avx_vmcmp<mode>3): Ditto.
16609 (<sse>_vmmaskcmp<mode>3): Ditto.
16610 (<sse>_comi): Ditto.
16611 (<sse>_ucomi): Ditto.
16612 (*xop_vmfrcz_<mode>): Ditto.
16613 (*fmai_fmadd_<mode>): Ditto.
16614 (*fmai_fmsub_<mode>): Ditto.
16615 (*fmai_fnmadd_<mode>): Ditto.
16616 (*fmai_fnmsub_<mode>): Ditto.
16617 (*fma4i_vmfmadd_<mode>): Ditto.
16618 (*fma4i_vmfmsub_<mode>): Ditto.
16619 (*fma4i_vmfnmadd_<mode>): Ditto.
16620 (*fma4i_vmfnmsub_<mode>): Ditto.
16621 (*xop_vmfrcz_<mode>): Ditto.
16622 (sse_cvtps2pi): Ditto.
16623 (sse_cvttps2pi): Ditto.
16624 (sse_cvtss2si): Ditto.
16625 (sse_cvtss2si_2): Ditto.
16626 (sse_cvtss2siq_2): Ditto.
16627 (sse_cvttss2si): Ditto.
16628 (sse_cvttss2siq): Ditto.
16629 (sse_cvtsd2si): Ditto.
16630 (sse_cvtsd2si_2): Ditto.
16631 (sse_cvtsd2siq_2): Ditto.
16632 (sse_cvttsd2si): Ditto.
16633 (sse_cvttsd2siq): Ditto.
16634 (sse_cvtsd2ss): Ditto.
16635 (sse_cvtss2sd): Ditto.
16636 (avx2_pbroadcast<mode>): Ditto.
16637 (avx2_pbroadcast<mode>_1): Ditto.
16638 (*avx_vperm_broadcast_v4sf): Ditto.
16640 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
16641 (sse_movlhps): Ditto.
16642 (sse_storehps): Ditto.
16643 (sse_loadhps): Ditto.
16644 (sse_storelps): Ditto.
16645 (sse_loadlps): Ditto.
16646 (*vec_concatv4sf): Ditto.
16647 (*vec_interleave_highv2df): Ditto.
16648 (*vec_interleave_lowv2df): Ditto.
16649 (*vec_extractv2df_1_sse): Ditto.
16650 (*vec_extractv2df_0_sse): Ditto.
16651 (sse2_storelpd): Ditto.
16652 (sse2_loadlpd): Ditto.
16653 (sse2_movsd): Ditto.
16654 (*vec_concatv4si): Ditto.
16655 (vec_concatv2di): Ditto.
16657 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
16658 for Intel asm dialect.
16659 (mmx_punpcklwd): Ditto.
16660 (mmx_punpckldq): Ditto.
16662 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
16663 for intel assembler dialect.
16665 2013-07-06 Jakub Jelinek <jakub@redhat.com>
16668 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
16669 for BUILT_IN_C{LZ,LRSB}*.
16670 * tree.h (CASE_INT_FN): Add FN##IMAX case.
16671 * tree-vrp.c (extract_range_basic): Handle
16672 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
16673 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
16674 fall thru to code calling set_value*.
16675 * builtins.c (expand_builtin): Remove *IMAX cases.
16676 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
16677 if width is bigger than 2*HWI.
16679 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
16681 PR rtl-optimization/55342
16682 * lra-int.h (lra_subreg_reload_pseudos): New.
16683 * lra.c: Add undoing optional reloads to the block diagram.
16684 (lra_subreg_reload_pseudos): New.
16685 (lra_optional_reload_pseudos): Change comments.
16686 (lra): Init and clear lra_subreg_reload_pseudos. Clear
16687 lra_optional_reload_pseudos after undo transformations.
16688 * lra-assigns.c (pseudo_prefix_title): New.
16689 (lra_setup_reg_renumber): Use it.
16690 (spill_for): Ditto. Check subreg reload pseudos too.
16691 (assign_by_spills): Consider subreg reload pseudos too.
16692 * lra-constraints.c (simplify_operand_subreg): Use
16693 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
16694 (curr_insn_transform): Recognize and do optional reloads.
16695 (undo_optional_reloads): New.
16696 (lra_undo_inheritance): Call undo_optional_reloads.
16698 2013-07-05 Thomas Quinot <quinot@adacore.com>
16700 * tree-complex.c (expand_complex_operations_1): Fix typo.
16702 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
16704 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
16705 (tune_params): New member 'const vec_costs'.
16706 * config/aarch64/aarch64.c (generic_vector_cost): New.
16707 (generic_tunings): New member 'generic_vector_cost'.
16708 (aarch64_builtin_vectorization_cost): New.
16709 (aarch64_add_stmt_cost): New.
16710 (TARGET_VECTORIZE_ADD_STMT_COST): New.
16711 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
16713 2013-07-03 Jakub Jelinek <jakub@redhat.com>
16716 * config/i386/predicates.md (vsib_address_operand): Disallow
16717 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
16719 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
16721 PR middle-end/55030
16722 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
16723 expand_builtin_setjmp_receiver.
16724 (expand_label): Adjust, call expand_builtin_setjmp_receiver
16725 with NULL for the label parameter.
16726 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
16727 the frame-pointer. Adjust comments.
16728 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
16729 only if LABEL is non-NULL.
16731 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
16733 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
16734 (ARM_ABI_AAPCS64): Ditto.
16736 (ARM_DEFAULT_ABI): Ditto.
16738 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
16740 * config/aarch64/aarch64-builtins.c
16741 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
16742 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
16744 * config/aarch64/aarch64-simd.md
16745 (aarch64_ld1<VALL:mode>): New.
16746 (aarch64_st1<VALL:mode>): Likewise.
16747 * config/aarch64/arm_neon.h
16748 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
16750 2013-07-02 Sriraman Tallam <tmsriram@google.com>
16752 * config/i386/i386.c (gate_insert_vzeroupper): Check if
16754 (ix86_option_override_internal):Turn on all -mavx target flags by
16755 default as they are dependent on AVX anyway.
16757 2013-07-02 Cary Coutant <ccoutant@google.com>
16759 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
16760 deterministic hash.
16761 (loc_checksum_ordered): Likewise.
16762 (hash_loc_operands): Remove inline keyword.
16764 2013-07-02 Jakub Jelinek <jakub@redhat.com>
16766 PR tree-optimization/57741
16767 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
16768 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
16769 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
16770 Allow REAL_CST step_exprs if flag_associative_math.
16771 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
16773 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16775 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
16777 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16779 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
16781 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16783 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
16785 (iorsi3_insn): Likewise.
16786 (arm_xorsi3): Likewise.
16788 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
16790 * arm.md (attribute "wtype"): Delete. Move attribute values from here
16792 (attribute "type"): ... here, and prefix with "wmmx_".
16793 (attribute "core_cycles"): Update for attribute changes.
16794 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
16795 (tbcstv4hi): Likewise.
16796 (tbcstv2si): Likewise.
16797 (iwmmxt_iordi3): Likewise.
16798 (iwmmxt_xordi3): Likewise.
16799 (iwmmxt_anddi3): Likewise.
16800 (iwmmxt_nanddi3): Likewise.
16801 (iwmmxt_arm_movdi): Likewise.
16802 (iwmmxt_movsi_insn): Likewise.
16803 (mov<mode>_internal): Likewise.
16804 (and<mode>3_iwmmxt): Likewise.
16805 (ior<mode>3_iwmmxt): Likewise.
16806 (xor<mode>3_iwmmxt): Likewise.
16807 (add<mode>3_iwmmxt): Likewise.
16808 (ssaddv8qi3): Likewise.
16809 (ssaddv4hi3): Likewise.
16810 (ssaddv2si3): Likewise.
16811 (usaddv8qi3): Likewise.
16812 (usaddv4hi3): Likewise.
16813 (usaddv2si3): Likewise.
16814 (sub<mode>3_iwmmxt): Likewise.
16815 (sssubv8qi3): Likewise.
16816 (sssubv4hi3): Likewise.
16817 (sssubv2si3): Likewise.
16818 (ussubv8qi3): Likewise.
16819 (ussubv4hi3): Likewise.
16820 (ussubv2si3): Likewise.
16821 (mulv4hi3_iwmmxt): Likewise.
16822 (smulv4hi3_highpart): Likewise.
16823 (umulv4hi3_highpart): Likewise.
16824 (iwmmxt_wmacs): Likewise.
16825 (iwmmxt_wmacsz): Likewise.
16826 (iwmmxt_wmacu): Likewise.
16827 (iwmmxt_wmacuz): Likewise.
16828 (iwmmxt_clrdi): Likewise.
16829 (iwmmxt_clrv8qi): Likewise.
16830 (iwmmxt_clr4hi): Likewise.
16831 (iwmmxt_clr2si): Likewise.
16832 (iwmmxt_uavgrndv8qi3): Likewise.
16833 (iwmmxt_uavgrndv4hi3): Likewise.
16834 (iwmmxt_uavgv8qi3): Likewise.
16835 (iwmmxt_uavgv4hi3): Likewise.
16836 (iwmmxt_tinsrb): Likewise.
16837 (iwmmxt_tinsrh): Likewise.
16838 (iwmmxt_tinsrw): Likewise.
16839 (iwmmxt_textrmub): Likewise.
16840 (iwmmxt_textrmsb): Likewise.
16841 (iwmmxt_textrmuh): Likewise.
16842 (iwmmxt_textrmsh): Likewise.
16843 (iwmmxt_textrmw): Likewise.
16844 (iwmxxt_wshufh): Likewise.
16845 (eqv8qi3): Likewise.
16846 (eqv4hi3): Likewise.
16847 (eqv2si3): Likewise.
16848 (gtuv8qi3): Likewise.
16849 (gtuv4hi3): Likewise.
16850 (gtuv2si3): Likewise.
16851 (gtv8qi3): Likewise.
16852 (gtv4hi3): Likewise.
16853 (gtv2si3): Likewise.
16854 (smax<mode>3_iwmmxt): Likewise.
16855 (umax<mode>3_iwmmxt): Likewise.
16856 (smin<mode>3_iwmmxt): Likewise.
16857 (umin<mode>3_iwmmxt): Likewise.
16858 (iwmmxt_wpackhss): Likewise.
16859 (iwmmxt_wpackwss): Likewise.
16860 (iwmmxt_wpackdss): Likewise.
16861 (iwmmxt_wpackhus): Likewise.
16862 (iwmmxt_wpackwus): Likewise.
16863 (iwmmxt_wpackdus): Likewise.
16864 (iwmmxt_wunpckihb): Likewise.
16865 (iwmmxt_wunpckihh): Likewise.
16866 (iwmmxt_wunpckihw): Likewise.
16867 (iwmmxt_wunpckilb): Likewise.
16868 (iwmmxt_wunpckilh): Likewise.
16869 (iwmmxt_wunpckilw): Likewise.
16870 (iwmmxt_wunpckehub): Likewise.
16871 (iwmmxt_wunpckehuh): Likewise.
16872 (iwmmxt_wunpckehuw): Likewise.
16873 (iwmmxt_wunpckehsb): Likewise.
16874 (iwmmxt_wunpckehsh): Likewise.
16875 (iwmmxt_wunpckehsw): Likewise.
16876 (iwmmxt_wunpckelub): Likewise.
16877 (iwmmxt_wunpckeluh): Likewise.
16878 (iwmmxt_wunpckeluw): Likewise.
16879 (iwmmxt_wunpckelsb): Likewise.
16880 (iwmmxt_wunpckelsh): Likewise.
16881 (iwmmxt_wunpckelsw): Likewise.
16882 (ror<mode>3): Likewise.
16883 (ashr<mode>3_iwmmxt): Likewise.
16884 (lshr<mode>3_iwmmxt): Likewise.
16885 (ashl<mode>3_iwmmxt): Likewise.
16886 (ror<mode>3_di): Likewise.
16887 (ashr<mode>3_di): Likewise.
16888 (lshr<mode>3_di): Likewise.
16889 (ashl<mode>3_di): Likewise.
16890 (iwmmxt_wmadds): Likewise.
16891 (iwmmxt_wmaddu): Likewise.
16892 (iwmmxt_tmia): Likewise.
16893 (iwmmxt_tmiaph): Likewise.
16894 (iwmmxt_tmiabb): Likewise.
16895 (iwmmxt_tmiatb): Likewise.
16896 (iwmmxt_tmiabt): Likewise.
16897 (iwmmxt_tmiatt): Likewise.
16898 (iwmmxt_tmovmskb): Likewise.
16899 (iwmmxt_tmovmskh): Likewise.
16900 (iwmmxt_tmovmskw): Likewise.
16901 (iwmmxt_waccb): Likewise.
16902 (iwmmxt_wacch): Likewise.
16903 (iwmmxt_waccw): Likewise.
16904 (iwmmxt_waligni): Likewise.
16905 (iwmmxt_walignr): Likewise.
16906 (iwmmxt_walignr0): Likewise.
16907 (iwmmxt_walignr1): Likewise.
16908 (iwmmxt_walignr2): Likewise.
16909 (iwmmxt_walignr3): Likewise.
16910 (iwmmxt_wsadb): Likewise.
16911 (iwmmxt_wsadh): Likewise.
16912 (iwmmxt_wsadbz): Likewise.
16913 (iwmmxt_wsadhz): Likewise.
16914 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16915 (iwmmxt_wabsdiffb): Likewise.
16916 (iwmmxt_wabsdiffh): Likewise.
16917 (iwmmxt_wabsdiffw): Likewise.
16918 (iwmmxt_waddsubhx): Likewise
16919 (iwmmxt_wsubaddhx): Likewise.
16920 (addc<mode>3): Likewise.
16921 (iwmmxt_avg4): Likewise.
16922 (iwmmxt_avg4r): Likewise.
16923 (iwmmxt_wmaddsx): Likewise.
16924 (iwmmxt_wmaddux): Likewise.
16925 (iwmmxt_wmaddsn): Likewise.
16926 (iwmmxt_wmaddun): Likewise.
16927 (iwmmxt_wmulwsm): Likewise.
16928 (iwmmxt_wmulwum): Likewise.
16929 (iwmmxt_wmulsmr): Likewise.
16930 (iwmmxt_wmulumr): Likewise.
16931 (iwmmxt_wmulwsmr): Likewise.
16932 (iwmmxt_wmulwumr): Likewise.
16933 (iwmmxt_wmulwl): Likewise.
16934 (iwmmxt_wqmulm): Likewise.
16935 (iwmmxt_wqmulwm): Likewise.
16936 (iwmmxt_wqmulmr): Likewise.
16937 (iwmmxt_wqmulwmr): Likewise.
16938 (iwmmxt_waddbhusm): Likewise.
16939 (iwmmxt_waddbhusl): Likewise.
16940 (iwmmxt_wqmiabb): Likewise.
16941 (iwmmxt_wqmiabt): Likewise.
16942 (iwmmxt_wqmiatb): Likewise.
16943 (iwmmxt_wqmiatt): Likewise.
16944 (iwmmxt_wqmiabbn): Likewise.
16945 (iwmmxt_wqmiabtn): Likewise.
16946 (iwmmxt_wqmiatbn): Likewise.
16947 (iwmmxt_wqmiattn): Likewise.
16948 (iwmmxt_wmiabb): Likewise.
16949 (iwmmxt_wmiabt): Likewise.
16950 (iwmmxt_wmiatb): Likewise.
16951 (iwmmxt_wmiatt): Likewise.
16952 (iwmmxt_wmiabbn): Likewise.
16953 (iwmmxt_wmiabtn): Likewise.
16954 (iwmmxt_wmiatbn): Likewise.
16955 (iwmmxt_wmiattn): Likewise.
16956 (iwmmxt_wmiawbb): Likewise.
16957 (iwmmxt_wmiawbt): Likewise.
16958 (iwmmxt_wmiawtb): Likewise.
16959 (iwmmxt_wmiawtt): Likewise.
16960 (iwmmxt_wmiawbbn): Likewise.
16961 (iwmmxt_wmiawbtn): Likewise.
16962 (iwmmxt_wmiawtbn): Likewise.
16963 (iwmmxt_wmiawttn): Likewise.
16964 (iwmmxt_wmerge): Likewise.
16965 (iwmmxt_tandc<mode>3): Likewise.
16966 (iwmmxt_torc<mode>3): Likewise.
16967 (iwmmxt_torvsc<mode>3): Likewise.
16968 (iwmmxt_textrc<mode>3): Likewise.
16969 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16970 (wmmxt_pack): Likewise.
16971 (wmmxt_mult_c1): Likewise.
16972 (wmmxt_mult_c2): Likewise.
16973 (wmmxt_alu_c1): Likewise.
16974 (wmmxt_alu_c2): Likewise.
16975 (wmmxt_alu_c3): Likewise.
16976 (wmmxt_transfer_c1): Likewise.
16977 (wmmxt_transfer_c2): Likewise.
16978 (wmmxt_transfer_c3): Likewise.
16979 (marvell_f_iwmmxt_wstr): Likewise.
16980 (marvell_f_iwmmxt_wldr): Likewise.
16982 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16984 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16986 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16989 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16990 * lra-constraints.c (need_for_split_p): Check call used hard regs
16991 living through calls.
16993 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16994 call used regs for call insn.
16996 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16999 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
17000 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
17002 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
17004 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
17006 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
17008 * lra-constraints.c (need_for_split_p): Check call used hard regs
17009 living through calls.
17011 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
17014 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
17015 to tie with any other modes. Eliminate Altivec vector mode tests,
17016 since these are a subset of ALTIVEC or VSX vector modes. Simplify
17017 code, to return 0 if testing MODE2 for a condition, if we've
17018 already tested MODE1 for the same condition.
17020 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
17022 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
17025 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
17027 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
17028 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
17030 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
17032 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
17034 (aarch64_symbolic_constant_p): Remove.
17035 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
17036 static. Fix line length and white space.
17037 (aarch64_symbolic_constant_p): Remove.
17038 * config/aarch64/predicates.md (aarch64_valid_symref):
17039 Use aarch64_classify_symbol_expression.
17041 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17043 * config/arm/constraints.md (Ts): New constraint.
17044 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
17046 (compare_scc): Use "Ts" constraint for operand 0.
17047 (ior_scc_scc): Likewise.
17048 (and_scc_scc): Likewise.
17049 (and_scc_scc_nodom): Likewise.
17050 (ior_scc_scc_cmp): Likewise for operand 7.
17051 (and_scc_scc_cmp): Likewise.
17052 * config/arm/thumb2.md (thumb2_movsi_insn):
17053 Add alternatives for 16-bit encodings.
17054 (thumb2_movhi_insn): Likewise.
17055 (thumb2_movsicc_insn): Likewise.
17056 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
17057 (thumb2_negscc): Use "Ts" constraint.
17058 Move mvn instruction outside cond_exec block.
17059 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
17060 for 16-bit encodings.
17062 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17064 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
17066 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
17067 (mulsi3subsi): Likewise.
17068 (mulsidi3adddi): Likewise.
17069 (mulsidi3_v6): Likewise.
17070 (umulsidi3_v6): Likewise.
17071 (umulsidi3adddi_v6): Likewise.
17072 (smulsi3_highpart_v6): Likewise.
17073 (umulsi3_highpart_v6): Likewise.
17074 (mulhisi3tb): Likewise.
17075 (mulhisi3bt): Likewise.
17076 (mulhisi3tt): Likewise.
17077 (maddhisi4): Likewise.
17078 (maddhisi4tb): Likewise.
17079 (maddhisi4tt): Likewise.
17080 (maddhidi4): Likewise.
17081 (maddhidi4tb): Likewise.
17082 (maddhidi4tt): Likewise.
17083 (zeroextractsi_compare0_scratch): Likewise.
17084 (insv_zero): Likewise.
17085 (insv_t2): Likewise.
17086 (anddi_notzesidi_di): Likewise.
17087 (anddi_notsesidi_di): Likewise.
17088 (andsi_notsi_si): Likewise.
17089 (iordi_zesidi_di): Likewise.
17090 (xordi_zesidi_di): Likewise.
17091 (andsi_iorsi3_notsi): Likewise.
17092 (smax_0): Likewise.
17093 (smax_m1): Likewise.
17094 (smin_0): Likewise.
17095 (not_shiftsi): Likewise.
17096 (unaligned_loadsi): Likewise.
17097 (unaligned_loadhis): Likewise.
17098 (unaligned_loadhiu): Likewise.
17099 (unaligned_storesi): Likewise.
17100 (unaligned_storehi): Likewise.
17101 (extv_reg): Likewise.
17102 (extzv_t2): Likewise.
17103 (divsi3): Likewise.
17104 (udivsi3): Likewise.
17105 (arm_zero_extendhisi2addsi): Likewise.
17106 (arm_zero_extendqisi2addsi): Likewise.
17107 (compareqi_eq0): Likewise.
17108 (arm_extendhisi2_v6): Likewise.
17109 (arm_extendqisi2addsi): Likewise.
17110 (arm_movt): Likewise.
17111 (thumb2_ldrd): Likewise.
17112 (thumb2_ldrd_base): Likewise.
17113 (thumb2_ldrd_base_neg): Likewise.
17114 (thumb2_strd): Likewise.
17115 (thumb2_strd_base): Likewise.
17116 (thumb2_strd_base_neg): Likewise.
17117 (arm_negsi2): Add alternative for 16-bit encoding.
17118 (arm_one_cmplsi2): Likewise.
17120 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17122 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
17123 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
17124 (movdfcc): Likewise.
17125 * config/arm/vfp.md (*thumb2_movsf_vfp):
17126 Disable predication for arm_restrict_it.
17127 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
17128 (*thumb2_movdfcc_vfp): Likewise.
17129 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
17130 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
17131 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
17132 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
17133 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
17134 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
17135 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
17136 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
17137 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
17138 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
17139 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
17140 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
17141 Disable predication for arm_restrict_it.
17143 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
17145 * config/i386/bmiintrin.h (_bextr_u32): New.
17146 (_bextr_u64): Ditto.
17148 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
17150 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
17151 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
17152 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
17153 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
17154 * config/mips/n32-elf.h: ...this new file.
17156 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
17159 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
17160 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
17162 2013-06-27 Catherine Moore <clm@codesourcery.com>
17164 * config/mips/mips-tables.opt: Regenerate.
17165 * config/mips/mips-cpus.def: Add m14ke and m14kec.
17166 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
17167 * doc/invoke.texi: Add -m14kc.
17169 2013-06-27 Jakub Jelinek <jakub@redhat.com>
17172 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
17173 constraints of operand 1 and 2.
17176 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
17177 to match RTL canonicalization. Swap predicates and
17178 constraints of operand 1 and 2.
17180 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
17182 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
17183 Process OP_INOUT regs for splitting too.
17185 2013-06-27 Jakub Jelinek <jakub@redhat.com>
17187 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
17188 decl before the loop, initialize to NULL.
17189 (vectorizable_load): Initialize ptr_incr to NULL.
17191 2013-06-27 Martin Jambor <mjambor@suse.cz>
17194 * ipa-ref.h (ipa_maybe_record_reference): Declare.
17195 * ipa-ref.c (ipa_maybe_record_reference): New function.
17196 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
17197 * ipa-cp.c (create_specialized_node): Record potential references from
17199 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
17201 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
17203 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
17204 parameter 'mode' of type 'enum machine_mode mode'; change to pass
17205 'mode' to force_reg.
17206 (aarch64_add_offset): Update calls to aarch64_force_temporary.
17207 (aarch64_expand_mov_immediate): Likewise.
17209 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
17211 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
17212 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
17214 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17216 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
17217 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
17218 (struct machine_function): Add tbegin_p.
17219 (s390_canonicalize_comparison): Fold CC mode compares to
17220 conditional jump if possible.
17221 (s390_emit_jump): Return the emitted jump.
17222 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
17223 Handle CCRAWmode compares.
17224 (s390_option_override): Default to -mhtm if available.
17225 (s390_reg_clobbered_rtx): Handle floating point regs as well.
17226 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
17227 FPRs instead of df_regs_ever_live_p.
17228 (s390_optimize_nonescaping_tx): New function.
17229 (s390_init_frame_layout): Extend clobbered_regs array to cover
17231 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
17232 (s390_expand_tbegin): New function.
17233 (enum s390_builtin): New enum definition.
17234 (code_for_builtin): New array definition.
17235 (s390_init_builtins): New function.
17236 (s390_expand_builtin): New function.
17237 (TARGET_INIT_BUILTINS): Define.
17238 (TARGET_EXPAND_BUILTIN): Define.
17239 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
17240 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
17241 (s390_alc_comparison): Likewise.
17242 * config/s390/s390-modes.def: Add CCRAWmode.
17243 * config/s390/s390.h (processor_flags): Add PF_TX.
17244 (TARGET_CPU_HTM): Define macro.
17245 (TARGET_HTM): Define macro.
17246 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
17247 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
17248 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
17249 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
17250 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
17251 (TBEGIN_MASK, TBEGINC_MASK): New constants.
17252 ("*cc_to_int"): Move up.
17253 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
17254 constants other than 0.
17255 ("*ccraw_to_int"): New insn and splitter definition.
17256 ("tbegin", "tbegin_nofloat", "tbegin_retry")
17257 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
17258 ("tx_assist"): New expander.
17259 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
17260 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
17261 * config/s390/s390.opt: Add -mhtm option.
17262 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
17263 * config/s390/htmxlintrin.h: New file.
17264 * config/s390/htmintrin.h: New file.
17265 * config/s390/s390intrin.h: New file.
17266 * doc/extend.texi: Document htm builtins.
17267 * config.gcc: Add the new header files to extra_headers.
17269 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
17271 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
17272 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
17274 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
17275 Pat Haugen <pthaugen@us.ibm.com>
17276 Peter Bergner <bergner@vnet.ibm.com>
17278 * config/rs6000/power8.md: New.
17279 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
17280 setting for power8 entry.
17281 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
17282 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
17283 test for Power4/Power5 only.
17284 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
17286 (force_new_group): Adjust comment.
17287 * config/rs6000/rs6000.md: Include power8.md.
17289 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
17291 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
17292 * config/arm/arm-protos.h (arm_max_conditional_execute): New
17294 (tune_params): Update comment.
17295 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
17296 (arm_max_conditional_execute): New function.
17297 (thumb2_final_prescan_insn): Use max_insn_skipped and
17298 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
17300 2013-06-25 Jakub Jelinek <jakub@redhat.com>
17302 PR tree-optimization/57705
17303 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
17304 SSA_NAME step, provided that it is not defined inside the loop.
17305 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
17306 (get_initial_def_for_induction): Handle SSA_NAME IV step.
17308 2013-06-25 Martin Jambor <mjambor@suse.cz>
17310 PR middle-end/57670
17311 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
17312 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
17314 (ipa_note_param_call): Initialize member_ptr flag.
17315 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
17316 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
17317 (ipa_write_indirect_edge_info): Stream member_ptr flag.
17318 (ipa_read_indirect_edge_info): Likewise.
17320 2013-06-25 Richard Biener <rguenther@suse.de>
17322 PR middle-end/56977
17323 * passes.c (init_optimization_passes): Move pass_fold_builtins
17324 and pass_dce earlier with -Og.
17326 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
17328 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
17329 <BIT_FIELD_REF>: Remove trailing TAB.
17330 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
17333 2013-06-24 Martin Jambor <mjambor@suse.cz>
17335 PR tree-optimization/57358
17336 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
17337 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
17338 (ipa_analyze_params_uses): Generate pessimistic info when true.
17340 2013-06-24 Martin Jambor <mjambor@suse.cz>
17342 PR tree-optimization/57539
17343 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
17344 global.inlined_to of the new node to it. All callers changed.
17345 * ipa-inline-transform.c (clone_inlined_nodes): New variable
17346 inlining_into, pass it to cgraph_clone_node.
17347 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
17348 ipa_free_edge_args_substructures.
17349 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
17350 rdesc linked list. Do not assert rdesc edges have inlined caller.
17351 Assert we have found an rdesc in the rdesc list.
17353 2013-06-24 Richard Biener <rguenther@suse.de>
17355 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
17356 (pointer_set_lookup): Declare.
17357 (class pointer_map): New template class implementing a
17358 generic pointer to T map.
17359 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
17360 pointer_map<T>::contains, pointer_map<T>::insert,
17361 pointer_map<T>::traverse): New functions.
17362 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
17363 (pointer_set_lookup): New function.
17364 (pointer_set_contains): Use pointer_set_lookup.
17365 (pointer_set_insert): Likewise.
17366 (insert_aux): Remove.
17367 (struct pointer_map_t): Embed a pointer_set_t.
17368 (pointer_map_create): Adjust.
17369 (pointer_map_destroy): Likewise.
17370 (pointer_map_contains): Likewise.
17371 (pointer_map_insert): Likewise.
17372 (pointer_map_traverse): Likewise.
17373 * tree-streamer.h (struct streamer_tree_cache_d): Use a
17374 pointer_map<unsigned> instead of a pointer_map_t.
17375 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
17376 (streamer_tree_cache_lookup): Likewise.
17377 (streamer_tree_cache_create): Likewise.
17378 (streamer_tree_cache_delete): Likewise.
17379 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
17380 pointer_map<unsigned> instead of a pointer_map_t.
17381 (lto_init_tree_ref_encoder): Adjust.
17382 (lto_destroy_tree_ref_encoder): Likewise.
17383 * lto-section-out.c (lto_output_decl_index): Likewise.
17384 (lto_record_function_out_decl_state): Likewise.
17385 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
17387 2013-06-24 Richard Biener <rguenther@suse.de>
17389 PR tree-optimization/57488
17390 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
17392 2013-06-24 Alan Modra <amodra@gmail.com>
17394 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
17395 (gen_easy_altivec_constant): Likewise.
17396 * config/rs6000/predicates.md (easy_vector_constant_add_self,
17397 easy_vector_constant_msb): Likewise.
17399 2013-06-23 Jakub Jelinek <jakub@redhat.com>
17402 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
17403 add missing return true.
17405 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
17408 * config/sh/predicates.md (general_extend_operand): Invoke
17409 general_movsrc_operand for memory operands.
17410 (general_movsrc_operand): Allow reg+reg addressing, do not use
17411 general_operand for memory operands.
17413 2013-06-23 Sriraman Tallam <tmsriram@google.com>
17415 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
17416 when current target options does not apply.
17417 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
17418 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
17419 * config/i386/bmiintrin.h: Pass appropriate target
17420 attributes to header.
17421 * config/i386/mmintrin.h: Ditto.
17422 * config/i386/nmmintrin.h: Ditto.
17423 * config/i386/avx2intrin.h: Ditto.
17424 * config/i386/fxsrintrin.h: Ditto.
17425 * config/i386/tbmintrin.h: Ditto.
17426 * config/i386/xsaveintrin.h: Ditto.
17427 * config/i386/f16cintrin.h: Ditto.
17428 * config/i386/xtestintrin.h: Ditto.
17429 * config/i386/xsaveoptintrin.h: Ditto.
17430 * config/i386/bmi2intrin.h: Ditto.
17431 * config/i386/lzcntintrin.h: Ditto.
17432 * config/i386/smmintrin.h: Ditto.
17433 * config/i386/wmmintrin.h: Ditto.
17434 * config/i386/x86intrin.h: Remove all header include guards.
17435 * config/i386/prfchwintrin.h: Ditto.
17436 * config/i386/pmmintrin.h: Ditto.
17437 * config/i386/tmmintrin.h: Ditto.
17438 * config/i386/xmmintrin.h: Ditto.
17439 * config/i386/popcntintrin.h: Ditto.
17440 * config/i386/rdseedintrin.h: Ditto.
17441 * config/i386/ammintrin.h: Ditto.
17442 * config/i386/emmintrin.h: Ditto.
17443 * config/i386/immintrin.h: Remove all header include guards.
17444 * config/i386/fma4intrin.h: Ditto.
17445 * config/i386/lwpintrin.h: Ditto.
17446 * config/i386/xopintrin.h: Ditto.
17447 * config/i386/ia32intrin.h: Ditto.
17448 * config/i386/avxintrin.h: Ditto.
17449 * config/i386/rtmintrin.h: Ditto.
17450 * config/i386/fmaintrin.h: Ditto.
17451 * config/i386/mm3dnow.h: Ditto.
17453 2013-06-22 Sriraman Tallam <tmsriram@google.com>
17455 * common/config/i386/i386-common.c: Handle LZCNT.
17457 2013-06-22 Andi Kleen <ak@linux.intel.com>
17459 * doc/extend.texi: Use __atomic_store_n instead of
17460 __atomic_store in HLE example.
17462 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
17464 * config/sh/sh.c: Remove <cstdlib> workaround.
17466 2013-06-21 Andi Kleen <ak@linux.intel.com>
17468 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
17470 2013-06-21 Andi Kleen <ak@linux.intel.com>
17472 * doc/extend.texi: Document that __atomic_clear and
17473 __atomic_test_and_set should only be used with bool.
17475 2013-06-20 Jan Hubicka <jh@suse.cz>
17477 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
17478 types_same_for_odr.
17479 * tree.c (decls_same_for_odr): New function.
17480 (same_for_edr): New function.
17481 (types_same_for_odr): New function.
17482 (get_binfo_at_offset): Use it.
17483 * tree.h (types_same_for_odr): Declare.
17485 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
17486 Jason Merrill <jason@redhat.com>
17488 * system.h: Include <cstdlib> as well as <stdlib.h>.
17490 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
17493 * config/i386/i386.c (construct_container): Report error if
17494 long double is used with disabled x87 float returns.
17496 2013-06-20 Jan Hubicka <jh@suse.cz>
17498 * lto-cgraph.c (input_symtab): Do not set cgraph state.
17500 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
17502 PR rtl-optimization/57425
17503 PR rtl-optimization/57569
17504 * alias.c (write_dependence_p): Remove parameters mem_mode and
17505 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
17506 Changed all callers.
17507 (canon_anti_dependence): Get comments and semantics in sync.
17508 Add parameter mem_canonicalized. Changed all callers.
17509 * rtl.h (canon_anti_dependence): Update prototype.
17511 2013-06-20 Richard Biener <rguenther@suse.de>
17513 * data-streamer-in.c (streamer_read_uhwi): Optimize single
17514 byte case, inline streamer_read_uchar and defer section
17517 2013-06-20 Richard Biener <rguenther@suse.de>
17519 PR tree-optimization/57584
17520 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
17521 SSA names into the expanded expression that take part in
17522 abnormal coalescing.
17524 2013-06-19 Sharad Singhai <singhai@google.com>
17526 * gcov.c (print_usage): Handle new option.
17527 (process_args): Ditto.
17528 (get_gcov_intermediate_filename): New function.
17529 (output_intermediate_file): New function.
17530 (output_gcov_file): New function
17531 (generate_results): Handle new option.
17532 (release_function): Relase demangled name.
17533 (read_graph_file): Handle demangled name.
17534 (output_lines): Ditto.
17535 * doc/gcov.texi: Document gcov intermediate format.
17537 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
17540 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
17541 (lra_emit_add): Use the functions. Add comment about Y as an
17544 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
17547 * collect2.c (collect_atexit): New.
17548 (collect_exit): Delete.
17549 (main): Register collect_atexit with atexit.
17550 (collect_wait): Change collect_exit to exit.
17552 * collect2.h (collect_exit): Delete.
17553 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
17555 2013-06-19 Wei Mi <wmi@google.com>
17557 PR rtl-optimization/57518
17558 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
17559 if regno is used in paradoxical subreg.
17560 (update_equiv_regs): Check pdx_subregs[regno] before
17561 set a reg to be equivalent with a mem.
17563 2013-06-19 Matthias Klose <doko@ubuntu.com>
17566 * file-find.h (find_a_file): Add a mode parameter.
17567 * file-find.c (find_a_file): Likewise.
17568 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
17569 with X_OK for the executables.
17570 * collect2.c (main): Call find_a_file with X_OK.
17572 2013-06-19 Steve Ellcey <sellcey@mips.com>
17575 * config/mips/mips.md (casesi_internal_mips16_<mode>):
17576 Use NEXT_INSN instead of next_real_insn.
17578 2013-06-19 Jan Hubicka <jh@suse.cz>
17580 * cgraph.h (const_value_known_p): Replace by ...
17581 (ctor_for_folding): .. this one.
17582 * cgraphunit.c (process_function_and_variable_attributes): Use it.
17583 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
17584 * expr.c (expand_expr_real_1): Likewise.
17585 (string_constant): Likewise.
17586 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
17587 * ipa.c (process_references): Likewise.
17588 (symtab_remove_unreachable_nodes): Likewise.
17589 * ipa-inline-analysis.c (param_change_prob): Likewise.
17590 * gimple-fold.c (canonicalize_constructor_val): Likewise.
17591 (get_base_constructor): Likwise.
17592 * varpool.c (varpool_remove_node): Likewise.
17593 (varpool_remove_initializer): LIkewise.
17594 (dump_varpool_node): LIkwise.
17595 (const_value_known_p): Rewrite to ...
17596 (ctor_for_folding): ... this one.
17598 2013-06-19 Jakub Jelinek <jakub@redhat.com>
17601 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
17602 PERSONALITY in $PATH derived prefixes.
17604 2013-06-19 Jeff Law <law@redhat.com>
17606 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
17609 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
17610 (simplify_bitwise_binary): Use it to simpify certain binary ops on
17613 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
17615 * config/arm/vfp.md: Move VFP instruction classification documentation
17617 * config/arm/arm.md: ... here. Update instruction classification
17620 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
17622 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
17623 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
17624 pattern. Use more efficient sequences on ARMv5 and Thumb2.
17626 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
17629 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17630 with NEXT_INSN. Use tablejump_p to check for jump table data
17633 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
17636 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
17637 that now in C++ the value is correct per the C++ standards.
17639 2013-06-19 Richard Biener <rguenther@suse.de>
17641 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
17642 for global context.
17644 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17647 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17650 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17651 with next_active_insn.
17653 2013-06-18 Sriraman Tallam <tmsriram@google.com>
17655 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
17656 functions are inlined during failures to flag an error.
17657 * tree-inline.c (expand_call_inline): Allow the error to be flagged
17658 in early inline pass.
17660 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
17662 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
17665 2013-06-18 Julian Brown <julian@codesourcery.com>
17667 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
17668 Permit virtual register pre-reload if !strict.
17669 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
17671 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
17673 * config/arm/neon.md (movmisalign<mode>): Use
17674 neon_perm_struct_or_reg_operand instead of
17675 neon_struct_or_register_operand.
17676 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
17677 neon_permissive_struct_operand instead of neon_struct_operand.
17678 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
17679 neon_vector_mem_operand.
17680 * config/arm/predicates.md (neon_struct_operand): Adjust call to
17681 neon_vector_mem_operand.
17682 (neon_permissive_struct_operand): New.
17683 (neon_struct_or_register_operand): Rename to...
17684 (neon_perm_struct_or_reg_operand): This. Adjust call to
17685 neon_vector_mem_operand.
17687 2013-06-18 Richard Biener <rguenther@suse.de>
17689 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
17690 * lto-streamer.h: Include pointer-set.h.
17691 (struct lto_decl_slot): Remove.
17692 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
17693 Remove next_index entry.
17694 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17695 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17696 (lto_init_tree_ref_encoder): Adjust.
17697 (lto_destroy_tree_ref_encoder): Likewise.
17698 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17699 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17700 (lto_output_decl_index): Adjust.
17701 (lto_new_out_decl_state): Likewise.
17702 (lto_record_function_out_decl_state): Likewise.
17703 * lto-streamer-out.c (copy_function): Likewise.
17705 2013-06-18 Richard Biener <rguenther@suse.de>
17707 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
17708 * cgraphunit.c: Include cfgloop.h.
17709 (init_lowered_empty_function): Initialize the loop tree.
17710 (assemble_thunk): Insert new BBs into loops.
17712 2013-06-18 Richard Biener <rguenther@suse.de>
17714 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
17715 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
17716 the map from cache entry to cache index optional.
17717 (streamer_tree_cache_replace_tree): Adjust accordingly.
17718 (streamer_tree_cache_append): Likewise.
17719 (streamer_tree_cache_delete): Likewise.
17720 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
17721 streamer cache map from cache entry to cache index.
17722 * lto-streamer-out.c (create_output_block): Adjust.
17724 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
17726 * config/arm/arm.md (attribute "insn"): Move multiplication and
17727 division attributes to...
17728 (attribute "type"): ... here. Remove mult.
17729 (attribute "mul32"): New attribute.
17730 (attribute "mul64"): Add umaal.
17731 (*arm_mulsi3): Update attributes.
17732 (*arm_mulsi3_v6): Likewise.
17733 (*thumb_mulsi3): Likewise.
17734 (*thumb_mulsi3_v6): Likewise.
17735 (*mulsi3_compare0): Likewise.
17736 (*mulsi3_compare0_v6): Likewise.
17737 (*mulsi_compare0_scratch): Likewise.
17738 (*mulsi_compare0_scratch_v6): Likewise.
17739 (*mulsi3addsi): Likewise.
17740 (*mulsi3addsi_v6): Likewise.
17741 (*mulsi3addsi_compare0): Likewise.
17742 (*mulsi3addsi_compare0_v6): Likewise.
17743 (*mulsi3addsi_compare0_scratch): Likewise.
17744 (*mulsi3addsi_compare0_scratch_v6): Likewise.
17745 (*mulsi3subsi): Likewise.
17746 (*mulsidi3adddi): Likewise.
17747 (*mulsi3addsi_v6): Likewise.
17748 (*mulsidi3adddi_v6): Likewise.
17749 (*mulsidi3_nov6): Likewise.
17750 (*mulsidi3_v6): Likewise.
17751 (*umulsidi3_nov6): Likewise.
17752 (*umulsidi3_v6): Likewise.
17753 (*umulsidi3adddi): Likewise.
17754 (*umulsidi3adddi_v6): Likewise.
17755 (*smulsi3_highpart_nov6): Likewise.
17756 (*smulsi3_highpart_v6): Likewise.
17757 (*umulsi3_highpart_nov6): Likewise.
17758 (*umulsi3_highpart_v6): Likewise.
17759 (mulhisi3): Likewise.
17760 (*mulhisi3tb): Likewise.
17761 (*mulhisi3bt): Likewise.
17762 (*mulhisi3tt): Likewise.
17763 (maddhisi4): Likewise.
17764 (*maddhisi4tb): Likewise.
17765 (*maddhisi4tt): Likewise.
17766 (maddhidi4): Likewise.
17767 (*maddhidi4tb): Likewise.
17768 (*maddhidi4tt): Likewise.
17769 (divsi3): Likewise.
17770 (udivsi3): Likewise.
17771 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
17772 (thumb2_mulsi_short_compare0): Likewise.
17773 (thumb2_mulsi_short_compare0_scratch): Likewise.
17774 * config/arm/arm1020e.md (1020mult1): Update attribute change.
17775 (1020mult2): Likewise.
17776 (1020mult3): Likewise.
17777 (1020mult4): Likewise.
17778 (1020mult5): Likewise.
17779 (1020mult6): Likewise.
17780 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
17782 (cortex_a15_mult64): Likewise.
17783 (cortex_a15_sdiv): Likewise.
17784 (cortex_a15_udiv): Likewise.
17785 * config/arm/arm1026ejs.md (mult1): Update attribute change.
17791 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
17792 (pj4_ir_div): Likewise.
17793 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
17794 (11_mult2): Likewise.
17795 (11_mult3): Likewise.
17796 (11_mult4): Likewise.
17797 (11_mult5): Likewise.
17798 (11_mult6): Likewise.
17799 (11_mult7): Likewise.
17800 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
17801 (cortex_a8_mla): Likewise.
17802 (cortex_a8_mull): Likewise.
17803 (cortex_a8_smulwy): Likewise.
17804 (cortex_a8_smlald): Likewise.
17805 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
17806 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
17807 (cortex_r4_mul_3): Likewise.
17808 (cortex_r4_mla_4): Likewise.
17809 (cortex_r4_mla_3): Likewise.
17810 (cortex_r4_smlald): Likewise.
17811 (cortex_r4_mull): Likewise.
17812 (cortex_r4_sdiv): Likewise.
17813 (cortex_r4_udiv): Likewise.
17814 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
17815 (cortex_a7_idiv): Likewise.
17816 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
17817 (9_mult2): Likewise.
17818 (9_mult3): Likewise.
17819 (9_mult4): Likewise.
17820 (9_mult5): Likewise.
17821 (9_mult6): Likewise.
17822 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
17823 (cortex_a53_sdiv): Likewise.
17824 (cortex_a53_udiv): Likewise.
17825 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
17826 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
17827 (mp626_mult2): Likewise.
17828 (mp626_mult3): Likewise.
17829 (mp626_mult4): Likewise.
17830 * config/arm/fa526.md (526_mult1): Update attribute change.
17831 (526_mult2): Likewise.
17832 * config/arm/arm-generic.md (mult): Update attribute change.
17833 (mult_ldsched_strongarm): Likewise.
17834 (mult_ldsched): Likewise.
17835 (multi_cycle): Likewise.
17836 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
17837 * config/arm/fa606te.md (606te_mult1): Update attribute change.
17838 (606te_mult2): Likewise.
17839 (606te_mult3): Likewise.
17840 (606te_mult4): Likewise.
17841 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
17842 (cortex_a9_mac16): Likewise.
17843 (cortex_a9_multiply): Likewise.
17844 (cortex_a9_mac): Likewise.
17845 (cortex_a9_multiply_long): Likewise.
17846 * config/arm/fa626te.md (626te_mult1): Update attribute change.
17847 (626te_mult2): Likewise.
17848 (626te_mult3): Likewise.
17849 (626te_mult4): Likewise.
17851 2013-06-18 Richard Biener <rguenther@suse.de>
17854 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17856 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17859 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17860 with next_active_insn.
17862 2013-06-18 Alan Modra <amodra@gmail.com>
17864 * config/rs6000/rs6000.h (enum data_align): New.
17865 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17866 (DATA_ABI_ALIGNMENT): Define.
17867 (CONSTANT_ALIGNMENT): Correct comment.
17868 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17869 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17871 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17873 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17874 ATTRIBUTE_UNUSED marking.
17876 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17878 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17879 alternative and update.
17880 (aarch64_dup_lanedi): Delete.
17881 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17882 * config/aarch64/aarch64-simd-builtins.def: Update.
17884 2013-06-17 Richard Biener <rguenther@suse.de>
17886 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17887 (lto_input_scc): Declare.
17888 (lto_input_tree_1): Likewise.
17889 (struct lto_stats_d): Add num_tree_bodies_output and
17890 num_pickle_refs_output.
17891 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17892 (lto_read_tree_1): Split out from ...
17893 (lto_read_tree): ... this.
17894 (lto_input_scc): New function.
17895 (lto_input_tree_1): Split out from ...
17896 (lto_input_tree): ... this. Handle LTO_tree_scc.
17897 (lto_data_in_create): Create the streamer cache without hashes.
17898 * lto-streamer-out.c (create_output_block): Create the streamer
17899 cache with hashes when not doing WPA.
17900 (lto_write_tree_1): Split out from ...
17901 (lto_write_tree): ... this.
17902 (get_symbol_initial_value): New function.
17903 (lto_output_tree_1): Split out from ...
17904 (lto_output_tree): ... this. Write trees as series of SCCs
17905 using a DFS walk via DFS_write_tree.
17906 (struct sccs, struct scc_entry): New types.
17907 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17908 (DFS_write_tree_body): New function.
17909 (DFS_write_tree): Likewise.
17910 (hash_tree): Likewise.
17911 (scc_entry_compare): Likewise.
17912 (hash_scc): Likewise.
17913 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17914 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17915 TREE_CHAIN as regular reference.
17916 (streamer_read_integer_cst): Remove.
17917 (streamer_get_pickled_tree): Adjust.
17918 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17919 of DECL_EXTERNALs in BLOCK_VARS for now.
17920 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17922 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17923 Add hash value argument and record that if hashes are recorded
17925 (streamer_tree_cache_insert_1): Adjust.
17926 (streamer_tree_cache_insert): Likewise.
17927 (streamer_tree_cache_insert_at): Rename to ...
17928 (streamer_tree_cache_replace_tree): ... this and adjust.
17929 (streamer_tree_cache_append): Adjust.
17930 (record_common_node): Likewise.
17931 (streamer_tree_cache_create): Add argument whether to
17932 record hash values together with trees.
17933 (streamer_tree_cache_delete): Adjust.
17934 * tree-streamer.h (struct streamer_tree_cache_d): Add
17936 (streamer_read_integer_cst): Remove.
17937 (streamer_tree_cache_insert): Adjust.
17938 (streamer_tree_cache_append): Likewise.
17939 (streamer_tree_cache_insert_at): Rename to ...
17940 (streamer_tree_cache_replace_tree): ... this and adjust.
17941 (streamer_tree_cache_create): Add argument whether to record hashes.
17942 (streamer_tree_cache_get): Rename to ...
17943 (streamer_tree_cache_get_tree): ... this.
17944 (streamer_tree_cache_get_hash): New function.
17945 * tree.c (cache_integer_cst): New function.
17946 * tree.h (cache_integer_cst): Declare.
17947 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17948 * lto-symtab.c (lto_varpool_replace_node): Only release
17949 DECL_INITIAL of non-prevailing decls.
17950 * varpool.c (varpool_remove_initializer): Do not release
17951 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17953 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17955 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17956 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17957 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17958 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17959 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17960 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17961 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17962 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17963 instead of TARGET_64BIT.
17964 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17965 Require ISA_HAS_<D>DIV.
17967 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17969 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17970 (mips*-*-linux*): Move default with_llsc setting to where other
17972 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17974 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17975 Likewise. Remove default with_tune setting. Move default float
17976 setting to its own block. Handle with_llsc in the same block as above.
17978 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17980 PR rtl-optimization/57425
17981 PR rtl-optimization/57569
17982 * alias.c (write_dependence_p): Add new parameters mem_mode,
17983 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17984 Changed all callers.
17985 (canon_anti_dependence): New function.
17986 * cse.c (check_dependence): Use canon_anti_dependence.
17987 * cselib.c (cselib_invalidate_mem): Likewise.
17988 * rtl.h (canon_anti_dependence): Declare.
17990 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17992 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17993 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17994 ".set mips3" for 64-bit targets.
17996 2013-06-15 Dehao Chen <dehao@google.com>
17998 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17999 * gimple-low.c (gimple_check_call_matching_types): Likewise.
18000 (gimple_check_call_args): Likewise.
18001 * value-prof.c (check_ic_target): Likewise.
18002 * ipa-inline.c (early_inliner): Likewise.
18003 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
18004 * cgraph.c (cgraph_create_edge_1): Likewise.
18005 (cgraph_make_edge_direct): Likewise.
18007 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
18010 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
18011 rs6000_output_move_128bit to handle emitting quad memory
18012 operations. Set attribute length to 8 bytes.
18014 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
18016 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
18018 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
18019 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
18020 (aarch64_<su>mlsl<mode>): Likewise.
18022 2013-06-14 Mike Stump <mikestump@comcast.net>
18024 * Makefile.in (TARGET_H): Add insn-codes.h.
18026 2013-06-14 Alan Modra <amodra@gmail.com>
18028 PR middle-end/57134
18029 PR middle-end/57586
18030 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
18031 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
18032 bitfield expansion when EXPAND_MEMORY.
18033 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
18035 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
18037 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
18038 test for clearing quad memory on 32-bit later.
18040 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
18042 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
18043 (fold_negate_expr): Likewise.
18044 (fold_real_zero_addition_p): Handle vectors.
18045 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
18047 2013-06-14 Alan Modra <amodra@gmail.com>
18049 * varasm.c (force_const_mem): Revert 2013-06-07 change.
18051 2013-06-13 Jan Hubicka <jh@suse.cz>
18053 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
18054 Local comdats are not externally visible.
18055 * symtab.c (dump_symtab_base): Dump externally visible.
18056 (verify_symtab_base): Verify back links in the symtab hash.
18058 2013-06-13 Bin Cheng <bin.cheng@arm.com>
18060 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
18061 CONVERT_EXPR as equal nodes.
18063 2013-06-13 Bin Cheng <bin.cheng@arm.com>
18065 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
18067 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
18069 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
18070 Generalize to complex and vector.
18071 * tree.c (build_all_ones_cst): New function.
18072 * tree.h (build_all_ones_cst): Declare it.
18074 2013-06-13 Alan Modra <amodra@gmail.com>
18076 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
18077 * config/rs6000/rs6000.md (signbittf2): New insn.
18078 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
18079 (abstf2_internal, cmptf_internal2): Likewise.
18080 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
18082 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
18083 Pat Haugen <pthaugen@us.ibm.com>
18084 Peter Bergner <bergner@vnet.ibm.com>
18086 * config/rs6000/rs6000.c (emit_load_locked): Add support for
18087 power8 byte, half-word, and quad-word atomic instructions.
18088 (emit_store_conditional): Likewise.
18089 (rs6000_expand_atomic_compare_and_swap): Likewise.
18090 (rs6000_expand_atomic_op): Likewise.
18092 * config/rs6000/sync.md (larx): Add new modes for power8.
18094 (AINT): New mode iterator to include TImode as well as normal
18095 integer modes on power8.
18096 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
18097 that VSX registers are not considered. Use AINT mode iterator
18098 instead of INT1 to allow inclusion of quad word atomic operations
18100 (load_locked<mode>): Likewise.
18101 (store_conditional<mode>): Likewise.
18102 (atomic_compare_and_swap<mode>): Likewise.
18103 (atomic_exchange<mode>): Likewise.
18104 (atomic_nand<mode>): Likewise.
18105 (atomic_fetch_<fetchop_name><mode>): Likewise.
18106 (atomic_nand_fetch<mode>): Likewise.
18107 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
18109 (ATOMIC): On power8, add QImode, HImode modes.
18110 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
18111 modes that promote to SImode.
18112 (load_lockedti): Convert TImode arguments to PTImode, so that we
18113 get a guaranteed even/odd register pair.
18114 (load_lockedpti): Likewise.
18115 (store_conditionalti): Likewise.
18116 (store_conditionalpti): Likewise.
18118 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
18119 atomic load/store instructions.
18122 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
18124 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
18126 (insn_count): New attribute, with most cases extracted from...
18127 (length): ...here. Redefine most cases in terms of insn_count.
18128 (single_insn): Delete.
18129 (can_delay): Use insn_count to check for single instructions.
18130 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
18131 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
18132 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
18133 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
18134 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
18135 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
18136 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
18137 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
18138 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
18139 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
18140 rather than "length".
18141 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
18142 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
18143 Use "insn_count" rather than "length".
18144 * config/mips/mips-dsp.md
18145 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
18146 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
18149 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
18151 PR tree-optimization/57361
18152 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
18154 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
18156 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
18158 (aarch64_simd_combine<mode>): New instruction expansion.
18159 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
18160 function prototype.
18161 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
18162 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
18164 2013-06-12 Jan Hubicka <jh@suse.cz>
18166 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
18167 decl has when in streaming stage.
18168 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
18169 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
18171 2013-06-12 Roland Stigge <stigge@antcom.de>
18174 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
18176 2013-06-12 Jakub Jelinek <jakub@redhat.com>
18178 PR tree-optimization/57537
18179 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
18180 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
18182 2013-06-12 Richard Biener <rguenther@suse.de>
18184 * data-streamer.h (streamer_write_char_stream): CSE
18185 obs->current_pointer.
18186 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
18187 streamer_write_char_stream manually and optimize the resulting loop.
18188 (streamer_write_hwi_stream): Likewise.
18190 2013-06-12 Jan Hubicka <jh@suse.cz>
18192 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
18193 * cgraph.h (varpool_create_empty_node): Declare.
18194 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
18196 * symtab.c (symtab_unregister_node): Be lax about missin entries
18198 (symtab_get_node): Update comment.
18199 * varpool.c (varpool_create_empty_node): Break out from ...
18200 (varpool_node_for_decl): ... here.
18201 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
18203 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
18205 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
18206 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
18207 part. Use straight-line flow at the end.
18208 <COMPONENT_REF>: Remove superfluous else.
18209 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
18211 2013-06-12 Jakub Jelinek <jakub@redhat.com>
18214 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
18215 target hook even for !TREE_PUBLIC decls. If no resolution info
18216 is available, return false for common and external decls.
18218 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18220 * config/rl78/constraints.md (U): New constraint.
18221 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
18224 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18227 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
18228 to allow returning address to AT_PLATFORM name.
18230 2013-06-11 Jan Hubicka <jh@suse.cz>
18232 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
18233 * cgraph.h (symtab_node_base): Add weakref flag.
18234 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
18235 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
18236 (output_weakrefs): Use weakref flag.
18237 * fold-const.c (simple_operand_p): Handle WEAK.
18238 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
18239 * ipa.c (varpool_externally_visible_p): Drop weakref.
18240 (function_and_variable_visibility): Update comment; fix weakref
18241 sanity checks; do not clear DECL_WEAK on them.
18242 * lto-cgraph.c (lto_output_node): update.
18243 (lto_output_varpool_node): Update.
18244 (input_overwrite_node): Update.
18245 (input_node): Update.
18246 (input_varpool_node): Update.
18247 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
18248 (lto_symtab_merge_symbols): Add sanity check.
18249 (lto_symtab_prevailing_decl): Do not special case weakrefs.
18250 * passes.c (rest_of_decl_compilation): Set static flag, too.
18251 * symtab.c (dump_symtab_base): Dump weakref.
18252 (verify_symtab_base): Sanity check weakrefs.
18253 (symtab_make_decl_local): Remove duplicated code.
18254 (symtab_alias_ultimate_target): Simplify.
18255 * varpool.c (varpool_create_variable_alias): Set weakref flag.
18257 2013-06-11 Tom de Vries <tom@codesourcery.com>
18259 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
18260 sequence_vect == NULL.
18262 2013-06-11 DJ Delorie <dj@redhat.com>
18264 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
18265 (rl78_unwind_word_mode): New.
18267 2013-06-11 David Malcolm <dmalcolm@redhat.com>
18269 * final.c (debug_prefix_maps): Make static.
18271 2013-06-11 David Malcolm <dmalcolm@redhat.com>
18273 * function.c (initial_trampoline): Remove stray copy.
18275 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
18277 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
18279 2013-06-11 Martin Jambor <mjambor@suse.cz>
18281 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
18282 within bounds at the beginning of the function.
18284 2013-06-11 Alan Modra <amodra@gmail.com>
18286 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
18288 (get_named_section): Don't NULL !DECL_P decl.
18290 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
18292 * doc/invoke.texi (core-avx2): Document.
18294 (atom): Updated with MOVBE.
18296 2013-06-11 Richard Biener <rguenther@suse.de>
18298 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
18300 2013-06-11 Anton Blanchard <anton@samba.org>
18302 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
18303 correct shift value in little-endian mode.
18305 2013-06-11 Jakub Jelinek <jakub@redhat.com>
18308 * varasm.c (get_variable_align): Move #endif to the right place.
18310 2013-06-10 Cary Coutant <ccoutant@google.com>
18312 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
18313 for hash so that hash table traversal order is deterministic.
18315 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
18316 Pat Haugen <pthaugen@us.ibm.com>
18317 Peter Bergner <bergner@vnet.ibm.com>
18319 * config/rs6000/vector.md (GPR move splitter): Do not split moves
18320 of vectors in GPRS if they are direct moves or quad word load or
18323 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
18325 (direct_move_p): Likewise.
18326 (quad_load_store_p): Likewise.
18328 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
18329 classes into bins based on the physical register type.
18330 (reg_class_to_reg_type): Likewise.
18331 (IS_STD_REG_TYPE): Likewise.
18332 (IS_FP_VECT_REG_TYPE): Likewise.
18333 (reload_fpr_gpr): Arrays to determine what insn to use if we can
18334 use direct move instructions.
18335 (reload_gpr_vsx): Likewise.
18336 (reload_vsx_gpr): Likewise.
18337 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
18338 information that is a simplification of register classes. Also
18339 precalculate direct move reload helpers.
18340 (direct_move_p): New function to return true if the operation can
18341 be done as a direct move instruciton.
18342 (quad_load_store_p): New function to return true if the operation
18343 is a quad memory operation.
18344 (rs6000_legitimize_address): If quad memory, only allow register
18345 indirect for TImode addresses.
18346 (rs6000_legitimate_address_p): Likewise.
18347 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
18348 (rs6000_reload_register_type): Likewise.
18349 (register_to_reg_type): Return register type.
18350 (rs6000_secondary_reload_simple_move): New helper function for
18351 secondary reload and secondary memory needed to identify anything
18352 that is a simple move, and does not need reloading.
18353 (rs6000_secondary_reload_direct_move): New helper function for
18354 secondary reload to identify cases that can be done with several
18355 instructions via the direct move instructions.
18356 (rs6000_secondary_reload_move): New helper function for secondary
18357 reload to identify moves between register types that can be done.
18358 (rs6000_secondary_reload): Add support for quad memory operations
18359 and for direct move.
18360 (rs6000_secondary_memory_needed): Likewise.
18361 (rs6000_debug_secondary_memory_needed): Change argument names.
18362 (rs6000_output_move_128bit): New function to return the move to
18363 use for 128-bit moves, including knowing about the various
18364 limitations of quad memory operations.
18366 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
18367 memory operations. call rs6000_output_move_128bit for the actual
18368 instruciton(s) to generate.
18369 (vsx_movti_64bit): Likewise.
18371 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
18372 (UNSPEC_P8V_MTVSRWZ): Likewise.
18373 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
18374 (UNSPEC_P8V_MTVSRD): Likewise.
18375 (UNSPEC_P8V_XXPERMDI): Likewise.
18376 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
18377 (UNSPEC_FUSION_GPR): Likewise.
18378 (FMOVE128_GPR): New iterator for direct move.
18379 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
18380 (f32_sv): Likewise.
18381 (f32_dm): Likewise.
18382 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
18383 loads and direct move instructions.
18384 (zero_extendsidi2_lfiwzx): Likewise.
18385 (extendsidi2_lfiwax): Likewise.
18386 (extendsidi2_nocell): Likewise.
18387 (floatsi<mode>2_lfiwax): Likewise.
18388 (lfiwax): Likewise.
18389 (floatunssi<mode>2_lfiwzx): Likewise.
18390 (lfiwzx): Likewise.
18391 (fix_trunc<mode>_stfiwx): Likewise.
18392 (fixuns_trunc<mode>_stfiwx): Likewise.
18393 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
18394 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
18395 (parity<mode>2_cmpb): Set length/type attr.
18396 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
18397 for 'mr.' to fast_compare.
18398 (bpermd_<mode>): Change type attr to popcnt.
18399 (p8_fmrgow_<mode>): New insns for power8 direct move support.
18400 (p8_mtvsrwz_1): Likewise.
18401 (p8_mtvsrwz_2): Likewise.
18402 (reload_fpr_from_gpr<mode>): Likewise.
18403 (p8_mtvsrd_1): Likewise.
18404 (p8_mtvsrd_2): Likewise.
18405 (p8_xxpermdi_<mode>): Likewise.
18406 (reload_vsx_from_gpr<mode>): Likewise.
18407 (reload_vsx_from_gprsf): Likewise.
18408 (p8_mfvsrd_3_<mode>): LIkewise.
18409 (reload_gpr_from_vsx<mode>): Likewise.
18410 (reload_gpr_from_vsxsf): Likewise.
18411 (p8_mfvsrd_4_disf): Likewise.
18412 (multi-word GPR splits): Do not split direct moves or quad memory
18415 2013-06-10 David Malcolm <dmalcolm@redhat.com>
18417 * tree-into-ssa.c (interesting_blocks): Make static.
18419 2013-06-10 Jakub Jelinek <jakub@redhat.com>
18422 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
18423 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
18424 Use DATA_ABI_ALIGNMENT for that case instead if defined.
18425 (get_variable_align): New function.
18426 (get_variable_section, emit_bss, emit_common,
18427 assemble_variable_contents, place_block_symbol): Use
18428 get_variable_align instead of DECL_ALIGN.
18429 (assemble_noswitch_variable): Add align argument, use it
18430 instead of DECL_ALIGN.
18431 (assemble_variable): Adjust caller. Use get_variable_align
18432 instead of DECL_ALIGN.
18433 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
18435 (DATA_ABI_ALIGNMENT): Define.
18436 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
18437 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
18438 opt is false, only return the psABI mandated alignment increase.
18439 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
18440 (DATA_ABI_ALIGNMENT): ... this.
18441 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
18442 (DATA_ABI_ALIGNMENT): ... this.
18443 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
18444 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
18445 (DATA_ABI_ALIGNMENT): ... this.
18446 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
18447 * doc/tm.texi: Regenerated.
18449 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
18451 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
18452 cmp_code to construct REG_EQUAL note.
18454 2013-06-09 Jakub Jelinek <jakub@redhat.com>
18457 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
18458 that operands[2] doesn't overlap with operands[0].
18460 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
18461 Jan Hubicka <jh@suse.cz>
18463 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
18464 hack to mark symbols as used.
18466 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
18468 PR rtl-optimization/57559
18469 * lra-constraints.c (process_alt_operands): Don't discourage
18470 memory with known offset for offsetable memory constraint.
18471 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
18473 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
18475 * varasm.c (struct oc_local_state): Reorder fields.
18476 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
18477 and adjust accordingly.
18478 (output_constructor): Reorder initialization code and adjust call to
18479 output_constructor_bitfield.
18481 2013-06-07 Jan Hubicka <jh@suse.cz>
18483 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
18485 2013-06-07 David Malcolm <dmalcolm@redhat.com>
18487 * tree-object-size.c (unknown): Make const.
18489 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18491 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
18492 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
18493 for last alternative in the cpu_facility attribute.
18495 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18498 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
18499 (xordi3): Change operand 2 constraint to arm_xordi_operand.
18500 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
18501 * config/arm/constraints.md (Dg): New constraint.
18502 * config/arm/neon.md (xordi3_neon): Remove.
18503 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
18504 * config/arm/predicates.md (arm_xordi_operand): New predicate.
18506 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18508 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
18509 Clean up alternatives.
18511 2013-06-07 Alan Modra <amodra@gmail.com>
18513 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
18516 2013-06-07 Alan Modra <amodra@gmail.com>
18518 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
18520 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18522 * config/arm/constraints.md (Df): New constraint.
18523 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
18524 Correct length attribute for last two alternatives.
18526 2013-06-07 Alan Modra <amodra@gmail.com>
18528 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
18529 override user -mfp-in-toc.
18530 (offsettable_ok_by_alignment): Consider just the current access
18531 rather than the whole object, unless BLKmode. Handle
18532 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
18533 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
18534 for -mcmodel=medium.
18535 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
18536 override user -mfp-in-toc or -msum-in-toc. Default to
18537 -mno-fp-in-toc for -mcmodel=medium.
18539 2013-06-06 DJ Delorie <dj@redhat.com>
18541 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
18542 TARGET_VALID_POINTER_MODE.
18544 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
18545 Pat Haugen <pthaugen@us.ibm.com>
18546 Peter Bergner <bergner@vnet.ibm.com>
18548 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
18549 Document new power8 builtins.
18551 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
18552 condition code register, to allow 128-bit logical operations to be
18553 done in the VSX or GPR registers.
18554 (nor<mode>3): Use the canonical form for nor.
18555 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
18556 vclz*, and vpopcnt* vector instructions.
18557 (nand<mode>3): Likewise.
18558 (orc<mode>3): Likewise.
18559 (clz<mode>2): LIkewise.
18560 (popcount<mode>2): Likewise.
18562 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
18563 that only the GPRs are recognized.
18565 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18566 support for new power8 builtins.
18568 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
18570 (xscvdpspn): Likewise.
18576 (vpopcnt): Likewise.
18577 (vpopcntb): Likewise.
18578 (vpopcnth): Likewise.
18579 (vpopcntw): Likewise.
18580 (vpopcntd): Likewise.
18582 (vmrgew): Likewise.
18583 (vmrgow): Likewise.
18585 (eqv_v16qi3): Likewise.
18586 (eqv_v8hi3): Likewise.
18587 (eqv_v4si3): Likewise.
18588 (eqv_v2di3): Likewise.
18589 (eqv_v4sf3): Likewise.
18590 (eqv_v2df3): Likewise.
18592 (nand_v16qi3): Likewise.
18593 (nand_v8hi3): Likewise.
18594 (nand_v4si3): Likewise.
18595 (nand_v2di3): Likewise.
18596 (nand_v4sf3): Likewise.
18597 (nand_v2df3): Likewise.
18599 (orc_v16qi3): Likewise.
18600 (orc_v8hi3): Likewise.
18601 (orc_v4si3): Likewise.
18602 (orc_v2di3): Likewise.
18603 (orc_v4sf3): Likewise.
18604 (orc_v2df3): Likewise.
18606 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
18607 allow power8 quad mode in 64-bit.
18608 (rs6000_builtin_vectorized_function): Add support to vectorize
18609 ISA 2.07 count leading zeros, population count builtins.
18610 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
18611 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
18612 (builtin_function_type): Add vgbbd builtin function which takes an
18614 (altivec_expand_vec_perm_const): Add support for new power8 merge
18617 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
18618 that does not include TImdoe for use with 32-bit.
18619 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
18621 (UNSPEC_VSX_CVDPSPN): Likewise.
18622 (vsx_xscvdpspn): Likewise.
18623 (vsx_xscvspdpn): Likewise.
18624 (vsx_xscvdpspn_scalar): Likewise.
18625 (vsx_xscvspdpn_directmove): Likewise.
18626 (vsx_and<mode>3): Split logical operations into 32-bit and
18627 64-bit. Add support to do logical operations on TImode as well as
18628 VSX vector types. Allow logical operations to be done in either
18629 VSX registers or in general purpose registers in 64-bit mode. Add
18630 splitters if GPRs were used. For AND, add clobber of CCmode to
18631 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
18633 (vsx_and<mode>3_32bit): Likewise.
18634 (vsx_and<mode>3_64bit): Likewise.
18635 (vsx_ior<mode>3): Likewise.
18636 (vsx_ior<mode>3_32bit): Likewise.
18637 (vsx_ior<mode>3_64bit): Likewise.
18638 (vsx_xor<mode>3): Likewise.
18639 (vsx_xor<mode>3_32bit): Likewise.
18640 (vsx_xor<mode>3_64bit): Likewise.
18641 (vsx_one_cmpl<mode>2): Likewise.
18642 (vsx_one_cmpl<mode>2_32bit): Likewise.
18643 (vsx_one_cmpl<mode>2_64bit): Likewise.
18644 (vsx_nor<mode>3): Likewise.
18645 (vsx_nor<mode>3_32bit): Likewise.
18646 (vsx_nor<mode>3_64bit): Likewise.
18647 (vsx_andc<mode>3): Likewise.
18648 (vsx_andc<mode>3_32bit): Likewise.
18649 (vsx_andc<mode>3_64bit): Likewise.
18650 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
18651 and xxlorc instructions.
18652 (vsx_eqv<mode>3_64bit): Likewise.
18653 (vsx_nand<mode>3_32bit): Likewise.
18654 (vsx_nand<mode>3_64bit): Likewise.
18655 (vsx_orc<mode>3_32bit): Likewise.
18656 (vsx_orc<mode>3_64bit): Likewise.
18658 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
18660 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
18662 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
18663 (p8_vmrgow): Likewise.
18664 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
18665 GPRs to be split under VSX.
18666 (p8v_clz<mode>2): Add power8 count leading zero support.
18667 (p8v_popcount<mode>2): Add power8 population count support.
18668 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
18671 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
18674 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
18676 (vec_nand): Likewise.
18677 (vec_vclz): Likewise.
18678 (vec_vclzb): Likewise.
18679 (vec_vclzd): Likewise.
18680 (vec_vclzh): Likewise.
18681 (vec_vclzw): Likewise.
18682 (vec_vgbbd): Likewise.
18683 (vec_vmrgew): Likewise.
18684 (vec_vmrgow): Likewise.
18685 (vec_vpopcnt): Likewise.
18686 (vec_vpopcntb): Likewise.
18687 (vec_vpopcntd): Likewise.
18688 (vec_vpopcnth): Likewise.
18689 (vec_vpopcntw): Likewise.
18691 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18693 PR rtl-optimization/57468
18694 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
18697 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18699 PR rtl-optimization/57459
18700 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
18701 type when setting live regs.
18703 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18705 * config/s390/s390.opt (mlra): New option.
18706 * config/s390/s390.c (s390_decompose_address): Check displacement
18707 for all registers for LRA.
18708 (s390_secondary_reload): Don't used secondary reloads for LRA.
18709 (s390_lra_p): New function.
18710 (TARGET_LRA_P): Define.
18711 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
18712 of attribute cpu_facility to zarch for the last alternative.
18713 (*cmpmem_short): Ditto.
18715 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
18717 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
18718 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
18719 (arm_expand_prologue): Likewise.
18721 2013-06-06 Teresa Johnson <tejohnson@google.com>
18724 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
18725 as this is now done by redirect_edge_and_branch_force.
18726 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
18727 barriers, and fix interaction with splitting.
18728 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
18729 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
18730 reflect changes made in the routine.
18731 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
18732 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
18733 since this is called in cfglayout mode, and replace partition fixup
18734 with assert as that is now done by force_nonfallthru_and_redirect.
18735 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
18736 already be marked with region crossing note.
18737 (insert_section_boundary_note): Make non-static, gate on flag
18738 has_bb_partition, rewrite to also check for multiple partitions.
18739 (rest_of_handle_reorder_blocks): Remove call to
18740 insert_section_boundary_note, now done later during free_cfg.
18741 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
18742 * bb-reorder.h (insert_section_boundary_note): Declare.
18743 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
18744 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
18745 invoke insert_section_boundary_note.
18746 (try_redirect_by_replacing_jump): Remove unnecessary
18747 check for region crossing note.
18748 (fixup_partition_crossing): New function.
18749 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
18750 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
18751 in non-cfglayout mode.
18752 (force_nonfallthru_and_redirect): Fixup partition boundaries,
18753 remove old code that tried to do this. Emit barrier correctly
18754 when we are in cfglayout mode.
18755 (last_bb_in_partition): New function.
18756 (rtl_split_edge): Correctly fixup partition boundaries.
18757 (commit_one_edge_insertion): Remove old code that tried to
18758 fixup region crossing edge since this is now handled in
18759 split_block, and set up insertion point correctly since
18760 block may now end in a jump.
18761 (verify_hot_cold_block_grouping): Guard against checking when not in
18762 linearized RTL mode.
18763 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
18765 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
18766 rtl_verify_flow_info, so not called in cfglayout mode.
18767 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
18768 (fixup_reorder_chain): Remove old code that attempted to fixup region
18769 crossing note as this is now handled in force_nonfallthru_and_redirect.
18770 (duplicate_insn_chain): Don't duplicate switch section notes.
18771 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
18773 * basic-block.h (emit_barrier_after_bb): Declare.
18775 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18777 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
18778 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
18779 arm_usatsihi): Adjust alternatives for arm_restrict_it.
18781 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18783 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
18785 * config/arm/ldmstm.md: Regenerate.
18787 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18789 * config/arm/sync.md (atomic_loaddi_1):
18790 Disable predication for arm_restrict_it.
18791 (arm_load_exclusive<mode>): Likewise.
18792 (arm_load_exclusivesi): Likewise.
18793 (arm_load_exclusivedi): Likewise.
18794 (arm_load_acquire_exclusive<mode>): Likewise.
18795 (arm_load_acquire_exclusivesi): Likewise.
18796 (arm_load_acquire_exclusivedi): Likewise.
18797 (arm_store_exclusive<mode>): Likewise.
18798 (arm_store_exclusive<mode>): Likewise.
18799 (arm_store_release_exclusivedi): Likewise.
18800 (arm_store_release_exclusive<mode>): Likewise.
18802 2013-06-06 Richard Biener <rguenther@suse.de>
18804 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
18806 (lto_tag_is_tree_code_p): Adjust.
18807 (lto_tag_is_gimple_code_p): Likewise.
18808 (lto_gimple_code_to_tag): Likewise.
18809 (lto_tag_to_gimple_code): Likewise.
18810 (lto_tree_code_to_tag): Likewise.
18811 (lto_tag_to_tree_code): Likewise.
18812 * data-streamer.h (streamer_write_hwi_in_range): Use
18813 uhwi streaming to stream the normalized range.
18814 (streamer_read_hwi_in_range): Likewise.
18816 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18818 * config/arm/arm.md (enabled_for_depr_it): New attribute.
18819 (predicable_short_it): Likewise.
18820 (predicated): Likewise.
18821 (enabled): Handle above.
18822 (define_cond_exec): Set predicated attribute to yes.
18824 2013-06-05 Mike Stump <mikestump@comcast.net>
18826 * gdbinit.in (__FUNCTION__): Add.
18828 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
18830 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
18831 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
18833 2013-06-05 Jan Hubicka <jh@suse.cz>
18835 * varasm.c (mark_decl_referenced): Revert the removal until targets
18838 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
18840 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
18841 instead of mark_decl_referenced.
18843 2013-06-05 Jan Hubicka <jh@suse.cz>
18845 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
18846 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
18847 and symtab_used_from_object_file_p.
18848 (cgraph_make_node_local_1): Clear forced_by_abi.
18849 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
18850 * cgraph.h (symtab_node_base): Add forced_by_abi.
18851 (decide_is_variable_needed): Remove.
18852 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
18853 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
18854 (decide_is_symbol_needed): ... this one; handle symbols in general;
18855 always analyze virtuals; honnor forced_by_abi.
18856 (cgraph_finalize_function): Update.
18857 (varpool_finalize_decl): Update.
18858 (symbol_defined_and_needed): Remove.
18859 (analyze_functions): Update.
18860 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18861 output_refs, input_overwrite_node): Handle forced_by_abi.
18862 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18863 (address_taken_from_non_vtable_p): ... this one.
18864 (comdat_can_be_unshared_p_1): New function.
18865 (cgraph_comdat_can_be_unshared_p): Rename to ...
18866 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18867 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18868 (function_and_variable_visibility): Clear forced_by_abi as needed.
18869 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18870 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18871 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18872 * varpool.c (decide_is_variable_needed): Remove.
18874 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18876 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18877 (arm_option_override): Override arm_restrict_it where appropriate.
18878 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18879 * config/arm/arm.opt (mrestrict-it): New command-line option.
18880 * doc/invoke.texi: Document -mrestrict-it.
18882 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18884 * tsan.c (tsan_atomic_table): Make const.
18886 2013-06-05 Richard Biener <rguenther@suse.de>
18888 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18889 index associated with the tree we are supposed to replace.
18890 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18891 TREE_ASM_WRITTEN as zero for everything but SSA names.
18893 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18895 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18897 2013-06-04 Jan Hubicka <jh@suse.cz>
18899 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18900 (input_node, input_varpool_node): Handle correctly external same
18902 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18903 nodes at ltrans stage.
18905 2013-06-04 Jan Hubicka <jh@suse.cz>
18907 * ipa-inline.c (update_caller_keys): Fix availability test.
18908 (update_callee_keys): Likewise.
18909 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18910 to follow ELF standard.
18912 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18914 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18915 (mips64r5900el-*-elf*): New configurations.
18916 * config/mips/mips-cpus.def (r5900): New processor.
18917 * config/mips/mips-tables.opt: Regenerate.
18918 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18919 (mips_issue_rate): Handle PROCESSOR_R5900.
18920 (mips_reorg_process_insns): Force reorder mode for the R5900.
18921 * config/mips/mips.h (TARGET_MIPS5900): Define.
18922 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18924 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18926 * config/mips/mips.md (processor): Add r5900.
18927 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18929 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18931 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18932 into function to generate MOVI instruction.
18933 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18934 (aarch64_preferred_simd_mode): Turn into wrapper.
18935 (aarch64_output_scalar_simd_mov_immediate): New function.
18936 * config/aarch64/aarch64-protos.h: Add prototype for above.
18938 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18940 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18941 element_char member.
18942 (sizetochar): Return signed char.
18943 (aarch64_simd_valid_immediate): Remove elchar and other
18944 unnecessary variables.
18945 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18946 Calculate element_char as required.
18947 * config/aarch64/aarch64-protos.h: Update and move prototype
18948 for aarch64_output_simd_mov_immediate.
18949 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18952 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18954 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18955 information completed by aarch64_simd_valid_immediate.
18956 (aarch64_legitimate_constant_p): Update arguments.
18957 (aarch64_simd_valid_immediate): Work with struct rather than many
18959 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18960 (aarch64_simd_make_constant): Update arguments.
18961 (aarch64_output_simd_mov_immediate): Work with struct rather than
18962 many pointers. Output immediate directly rather than as operand.
18963 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18965 * config/aarch64/constraints.md (Dn): Update arguments.
18967 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18969 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18971 (aarch64_simd_immediate_valid_for_move): Remove.
18972 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18973 (aarch64_simd_make_constant): Update call.
18974 (aarch64_output_simd_mov_immediate): Update call.
18975 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18977 * config/aarch64/constraints.md (Dn): Update call.
18979 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18981 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18982 return type to bool for prototype.
18983 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18984 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18985 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18987 2013-06-04 Catherine Moore <clm@codesourcery.com>
18989 * config/mips/mips.opt (meva): New.
18990 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18991 (ASM_SPEC): Handle -meva.
18992 * doc/invoke.texi (meva): Document.
18994 2013-06-04 Alan Modra <amodra@gmail.com>
18996 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18999 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19001 * rtl.def: Add extra fourth optional field to define_cond_exec.
19002 * gensupport.c (process_one_cond_exec): Process attributes from
19004 * doc/md.texi: Document fourth field in define_cond_exec.
19006 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
19008 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
19009 out the processing order as in store_bit_field_1.
19011 2013-06-04 Jan Hubicka <jh@suse.cz>
19013 PR middle-end/57500
19014 * cgraphunit.c (cgraph_process_same_body_aliases): Create
19015 non-VAR_DECL node if it does not exist yet.
19017 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
19019 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
19020 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
19021 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
19022 target_cpu_default setting.
19024 2013-06-03 Teresa Johnson <tejohnson@google.com>
19026 * dumpfile.c (opt_info_switch_p): Change -fopt-info
19027 default to -fopt-info=optimized instead of all.
19028 * doc/invoke.texi: Ditto.
19029 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
19030 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
19031 (execute_vect_slp): Emit BB vectorization success under
19032 MSG_OPTIMIZED_LOCATIONS.
19033 * tree-vect-slp.c (vect_slp_transform_bb): Change
19034 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
19035 * tree-vect-loop.c (vect_transform_loop): Ditto.
19037 2013-06-03 Jason Merrill <jason@redhat.com>
19040 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
19041 Use TARGET_EXPR for C++.
19043 2013-06-03 Jakub Jelinek <jakub@redhat.com>
19045 PR rtl-optimization/57268
19046 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
19047 if DEBUG_INSN_P (insn).
19050 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
19052 PR rtl-optimization/57268
19053 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19054 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19056 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19058 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
19059 (ix86_avoid_lea_for_addr): Likewise.
19060 (exact_dependency_1): Likewise.
19061 (ix86_adjust_cost): Likewise.
19062 (swap_top_of_ready_list): Fix formatting and !reload_completed check
19064 (do_reorder_for_imul): Fix typo, formatting and
19065 !reload_completed check removed.
19066 (ix86_sched_reorder): Fix typo and formatting.
19067 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
19070 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
19072 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
19074 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
19076 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
19077 <STRING_CST>: Likewise.
19078 <VECTOR_CST>: Likewise.
19080 2013-06-01 Janus Weil <janus@gcc.gnu.org>
19081 Mikael Morin <mikael@gcc.gnu.org>
19083 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
19084 * config.in: Regenerated.
19085 * configure: Regenerated.
19087 2013-06-01 Jan Hubicka <jh@suse.cz>
19089 PR middle-end/57366
19090 * cgraphunit.c (compile): When weakref is not supported,
19091 set up transparent aliases before final output pass.
19092 * varasm.c (assemble_alias): Do not try to do it here.
19094 2013-06-01 Jan Hubicka <jh@suse.cz>
19096 PR middle-end/57467
19097 * passes.c (for_per_function): Skip unanalyzed functions.
19099 2013-06-01 Jan Hubicka <jh@suse.cz>
19101 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
19102 (lto_symtab_merge_symbols_1): ... this one.
19103 (lto_symtab_merge_cgraph_nodes): Rename to ...
19104 (lto_symtab_merge_symbols): ... this one; simplify.
19105 * cgraph.c (same_body_aliases_done): Rename to ...
19106 (cpp_implicit_aliases_done): ... this one.
19107 (cgraph_create_function_alias): Update.
19108 (cgraph_same_body_alias): Update.
19109 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
19110 (verify_edge_corresponds_to_fndecl): Simplify.
19111 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
19112 (cgraph_node): Remove same_body_alias.
19113 (varpool_node): Remove alias_of and extra_name_alias.
19114 (same_body_aliases_done): Rename to ..
19115 (cpp_implicit_aliases_done): ... this one.
19116 (symtab_alias_ultimate_target): Add default parameter.
19117 (symtab_resolve_alias): New function.
19118 (fixup_same_cpp_alias_visibility): Declare.
19119 (cgraph_function_node): Add default parameter.
19120 (cgraph_node_asm_name): Likewise.
19121 (cgraph_function_or_thunk_node): Add default parameter; do
19122 not ICE when it is NULL.
19123 (varpool_variable_node): Likewise.
19124 * tree-emutls.c (create_emultls_var): Update.
19125 (ipa_lower_emutls): Update.
19126 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
19127 (cgraph_reset_node): Reset alias info.
19128 (cgraph_finalize_function): Update.
19129 (fixup_same_cpp_alias_visibility): Move to symtab.c.
19130 (analyze_function): Simplify.
19131 (cgraph_process_same_body_aliases): Simplify.
19132 (analyze_functions): Fixup same body aliases.
19133 (handle_alias_pairs): Simplify.
19134 (assemble_thunk): Update.
19135 (assemble_thunks_and_aliases): Update.
19136 (output_weakrefs): Rewrite.
19137 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
19138 (lto_output_varpool_node): Likewise.
19139 (compute_ltrans_boundary): Remve assert.
19140 (get_alias_symbol): New functoin.
19141 (input_node): Rewrite alias handling.
19142 (input_varpool_node): Likewise.
19143 * ipa-pure-const.c (propagate_pure_const): Fix formating.
19144 * ipa.c (process_references): Handle weakrefs correctly.
19145 (symtab_remove_unreachable_nodes): Likewise.
19146 * trans-mem.c (get_cg_data): Update.
19147 (ipa_tm_create_version_alias): Update.
19148 (ipa_tm_execute): Update.
19149 * symtab.c (dump_symtab_base): Dump aliases.
19150 (verify_symtab_base): Verify aliases.
19151 (symtab_node_availability): New function.
19152 (symtab_alias_ultimate_target): Simplify.
19153 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
19154 handle all the fixup cases.
19155 (symtab_resolve_alias): New function.
19156 * passes.c (ipa_write_summaries): Handle weakrefs.
19157 * varpool.c (varpool_analyze_node): Simplify.
19158 (assemble_aliases): Update.
19159 (varpool_create_variable_alias): Simplify.
19160 (varpool_extra_name_alias): Simplify.
19161 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
19162 (lto_symtab_merge_symbols): ... this one.
19164 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
19167 PR rtl-optimization/57268
19168 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19169 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19171 2013-06-01 Tobias Burnus <burnus@net-b.de>
19173 Partially reverted:
19174 2013-05-31 Tobias Burnus <burnus@net-b.de>
19176 PR middle-end/57073
19177 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
19180 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
19182 PR rtl-optimization/57268
19183 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19184 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19186 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
19188 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
19189 unordered comparison operators when -fno-trapping-math is in effect
19191 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
19192 and implement unordered comparison operators properly on the e500.
19194 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
19196 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
19197 for constant scalar integers.
19198 (simplify_relational_operation_1): Likewise.
19200 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
19202 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
19203 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
19206 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19207 Igor Zamyatin <igor.zamyatin@intel.com>
19209 Silvermont (SLM) architecture performance tuning.
19210 * config/i386/i386.h (enum ix86_tune_indices): Add
19211 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
19212 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
19214 * config/i386/i386.c (initial_ix86_tune_features)
19215 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
19216 (ix86_lea_outperforms): Handle Silvermont tuning.
19217 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
19219 (ix86_use_lea_for_mov): Likewise.
19220 (ix86_avoid_lea_for_addr): Likewise.
19221 (ix86_lea_for_add_ok): Likewise.
19222 (exact_dependency_1): New function.
19223 (exact_store_load_dependency): Likewise.
19224 (ix86_adjust_cost): Handle Silvermont tuning.
19225 (do_reoder_for_imul): Likewise.
19226 (swap_top_of_ready_list): New function.
19227 (ix86_sched_reorder): Changed to handle Silvermont tuning.
19229 * config/i386/i386.md (peepholes that split memory operand in fp
19232 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
19234 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
19235 Remove un-necessary braces.
19237 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
19239 * config/aarch64/aarch64.c (aarch64_classify_symbol):
19240 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
19242 2013-05-31 Tobias Burnus <burnus@net-b.de>
19244 PR middle-end/57073
19245 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
19247 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19250 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
19251 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
19252 * config/arm/neon.md (iordi3_neon): Remove.
19253 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
19254 * config/arm/predicates.md (imm_for_neon_logic_operand):
19255 Move to earlier in the file.
19256 (neon_logic_op2): Likewise.
19257 (arm_iordi_operand_neon): New predicate.
19259 2013-05-31 Richard Biener <rguenther@suse.de>
19261 PR tree-optimization/57478
19262 PR tree-optimization/57453
19263 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
19266 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19268 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
19269 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
19271 2013-05-30 Tobias Burnus <burnus@net-b.de>
19272 Thomas Koenig <tkoenig@gcc.gnu.org>
19274 PR middle-end/57073
19275 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
19276 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
19278 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
19280 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
19282 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
19284 * target.def (register_usage_leveling_p): New hook.
19285 * targhooks.c (default_register_usage_leveling_p): New.
19286 * targhooks.h (default_register_usage_leveling_p): New prototype.
19287 * lra-assigns.c (register_usage_leveling_p): Use the hook.
19288 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
19289 * doc/tm.texi: Update.
19290 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
19292 2013-05-30 Ian Bolton <ian.bolton@arm.com>
19294 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
19295 (*insv_reg<mode>): New define_insn.
19297 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
19299 PR rtl-optimization/57439
19300 * postreload.c (move2add_valid_value_p): Check that we have
19301 a zero subreg_regno_offset when accessing the register in
19302 the requested mode.
19304 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19305 Igor Zamyatin <igor.zamyatin@intel.com>
19307 Silvermont (SLM) architecture pipeline model, tuning and
19309 * config.gcc: Add slm config options and target.
19311 * config/i386/slm.md: New.
19313 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
19315 * config/i386/i386-c.c (ix86_target_macros_internal): New case
19317 (ix86_target_macros_internal): Likewise.
19319 * config/i386/i386.c (slm_cost): New cost.
19320 (m_SLM): New macro flag.
19321 (initial_ix86_tune_features): Set m_SLM.
19322 (x86_accumulate_outgoing_args): Likewise.
19323 (x86_arch_always_fancy_math_387): Likewise.
19324 (processor_target_table): Add slm cost.
19325 (cpu_names): Add slm cpu name.
19326 (x86_option_override_internal): Set SLM ISA.
19327 (ix86_issue_rate): New case PROCESSOR_SLM.
19328 (ia32_multipass_dfa_lookahead): Likewise.
19329 (fold_builtin_cpu): Add slm.
19331 * config/i386/i386.h (TARGET_SLM): New target macro.
19332 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
19333 (processor_type): Add PROCESSOR_SLM.
19335 * config/i386/i386.md (cpu): Add new value "slm".
19336 (slm.md): Include slm.md.
19338 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
19339 Zhenqiang Chen <zhenqiang.chen@linaro.org>
19341 * config/arm/arm-protos.h: Add and update function protos.
19342 * config/arm/arm.c (use_simple_return_p): New added.
19343 (thumb2_expand_return): Check simple_return flag.
19344 * config/arm/arm.md: Add simple_return and conditional simple_return.
19345 * config/arm/iterators.md: Add iterator for return and simple_return.
19347 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
19349 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
19350 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
19351 (arm_emit_vfp_multi_reg_pop): Likewise.
19352 (thumb2_emit_ldrd_pop): Likewise.
19353 (arm_expand_epilogue): Add misc REG_CFA notes.
19354 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
19356 2013-05-29 Lawrence Crowl <crowl@google.com>
19358 * config/arm/t-arm: Update for below.
19360 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
19361 Change type to hash_table. Update dependent calls and types.
19363 * config/i386/t-cygming: Update for below.
19365 * config/i386/t-interix: Update for below.
19367 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
19368 Change type to hash_table. Update dependent calls and types.
19369 (i386_find_on_wrapper_list::wrappers): Likewise.
19371 * config/ia64/t-ia64: Update for below.
19373 * config/ia64/ia64.c (bundle_state_table):
19374 Change type to hash_table. Update dependent calls and types.
19376 * config/mips/mips.c (mips_reorg_process_insns::htab):
19377 Change type to hash_table. Update dependent calls and types.
19379 * config/sol2.c (solaris_comdat_htab):
19380 Change type to hash_table. Update dependent calls and types.
19382 * config/t-sol2: Update for above.
19384 2013-05-29 Teresa Johnson <tejohnson@google.com>
19386 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
19387 functions are not yet marked as defined.
19389 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
19390 Pat Haugen <pthaugen@us.ibm.com>
19391 Peter Bergner <bergner@vnet.ibm.com>
19393 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
19397 (vrotl<mode>3): Likewise.
19398 (vashl<mode>3): Likewise.
19399 (vlshr<mode>3): Likewise.
19400 (vashr<mode>3): Likewise.
19402 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
19403 support for power8 V2DI builtins.
19405 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
19406 power8 V2DI builtins.
19407 (vupkhsw): Likewise.
19408 (vupklsw): Likewise.
19409 (vaddudm): Likewise.
19410 (vminsd): Likewise.
19411 (vmaxsd): Likewise.
19412 (vminud): Likewise.
19413 (vmaxud): Likewise.
19414 (vpkudum): Likewise.
19415 (vpksdss): Likewise.
19416 (vpkudus): Likewise.
19417 (vpksdus): Likewise.
19422 (vsubudm): Likewise.
19423 (vcmpequd): Likewise.
19424 (vcmpgtsd): Likewise.
19425 (vcmpgtud): Likewise.
19426 (vcmpequd_p): Likewise.
19427 (vcmpgtsd_p): Likewise.
19428 (vcmpgtud_p): Likewise.
19429 (vupkhsw): Likewise.
19430 (vupklsw): Likewise.
19431 (vaddudm): Likewise.
19432 (vmaxsd): Likewise.
19433 (vmaxud): Likewise.
19434 (vminsd): Likewise.
19435 (vminud): Likewise.
19436 (vpksdss): Likewise.
19437 (vpksdus): Likewise.
19438 (vpkudum): Likewise.
19439 (vpkudus): Likewise.
19444 (vsubudm): Likewise.
19446 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
19447 support for power8 V2DI instructions.
19449 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
19450 power8 V2DI instructions. Combine pack and unpack insns to use an
19451 iterator for each mode. Check whether a particular mode supports
19452 Altivec instructions instead of just checking TARGET_ALTIVEC.
19453 (UNSPEC_VPKUWUM): Likewise.
19454 (UNSPEC_VPKSHSS): Likewise.
19455 (UNSPEC_VPKSWSS): Likewise.
19456 (UNSPEC_VPKUHUS): Likewise.
19457 (UNSPEC_VPKSHUS): Likewise.
19458 (UNSPEC_VPKUWUS): Likewise.
19459 (UNSPEC_VPKSWUS): Likewise.
19460 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
19461 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
19462 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
19463 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
19464 (UNSPEC_VUPKHSB): Likewise.
19465 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
19466 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
19467 (UNSPEC_VUPKHSH): Likewise.
19468 (UNSPEC_VUPKLSB): Likewise.
19469 (UNSPEC_VUPKLSH): Likewise.
19471 (VI_char): Likewise.
19472 (VI_scalar): Likewise.
19473 (VI_unit): Likewise.
19475 (VP_small): Likewise.
19476 (VP_small_lc): Likewise.
19477 (VU_char): Likewise.
19478 (add<mode>3): Likewise.
19479 (altivec_vaddcuw): Likewise.
19480 (altivec_vaddu<VI_char>s): Likewise.
19481 (altivec_vadds<VI_char>s): Likewise.
19482 (sub<mode>3): Likewise.
19483 (altivec_vsubcuw): Likewise.
19484 (altivec_vsubu<VI_char>s): Likewise.
19485 (altivec_vsubs<VI_char>s): Likewise.
19486 (altivec_vavgs<VI_char>): Likewise.
19487 (altivec_vcmpbfp): Likewise.
19488 (altivec_eq<mode>): Likewise.
19489 (altivec_gt<mode>): Likewise.
19490 (altivec_gtu<mode>): Likewise.
19491 (umax<mode>3): Likewise.
19492 (smax<mode>3): Likewise.
19493 (umin<mode>3): Likewise.
19494 (smin<mode>3): Likewise.
19495 (altivec_vpkuhum): Likewise.
19496 (altivec_vpkuwum): Likewise.
19497 (altivec_vpkshss): Likewise.
19498 (altivec_vpkswss): Likewise.
19499 (altivec_vpkuhus): Likewise.
19500 (altivec_vpkshus): Likewise.
19501 (altivec_vpkuwus): Likewise.
19502 (altivec_vpkswus): Likewise.
19503 (altivec_vpks<VI_char>ss): Likewise.
19504 (altivec_vpks<VI_char>us): Likewise.
19505 (altivec_vpku<VI_char>us): Likewise.
19506 (altivec_vpku<VI_char>um): Likewise.
19507 (altivec_vrl<VI_char>): Likewise.
19508 (altivec_vsl<VI_char>): Likewise.
19509 (altivec_vsr<VI_char>): Likewise.
19510 (altivec_vsra<VI_char>): Likewise.
19511 (altivec_vsldoi_<mode>): Likewise.
19512 (altivec_vupkhsb): Likewise.
19513 (altivec_vupkhs<VU_char>): Likewise.
19514 (altivec_vupkls<VU_char>): Likewise.
19515 (altivec_vupkhsh): Likewise.
19516 (altivec_vupklsb): Likewise.
19517 (altivec_vupklsh): Likewise.
19518 (altivec_vcmpequ<VI_char>_p): Likewise.
19519 (altivec_vcmpgts<VI_char>_p): Likewise.
19520 (altivec_vcmpgtu<VI_char>_p): Likewise.
19521 (abs<mode>2): Likewise.
19522 (vec_unpacks_hi_v16qi): Likewise.
19523 (vec_unpacks_hi_v8hi): Likewise.
19524 (vec_unpacks_lo_v16qi): Likewise.
19525 (vec_unpacks_hi_<VP_small_lc>): Likewise.
19526 (vec_unpacks_lo_v8hi): Likewise.
19527 (vec_unpacks_lo_<VP_small_lc>): Likewise.
19528 (vec_pack_trunc_v8h): Likewise.
19529 (vec_pack_trunc_v4si): Likewise.
19530 (vec_pack_trunc_<mode>): Likewise.
19532 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
19534 (vec_vmaxsd): Likewise.
19535 (vec_vmaxud): Likewise.
19536 (vec_vminsd): Likewise.
19537 (vec_vminud): Likewise.
19538 (vec_vpksdss): Likewise.
19539 (vec_vpksdus): Likewise.
19540 (vec_vpkudum): Likewise.
19541 (vec_vpkudus): Likewise.
19542 (vec_vrld): Likewise.
19543 (vec_vsld): Likewise.
19544 (vec_vsrad): Likewise.
19545 (vec_vsrd): Likewise.
19546 (vec_vsubudm): Likewise.
19547 (vec_vupkhsw): Likewise.
19548 (vec_vupklsw): Likewise.
19550 2013-05-29 Jan Hubicka <jh@suse.cz>
19552 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
19553 flags; reorder rest of fields in more consistent way.
19554 (varpool_node): Remove analyzed, finalized and alias.
19555 (cgraph_ndoe): Likewise.
19556 (symtab_alias_ultimate_target): New function.
19557 (cgraph_function_node): Move offline.
19558 (cgraph_reset_node): Declare.
19559 (cgraph_comdat_can_be_unshared_p): Remove.
19560 (varpool_remove_initializer): Declare.
19561 (varpool_first_defined_variable, varpool_next_defined_variable
19562 cgraph_first_defined_function, cgraph_next_defined_function): Update.
19563 (cgraph_function_with_gimple_body_p): Update.
19564 (varpool_all_refs_explicit_p): Update.
19565 (symtab_alias_target): New function.
19566 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
19567 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
19568 (cgraph_function_or_thunk_node): Simplify using
19569 symtab_alias_ultimate_target.
19570 (varpool_variable_node): Likewise.
19571 * cgraph.c (cgraph_create_function_alias): Update.
19572 (cgraph_add_thunk): Update.
19573 (cgraph_remove_node): Update.
19574 (dump_cgraph_node): Do not dump removed flags.
19575 (cgraph_function_body_availability): Update.
19576 (cgraph_propagate_frequency): Update.
19577 (verify_cgraph_node): Check sanity of local flag.
19578 (cgraph_function_node): Move here from cgraph.h; revamp for
19579 cgraph_function_or_thunk_node.
19580 * lto-symtab.c (lto_varpool_replace_node): Update.
19581 (lto_symtab_resolve_can_prevail_p): Update.
19582 (lto_symtab_merge_cgraph_nodes): Update.
19583 * ipa-cp.c (determine_versionability, initialize_node_lattices,
19584 propagate_constants_accross_call, devirtualization_time_bonus,
19585 ipcp_propagate_stage): Update.
19586 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
19587 * ipa-inline-transform.c (clone_inlined_nodes,
19588 preserve_function_body_p): Update.
19589 * ipa-reference.c (propagate): Update.
19590 (write_node_summary_p): Update.
19591 * toplev.c (wrapup_global_declaration_2): Update.
19592 * cgraphunit.c (cgraph_analyze_function): Rename to ...
19593 (analyze_function) ... this one.
19594 (cgraph_process_new_functions): Update.
19595 (cgraph_reset_node): Export.
19596 (cgraph_finalize_function): Update.
19597 (cgraph_add_new_function): Update.
19598 (process_function_and_variable_attributes): Update.
19599 (varpool_finalize_decl): Update.
19600 (symbol_finalized): Remove.
19601 (symbol_finalized_and_needed): Rename to ...
19602 (symbol_defined_and_needed): ... update.
19603 (cgraph_analyze_functions): Update.
19604 (handle_alias_pairs): Update.
19605 (mark_functions_to_output): Update.
19606 (assemble_thunk): Update.
19607 (output_in_order): Update.
19608 (output_weakrefs): Update.
19609 (finalize_compilation_unit): Update.
19610 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
19611 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
19612 input_node, input_varpool_node): Update.
19613 * dbxout.c (dbxout_expand_expr): Update.
19614 * cgraphclones.c (cgraph_clone_node): Update.
19615 (cgraph_copy_node_for_versioning): Update.
19616 (cgraph_materialize_clone): Update.
19617 (cgraph_materialize_all_clones): Update.
19618 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
19619 propagate_pure_const, propagate_nothrow): Update.
19620 * lto-streamer-out.c (lto_output, write_symbol): Update.
19621 * ipa-utils.c (ipa_reverse_postorder): Update.
19622 * ipa-inline.c (can_inline_edge_p): Update.
19623 (update_caller_keys, ipa_inline): Update.
19624 * dwarf2out.c (reference_to_unused,
19625 premark_types_used_by_global_vars_helper): Update.
19626 * tree-eh.c (tree_could_trap_p): Update.
19627 * ipa-split.c (consider_split, execute_split_functions): Update.
19628 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
19629 has_addr_references_p): Update; move ahead in file for better
19631 (process_references): Simplify.
19632 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
19633 bodies are removed.
19634 (cgraph_comdat_can_be_unshared_p): Make static.
19635 (cgraph_externally_visible_p): Update.
19636 (varpool_externally_visible_p): Update.
19637 (function_and_variable_visibility): Update.
19638 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
19639 ipa_tm_mark_force_output_node): Update.
19640 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
19641 estimate_edge_devirt_benefit, inline_generate_summary,
19642 inline_write_summary): Update.
19643 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
19644 * ipa-prop.c (ipa_compute_jump_functions): Update.
19645 (ipa_print_node_params, ipa_prop_read_section,
19646 ipa_update_after_lto_read, read_replacements_section): Update.
19647 * varasm.c (mark_decl_referenced): Update.
19648 (assemble_alias, dump_tm_clone_pairs): Update.
19649 * tree-inline.c (copy_bb): Update.
19650 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
19652 * symtab.c (dump_symtab_base): Print new flags.
19653 (verify_symtab_base): Verify new flags.
19654 (symtab_alias_ultimate_target): New function.
19655 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
19656 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
19658 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
19660 * i386.c (ix86_get_function_versions_dispatcher,
19661 ix86_generate_version_dispatcher_body): Update.
19662 (fold_builtin_cpu): Use varpool_add_new_variable.
19663 * varpool.c (varpool_remove_initializer): Break out from ...
19664 (varpool_remove_node): ... this one.
19665 (dump_varpool_node, varpool_node_for_asm,
19666 cgraph_variable_initializer_availability, varpool_analyze_node,
19667 varpool_assemble_decl, varpool_remove_unreferenced_decls,
19668 varpool_finalize_named_section_flags, varpool_create_variable_alias):
19671 2013-05-29 Jan Hubicka <jh@suse.cz>
19673 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
19675 2013-05-29 Easwaran Raman <eraman@google.com>
19677 PR tree-optimization/57442
19678 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
19679 when control exits the main loop.
19681 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
19683 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
19685 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
19686 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
19687 * rx/t-rx: Add rx100 under multi library matches option for nofpu
19690 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19692 PR tree-optimization/57441
19693 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
19694 Don't limit size of incr_vec to number of candidates.
19696 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
19698 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
19699 and mips16 directories.
19700 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
19701 (MULTILIB_DIRNAMES): Ditto.
19702 (MULTILIB_EXCEPTIONS): Add new exceptions.
19703 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
19704 (MULTILIB_DIRNAMES): Ditto.
19705 (MULTILIB_EXCEPTIONS): Add new exceptions.
19707 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19708 Marcus Shawcroft <marcus.shawcroft@arm.com>
19710 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
19711 SYMBOL_TINY_ABSOLUTE.
19712 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
19713 SYMBOL_TINY_ABSOLUTE.
19714 (aarch64_expand_mov_immediate): Likewise.
19715 (aarch64_classify_symbol): Likewise.
19716 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
19717 Permit SYMBOL_TINY_ABSOLUTE.
19718 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
19720 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19721 Marcus Shawcroft <marcus.shawcroft@arm.com>
19723 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
19724 Refactor if/switch. Replace gcc_assert with if.
19726 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19728 * config/i386/i386.c (initial_ix86_tune_features): Enable
19729 FP Reassociation for AMD bdver1 and bdver2.
19731 2013-05-29 Martin Jambor <mjambor@suse.cz>
19733 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
19734 and IMAGPART_EXPR do not occur within other handled_components.
19736 2013-05-29 Richard Biener <rguenther@suse.de>
19738 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
19739 access on whether the use is in the BB we currently try to
19741 (vect_bb_vectorization_profitable_p): Pass the BB we currently
19742 vectorize to vect_bb_slp_scalar_cost.
19744 2013-05-29 Richard Biener <rguenther@suse.de>
19746 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
19747 computing scalar cost offsetted by stmts that are kept live
19749 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
19750 for computation of scalar cost.
19752 2013-05-28 Steve Ellcey <sellcey@mips.com>
19754 * config/mips/mips-cpus.def (mips32r2): Change processor type.
19756 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
19758 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
19759 array notation built-in reduction functions.
19760 * doc/passes.texi (Passes): Added documentation about changes done
19762 * doc/invoke.texi (C Dialect Options): Added documentation about
19763 the -fcilkplus flag.
19764 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
19765 (BUILTINS_DEF): Depend on cilkplus.def.
19766 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
19767 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
19768 * cilkplus.def: New file.
19770 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
19772 PR rtl-optimization/57439
19773 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
19775 2013-05-28 Easwaran Raman <eraman@google.com>
19777 PR tree-optimization/57337
19778 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
19779 (find_insert_point): Correctly identify the insertion point
19780 when two statements with the same UID is compared.
19782 2013-05-28 Richard Biener <rguenther@suse.de>
19784 PR tree-optimization/56787
19785 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
19786 from the list of data references.
19787 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
19789 (vect_analyze_loop_operations): Likewise.
19790 (vect_transform_loop): Remove clobbers.
19792 2013-05-28 Martin Jambor <mjambor@suse.cz>
19794 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
19795 and REALPART_EXPRs have scalar type.
19797 2013-05-28 Richard Biener <rguenther@suse.de>
19799 PR tree-optimization/57411
19800 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
19802 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
19803 virtual operand propagation.
19805 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19807 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
19808 destination register for bmasksi_vis.
19809 (vector_init_bshuffle): Likewise.
19810 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
19812 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19814 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
19815 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
19816 mode if the instruction isn't available in the original mode.
19817 * config/sparc/sparc.opt (mfix-ut699): New option.
19818 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
19819 (divdf3): Turn into expander.
19820 (divdf3_nofix): New insn.
19821 (divdf3_fix): Likewise.
19822 (divsf3): Disable if -mfix-ut699.
19823 (sqrtdf2): Turn into expander.
19824 (sqrtdf2_nofix): New insn.
19825 (sqrtdf2_fix): Likewise.
19826 (sqrtsf2): Disable if -mfix-ut699.
19828 2013-05-27 Richard Biener <rguenther@suse.de>
19830 PR middle-end/57412
19831 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
19832 block for the new loop.
19834 2013-05-27 Richard Biener <rguenther@suse.de>
19836 PR tree-optimization/57343
19837 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
19838 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
19839 (number_of_iterations_cond): Do not build the folded tree.
19841 2013-05-27 Richard Biener <rguenther@suse.de>
19844 PR middle-end/57381
19845 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19846 OEP_CONSTANT_ADDRESS_OF retained.
19848 PR tree-optimization/57417
19849 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
19850 for unchanged base.
19851 (set_ssa_val_to): Compare addresses using
19852 get_addr_base_and_unit_offset.
19854 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19856 PR rtl-optimization/56833
19857 * postreload.c (move2add_record_mode): New function.
19858 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19859 (move2add_use_add2_insn): Use move2add_record_sym_value.
19860 (move2add_use_add3_insn): Likewise.
19861 (reload_cse_move2add): Use move2add_valid_value_p and
19862 move2add_record_mode. Invalidate call-clobbered and REG_INC
19863 affected regs by setting reg_mode to VOIDmode.
19864 (move2add_note_store): Don't pretend the inside of a SUBREG is
19865 the actual destination. Invalidate single/leading registers by
19866 setting reg_mode to VOIDmode.
19867 Use move2add_record_sym_value, move2add_valid_value_p and
19868 move2add_record_mode.
19870 2013-05-27 Richard Biener <rguenther@suse.de>
19872 PR tree-optimization/57396
19873 * tree-affine.c (double_int_constant_multiple_p): Properly
19874 return false for val == 0 and div != 0.
19876 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19878 * config/mips/mips.h: Use #elif in preprocessor conditions.
19880 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19883 * config/mips/constraints.md (kl): New constraint.
19884 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19885 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19886 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19887 from LO for MIPS16.
19888 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19890 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19893 * config/mips/mips.c (mips_can_inline_p): New function.
19894 (TARGET_CAN_INLINE_P): Define.
19896 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19898 * sched-int.h (ds_t, dw_t): Make unsigned int.
19899 Fix documentation that describes how all the ds_t bits are used.
19900 Reserve the last bit for delayed-branch scheduling.
19901 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19902 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19903 (gen_dep_weak_1): Remove prototype.
19904 * sched-deps.c (get_dep_weak_1): Make static.
19905 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19906 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19907 * doc/tm.texi: Regenerate.
19908 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19910 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19913 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19915 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19916 Sandra Loosemore <sandra@codesourcery.com>
19918 * config.gcc (powerpc-*): Allow native for with-cpu.
19920 2013-05-24 Jeff Law <law@redhat.com>
19922 PR tree-optimization/57124
19923 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19924 conversion feeding a condition if the range has an overflow
19925 if -fstrict-overflow. Add warnings for when we do make the
19928 2013-05-24 Dehao Chen <dehao@google.com>
19930 * tree-cfg.c (locus_discrim_map): Fix the typo.
19931 (locus_discrim_hasher): Likewise.
19932 (locus_discrim_hasher::hash): Likewise.
19933 (locus_discrim_hasher::equal): Likewise.
19935 2013-05-24 Martin Jambor <mjambor@suse.cz>
19937 PR tree-optimization/57294
19938 * cgraph.h (ipa_record_stmt_references): Declare.
19939 * cgraphbuild.c (ipa_record_stmt_references): New function.
19940 (build_cgraph_edges): Use ipa_record_stmt_references.
19941 (rebuild_cgraph_edges): Likewise.
19942 (cgraph_rebuild_references): Likewise.
19943 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19944 associated with the old statement and build references from the
19945 newly built statements.
19946 * ipa-ref.c (ipa_remove_stmt_references): New function.
19947 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19949 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19951 * lra-constraints.c (emit_spill_move): Use smaller mode for
19953 (check_and_process_move): Consider mem-reg moves for secondary
19955 (curr_insn_transform): Don't lose insns emitted before for
19956 secondary memory moves.
19957 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19958 reg set up in the current insn.
19960 2013-05-24 Dehao Chen <dehao@google.com>
19962 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19964 (locus_descrim_hasher::equal): Likewise.
19965 (build_gimple_cfg): New discriminator assignment algorithm.
19966 (make_edges): Likewise.
19967 (next_discriminator_for_locus): Likewise.
19968 (same_line_p): Likewise.
19969 (assign_discriminators): Likewise.
19970 (make_cond_expr_edges): Likewise.
19971 (make_gimple_switch_edges): Likewise.
19972 (make_goto_expr_edges): Likewise.
19973 (make_gimple_asm_edges): Likewise.
19975 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19977 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19978 X format specifier to only display bottom 16 bits.
19979 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19980 immediate to match for operand 2, since it will be masked.
19982 2013-05-24 Richard Biener <rguenther@suse.de>
19984 PR tree-optimization/57287
19985 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19986 all SSA names that occur in abnormal PHIs.
19988 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19991 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19992 that index is not negative.
19994 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19996 PR rtl-optimization/55177
19997 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19998 (simplify_byte_swapping_operation): New.
19999 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
20000 (simplify_relational_operation_1): Deal with BSWAP.
20002 2013-05-23 Richard Henderson <rth@redhat.com>
20005 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
20006 (ix86_reorg): Call it.
20008 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
20011 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
20012 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
20013 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
20015 2013-05-23 Christian Bruel <christian.bruel@st.com>
20018 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
20020 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
20021 Marcus Shawcroft <marcus.shawcroft@arm.com>
20023 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
20024 * config/aarch64/constraints.md (Usa): Remove.
20025 * doc/md.texi (AArch64 Usa): Remove.
20027 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
20028 Marcus Shawcroft <marcus.shawcroft@arm.com>
20030 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
20031 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
20032 * config/aarch64/predicates.md (aarch64_const_address): Remove.
20033 (aarch64_mov_operand): Use aarch64_mov_operand_p.
20035 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
20037 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
20038 instruction (AdvSIMD).
20039 * config/aarch64/aarch64-builtins.c
20040 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
20041 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
20043 2013-05-23 Martin Jambor <mjambor@suse.cz>
20045 PR middle-end/57347
20046 * tree.h (contains_bitfld_component_ref_p): Declare.
20047 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
20048 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
20050 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
20051 not access a bit-field. Assert all final offsets are byte-aligned.
20053 2013-05-23 Richard Biener <rguenther@suse.de>
20055 PR tree-optimization/57380
20056 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
20057 least one invariant or re-used load.
20058 * passes.c (init_optimization_passes): Move pass_phiprop before
20061 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
20063 * config/aarch64/aarch64-simd.md
20064 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
20066 2013-05-23 Richard Biener <rguenther@suse.de>
20068 PR middle-end/57381
20069 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
20070 OEP_CONSTANT_ADDRESS_OF retained.
20072 2013-05-23 Jakub Jelinek <jakub@redhat.com>
20074 PR middle-end/57344
20075 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
20076 don't lower unit. Handle unit not being always BITS_PER_WORD.
20078 2013-05-23 Richard Biener <rguenther@suse.de>
20080 PR rtl-optimization/57341
20081 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
20082 instead of true_dependence.
20084 2013-05-22 David Malcolm <dmalcolm@redhat.com>
20086 * bb-reorder.c (branch_threshold): Make const.
20087 (exec_threshold): Ditto.
20089 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
20090 Pat Haugen <pthaugen@us.ibm.com>
20091 Peter Bergner <bergner@vnet.ibm.com>
20093 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
20094 documentation for the power8 crypto builtins.
20096 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
20098 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
20099 macros for defining power8 builtin functions.
20100 (BU_P8V_AV_2): Likewise.
20101 (BU_P8V_AV_P): Likewise.
20102 (BU_P8V_VSX_1): Likewise.
20103 (BU_P8V_OVERLOAD_1): Likewise.
20104 (BU_P8V_OVERLOAD_2): Likewise.
20105 (BU_CRYPTO_1): Likewise.
20106 (BU_CRYPTO_2): Likewise.
20107 (BU_CRYPTO_3): Likewise.
20108 (BU_CRYPTO_OVERLOAD_1): Likewise.
20109 (BU_CRYPTO_OVERLOAD_2): Likewise.
20110 (XSCVSPDP): Fix typo, point to the correct instruction.
20111 (VCIPHER): Add power8 crypto builtins.
20112 (VCIPHERLAST): Likewise.
20113 (VNCIPHER): Likewise.
20114 (VNCIPHERLAST): Likewise.
20115 (VPMSUMB): Likewise.
20116 (VPMSUMH): Likewise.
20117 (VPMSUMW): Likewise.
20118 (VPERMXOR_V2DI): Likewise.
20119 (VPERMXOR_V4SI: Likewise.
20120 (VPERMXOR_V8HI: Likewise.
20121 (VPERMXOR_V16QI: Likewise.
20122 (VSHASIGMAW): Likewise.
20123 (VSHASIGMAD): Likewise.
20124 (VPMSUM): Likewise.
20125 (VPERMXOR): Likewise.
20126 (VSHASIGMA): Likewise.
20128 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
20129 __CRYPTO__ if the crypto instructions are available.
20130 (altivec_overloaded_builtins): Add support for overloaded power8
20133 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
20134 support for power8 crypto builtins.
20135 (builtin_function_type): Likewise.
20136 (altivec_init_builtins): Add support for builtins that take vector
20137 long long (V2DI) arguments.
20139 * config/rs6000/crypto.md: New file, define power8 crypto
20142 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
20143 Pat Haugen <pthaugen@us.ibm.com>
20144 Peter Bergner <bergner@vnet.ibm.com>
20146 * doc/invoke.texi (Option Summary): Add power8 options.
20147 (RS/6000 and PowerPC Options): Likewise.
20149 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
20150 constraints.md instead of rs6000.h. Reorder w* constraints. Add
20151 wm, wn, wr documentation.
20153 * config/rs6000/constraints.md (wm): New constraint for VSX
20154 registers if direct move instructions are enabled.
20155 (wn): New constraint for no registers.
20156 (wq): New constraint for quad word even GPR registers.
20157 (wr): New constraint if 64-bit instructions are enabled.
20158 (wv): New constraint if power8 vector instructions are enabled.
20159 (wQ): New constraint for quad word memory locations.
20161 * config/rs6000/predicates.md (const_0_to_15_operand): New
20162 constraint for 0..15 for crypto instructions.
20163 (gpc_reg_operand): If VSX allow registers in VSX registers as well
20164 as GPR and floating point registers.
20165 (int_reg_operand): New predicate to match only GPR registers.
20166 (base_reg_operand): New predicate to match base registers.
20167 (quad_int_reg_operand): New predicate to match even GPR registers
20168 for quad memory operations.
20169 (vsx_reg_or_cint_operand): New predicate to allow vector logical
20170 operations in both GPR and VSX registers.
20171 (quad_memory_operand): New predicate for quad memory operations.
20172 (reg_or_indexed_operand): New predicate for direct move support.
20174 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
20175 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
20176 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
20177 (POWERPC_MASKS): Add power8 options.
20178 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
20181 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
20182 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
20184 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
20185 (-mpower8-fusion): New power8 options.
20186 (-mpower8-fusion-sign): Likewise.
20187 (-mpower8-vector): Likewise.
20188 (-mcrypto): Likewise.
20189 (-mdirect-move): Likewise.
20190 (-mquad-memory): Likewise.
20192 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
20193 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
20195 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
20196 (rs6000_debug_vector_unit): Add p8_vector.
20197 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
20198 definitions. Also print fusion state.
20199 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
20200 (rs6000_builtin_mask_calculate): Add power8 builtin support.
20201 (rs6000_option_override_internal): Add support for power8.
20202 (rs6000_common_init_builtins): Add debugging for skipped builtins
20203 if -mdebug=builtin.
20204 (rs6000_adjust_cost): Add power8 support.
20205 (rs6000_issue_rate): Likewise.
20206 (insn_must_be_first_in_group): Likewise.
20207 (insn_must_be_last_in_group): Likewise.
20208 (force_new_group): Likewise.
20209 (rs6000_register_move_cost): Likewise.
20210 (rs6000_opt_masks): Likewise.
20212 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
20213 power8 capable assembler, default to power7 options.
20214 (TARGET_DIRECT_MOVE): Likewise.
20215 (TARGET_CRYPTO): Likewise.
20216 (TARGET_P8_VECTOR): Likewise.
20217 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
20218 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
20219 (VECTOR_MEM_P8_VECTOR_P): Likewise.
20220 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
20221 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
20222 (TARGET_XSCVDPSPN): Likewise.
20223 (TARGET_XSCVSPDPN): Likewsie.
20224 (TARGET_SYNC_HI_QI): Likewise.
20225 (TARGET_SYNC_TI): Likewise.
20226 (MASK_CRYPTO): Likewise.
20227 (MASK_DIRECT_MOVE): Likewise.
20228 (MASK_P8_FUSION): Likewise.
20229 (MASK_P8_VECTOR): Likewise.
20230 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
20231 temporary used by some of the direct move instructions to get two FP
20232 temporary registers does not force creation of a stack frame.
20233 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
20234 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
20235 that any VSX registers are tieable, even if they are also an
20236 Altivec vector mode.
20237 (r6000_reg_class_enum): Add wm, wr, wv constraints.
20238 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
20239 (RS6000_BTM_CRYPTO): Likewise.
20240 (RS6000_BTM_COMMON): Likewise.
20242 * config/rs6000/rs6000.md (cpu attribute): Add power8.
20243 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
20244 (enum rs6000_vector): Add power8 vector support.
20246 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20250 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
20251 (any_sibcall_could_use_r3): this and handle indirect calls.
20252 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
20254 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20256 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
20258 2013-05-22 Richard Biener <rguenther@suse.de>
20260 PR middle-end/57349
20261 * profile.c (branch_prob): Do not split blocks that are
20262 abnormally receiving from ECF_RETURNS_TWICE functions.
20264 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
20266 * recog.c (offsettable_address_addr_space_p): Fix calculation of
20267 address mode. Move pointer mode initialization to the same place.
20269 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
20271 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
20272 while it has any effect.
20274 2013-05-21 Easwaran Raman <eraman@google.com>
20276 PR tree-optimization/57322
20277 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
20278 UID of the statement added to the BB to be 1.
20280 2013-05-21 Jakub Jelinek <jakub@redhat.com>
20282 PR tree-optimization/57331
20283 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
20284 of conversion from pointer type to integral type with integer.
20286 2013-05-21 Martin Jambor <mjambor@suse.cz>
20289 * ipa-prop.c (ipa_read_node_info): Process param_used and
20290 controlled_uses in the same order as when writing.
20292 2013-05-21 Magnus Granberg <baldrick@free.fr>
20295 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
20297 2013-05-21 Richard Biener <rguenther@suse.de>
20299 PR tree-optimization/57318
20300 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
20301 estimate stmts with side-effects as likely eliminated.
20303 2013-05-21 Richard Biener <rguenther@suse.de>
20305 PR tree-optimization/57330
20306 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
20307 preserve the call stmts fntype.
20309 2013-05-21 Richard Biener <rguenther@suse.de>
20311 PR tree-optimization/57303
20312 * tree-ssa-sink.c (statement_sink_location): Improve killing
20313 stmt detection and properly handle self-assignments.
20315 2013-05-21 Christian Bruel <christian.bruel@st.com>
20317 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
20318 spanning registers. LEAF_REG_REMAP is supported only for contiguous
20319 registers. Set register size out of the PARALLEL loop.
20321 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
20324 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
20325 (*fmasf4, *fmasf4_media): New insns.
20327 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20329 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
20330 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
20331 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
20332 (mips_idiv_insns): Update the comments to say that the returned
20333 instruction counts are in units of BASE_INSN_LENGTH.
20334 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
20335 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
20336 using 2 rather than 4 as the length of indirect MIPS16 and
20337 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
20338 length of a NOP. Don't divide MIPS16 lengths by 2.
20339 (mips16_split_long_branches): Assume a branch is long if the
20340 length is greater than 4 rather than 8.
20341 * config/mips/mips.md (length): Give MIPS16 lengths directly,
20342 rather than multiplying them by 2. Multiply instruction counts
20343 by BASE_INSN_LENGTH rather than 4.
20344 (*jump_mips16, tls_get_tp_mips16_<mode>)
20345 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
20347 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20349 * config/mips/mips.md (extended_mips16): Remove branch case.
20350 (length): Remove duplicated extended_mips16 test.
20352 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20354 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
20356 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
20358 * recog.h (Recog_data): Rename to...
20359 (recog_data_d): ...this.
20360 (recog_data): Update accordingly.
20361 * recog.c (recog_data): Likewise.
20362 * reload.c (save_recog_data): Likewise.
20363 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
20364 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
20366 2013-05-17 Julian Brown <julian@codesourcery.com>
20368 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
20369 found in a REG_EQUAL note, invalidate it.
20371 2013-05-17 Easwaran Raman <eraman@google.com>
20373 * tree-ssa-reassoc.c (find_insert_point): New function.
20374 (insert_stmt_after): Likewise.
20375 (get_def_stmt): Likewise.
20376 (ensure_ops_are_available): Likewise.
20377 (not_dominated_by): Likewise.
20378 (rewrite_expr_tree): Do not move statements beyond what is
20379 necessary. Remove call to swap_ops_for_binary_stmt...
20380 (reassociate_bb): ... and move it here.
20381 (build_and_add_sum): Assign UIDs for new statements.
20382 (linearize_expr): Likewise.
20383 (do_reassoc): Renumber gimple statement UIDs.
20385 2013-05-17 Jan Hubicka <jh@suse.cz>
20387 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
20389 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
20390 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
20392 (output_weakrefs): Update.
20394 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
20395 Martin Jambor <mjambor@suse.cz>
20397 PR middle-end/57276
20398 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
20399 value that corresponds to the given aggval is found in values vector.
20401 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
20403 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
20404 sse, sse2, sse3, ssse3 and sse4a flags to options.
20406 2013-05-17 David Malcolm <dmalcolm@redhat.com>
20408 * gengtype-state.c: (s_expr_writer): New class, to handle
20409 prettifying of output layout of s-expressions.
20410 (state_writer): New class, to write out gtype.state.
20411 (state_written_type_count): Move this variable into member data of
20413 (s_expr_writer::s_expr_writer): New code: constructor for new class
20414 (state_writer::state_writer(): ditto
20415 (s_expr_writer::write_new_line): New function
20416 (s_expr_writer::write_any_indent): ditto
20417 (s_expr_writer::begin_s_expr): ditto
20418 (s_expr_writer::end_s_expr): ditto
20419 (write_state_fileloc): convert to method of state_writer...
20420 (state_writer:: write_state_fileloc): ...and use methods of
20421 s_expr_writer to write indentation into the gtype.state output file
20422 to visually represent the hierarchical structure of the list
20424 (write_state_fields): ditto, renaming to...
20425 (state_writer::write_state_fields)
20426 (write_state_a_string): ditto, renaming to...
20427 (state_writer::write_state_a_string)
20428 (write_state_string_option): ditto, renaming to...
20429 (state_writer::write_state_string_option)
20430 (write_state_type_option): ditto, renaming to...
20431 (state_writer::write_state_type_option)
20432 (write_state_nested_option): ditto, renaming to...
20433 (state_writer::write_state_nested_option)
20434 (write_state_option): ditto, renaming to...
20435 (state_writer::write_state_option)
20436 (write_state_options): ditto, renaming to...
20437 (state_writer::write_state_options)
20438 (write_state_lang_bitmap): ditto, renaming to...
20439 (state_writer::write_state_lang_bitmap)
20440 (write_state_version): ditto, renaming to...
20441 (state_writer::write_state_version)
20442 (write_state_scalar_type): ditto, renaming to...
20443 (state_writer::write_state_scalar_type)
20444 (write_state_string_type): ditto, renaming to...
20445 (state_writer::write_state_string_type)
20446 (write_state_undefined_type): ditto, renaming to...
20447 (state_writer::write_state_undefined_type)
20448 (write_state_struct_union_type): ditto, renaming to...
20449 (state_writer::write_state_struct_union_type)
20450 (write_state_struct_type): ditto, renaming to...
20451 (state_writer::write_state_struct_type)
20452 (write_state_user_struct_type): ditto, renaming to...
20453 (state_writer::write_state_user_struct_type)
20454 (write_state_lang_struct_type): ditto, renaming to...
20455 (state_writer::write_state_lang_struct_type)
20456 (write_state_param_struct_type): ditto, renaming to...
20457 (state_writer::write_state_param_struct_type)
20458 (write_state_pointer_type): ditto, renaming to...
20459 (state_writer::write_state_pointer_type)
20460 (write_state_array_type): ditto, renaming to...
20461 (state_writer::write_state_array_type)
20462 (write_state_gc_used): ditto, renaming to...
20463 (state_writer::write_state_gc_used)
20464 (write_state_common_type_content): ditto, renaming to...
20465 (state_writer::write_state_common_type_content)
20466 (write_state_type): ditto, renaming to...
20467 (state_writer::write_state_type)
20468 (write_state_pair_list): ditto, renaming to...
20469 (state_writer::write_state_pair_list)
20470 (write_state_pair): ditto, renaming to...
20471 (state_writer::write_state_pair)
20472 (write_state_typedefs): ditto, renaming to...
20473 (state_writer::write_state_typedefs)
20474 (write_state_structures): ditto, renaming to...
20475 (state_writer::write_state_structures)
20476 (write_state_param_structs): ditto, renaming to...
20477 (state_writer::write_state_param_structs)
20478 (write_state_variables): ditto, renaming to...
20479 (state_writer::write_state_variables)
20480 (write_state_srcdir): ditto, renaming to...
20481 (state_writer::write_state_srcdir)
20482 (write_state_files_list): ditto, renaming to...
20483 (state_writer::write_state_files_list)
20484 (write_state_languages): ditto, renaming to...
20485 (state_writer::write_state_languages)
20486 (write_state): create a state_writer instance and use it when
20487 writing out the state file
20489 2013-05-17 Mike Stump <mikestump@comcast.net>
20491 PR rtl-optimization/57304
20492 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
20493 accessing DF_REF_REAL_LOC.
20495 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20497 PR rtl-optimization/57281
20498 PR rtl-optimization/57300
20499 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
20500 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
20501 what the other splitter did if the registers are dead.
20503 2013-05-17 Richard Biener <rguenther@suse.de>
20505 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
20508 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20510 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
20513 2013-05-17 Marek Polacek <polacek@redhat.com>
20515 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
20516 length when doing non-zero store of storing '\0' to '\0'.
20518 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20520 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
20521 vect_external_def oprnd1 with loop_vinfo, try to emit
20522 optional cast, negation and and stmts on the loop preheader
20523 edge instead of into the pattern def seq.
20525 PR tree-optimization/57051
20526 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
20527 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
20529 2013-05-16 Nick Clifton <nickc@redhat.com>
20531 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
20532 (rl78_is_naked_func): New function.
20533 (rl78_expand_prologue): Skip prologue generation for naked functions.
20534 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
20535 * doc/extend.texi (naked): Add RL78 to the list of processors
20536 that supports this attribute.
20538 2013-05-16 Jeff Law <law@redhat.com>
20540 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
20542 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20544 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
20545 cache parameters using detect_caches_amd also for CYRIX,
20546 NSC and TM2 signatures.
20548 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20549 Dzianis Kahanovich <mahatma@eu.by>
20553 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
20554 VIA/Centaur processors and determine their cache parameters
20555 using detect_caches_amd.
20557 2013-05-16 Teresa Johnson <tejohnson@google.com>
20559 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
20560 (rtl_verify_edges): New function.
20561 (rtl_verify_bb_insns): Ditto.
20562 (rtl_verify_bb_pointers): Ditto.
20563 (rtl_verify_bb_insn_chain): Ditto.
20564 (rtl_verify_fallthru): Ditto.
20565 (rtl_verify_bb_layout): Ditto.
20566 (rtl_verify_flow_info_1): Outline checks into new functions.
20567 (rtl_verify_flow_info): Ditto.
20569 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
20571 * cfghooks.c (copy_bbs): Add update_dominance argument.
20572 * cfghooks.h (copy_bbs): Update prototype.
20573 * tree-cfg.c (gimple_duplicate_sese_region):
20574 Add update_dominance argument.
20575 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
20576 * tree-ssa-loop-ch.c (copy_loop_headers): Update
20577 gimple_duplicate_sese_region call.
20578 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
20579 Update copy_bbs call.
20580 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
20581 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
20583 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20585 * tree-vectorizer.h (NUM_PATTERNS): Increment.
20586 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
20587 vect_recog_rotate_pattern.
20588 (vect_recog_rotate_pattern): New function.
20590 2013-05-16 Jason Merrill <jason@redhat.com>
20592 * Makefile.in (LLINKER): New variable.
20593 (mostlyclean): Remove link mutex.
20594 * configure.ac: Handle --enable-link-mutex.
20595 * lock-and-run.sh: New script.
20597 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20600 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
20603 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20605 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
20607 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
20609 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
20610 * config/arm/arm.c (next_consecutive_mem): New function.
20611 (gen_movmem_ldrd_strd): Likewise.
20612 * config/arm/arm.md (movmemqi): Update condition and code.
20613 (unaligned_loaddi, unaligned_storedi): New patterns.
20615 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20617 * config.gcc: Obsolete *-*-solaris2.9*.
20618 * doc/install.texi (Specific, *-*-solaris2*): Document it.
20620 2013-05-16 Richard Biener <rguenther@suse.de>
20622 * passes.c (init_optimization_passes): Move pass_parallelize_loops
20623 earlier, after GRAPHITE transforms and IV canonicalization.
20625 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20627 * omp-low.c (extract_omp_for_data): For collapsed loops,
20628 if at least one of the loops is known at compile time to
20629 iterate zero times, set count to 0.
20630 (expand_omp_regimplify_p): New function.
20631 (expand_omp_for_generic): For collapsed loops, if at least
20632 one of the loops isn't known to iterate at least once,
20633 add runtime check with setting count to 0.
20634 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
20635 For unsigned types if it isn't known at compile time that
20636 the loop will iterate at least once, add runtime check to bypass
20637 the whole loop if initial condition isn't true.
20639 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
20641 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
20643 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
20645 PR middle-end/57286
20646 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
20647 transformations to avoid an infinite loop.
20649 2013-05-16 Marek Polacek <polacek@redhat.com>
20651 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
20653 2013-05-15 Leif Ekblad <leif@rdos.net>
20655 * config/i386/i386.c (ix86_decompose_address): Use
20656 DEFAULT_TLS_SEG_REG to access TLS segment register.
20657 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
20658 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
20659 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
20661 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
20664 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
20665 sibling calls to functions that would normally be lazily bound,
20666 unless $gp is call-clobbered.
20668 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
20670 * config/i386/i386.c (ix86_option_override_internal): Update
20671 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
20672 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
20673 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
20674 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
20676 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
20678 2013-05-15 Andreas Schwab <schwab@suse.de>
20680 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
20681 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
20684 2013-05-15 Teresa Johnson <tejohnson@google.com>
20686 * loop-unroll.c (report_unroll_peel): Check decision before
20687 emitting unroll/peel message.
20689 2013-05-15 Teresa Johnson <tejohnson@google.com>
20691 * function.h (has_bb_partition): New rtl_data flag.
20692 (bb_reorder_complete): Ditto.
20693 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
20694 instead of flag_reorder_blocks_and_partition.
20695 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
20696 with some enhancements.
20697 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
20698 * bb-reorder.c (connect_traces): Check for has_bb_partition
20699 instead of flag_reorder_blocks_and_partition.
20700 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
20701 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
20702 verify_hot_cold_block_grouping.
20703 (partition_hot_cold_basic_blocks): Set has_bb_partition.
20705 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20708 * config/arm/predicates.md (call_insn_operand): New predicate.
20709 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
20710 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
20711 if insn is not a tail call.
20712 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
20714 * config/arm/arm.h (enum reg_class): New caller save register class.
20715 (REG_CLASS_NAMES): Likewise.
20716 (REG_CLASS_CONTENTS): Likewise.
20717 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
20720 2013-05-15 Richard Biener <rguenther@suse.de>
20722 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
20723 of MSG_OPTIMIZED_LOCATIONS.
20724 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
20725 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
20727 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
20728 of MSG_OPTIMIZED_LOCATIONS.
20729 (execute_vect_slp): Likewise.
20730 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
20731 (vect_create_cond_for_alias_checks): Likewise.
20732 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
20733 (vect_recog_widen_mult_pattern): Likewise.
20734 (vect_recog_widen_sum_pattern): Likewise.
20735 (vect_recog_over_widening_pattern): Likewise.
20736 (vect_recog_widen_shift_pattern): Likewise.
20737 (vect_recog_vector_vector_shift_pattern): Likewise.
20738 (vect_recog_divmod_pattern): Likewise.
20739 (vect_recog_mixed_size_cond_pattern): Likewise.
20740 (vect_recog_bool_pattern): Likewise.
20741 (vect_pattern_recog_1): Likewise.
20743 2013-05-15 Martin Jambor <mjambor@suse.cz>
20745 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
20746 non-functions to builtin_unreachable.
20747 * ipa-inline-transform.c (inline_call): Do not assert estimates were
20748 correct when new direct edges were discovered.
20750 2013-05-15 Martin Jambor <mjambor@suse.cz>
20752 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
20753 header, print symbol order instead of node uid, print more information
20754 about indirect edge targets.
20755 (ipa_make_edge_direct_to_target): Print symbol order instead of node
20757 (ipa_make_edge_direct_to_target): Likewise.
20758 (remove_described_reference): Likewise.
20759 (propagate_controlled_uses): Likewise.
20760 (ipa_print_node_params): Also print symbol order.
20761 (ipcp_transform_function): Print symbol order instead of node uids.
20762 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
20763 (cgraph_get_create_real_symbol_node): Likewise.
20764 * ipa-cp.c (print_lattice): Likewise.
20765 (print_all_lattices): Likewise.
20766 (determine_versionability): Likewise.
20767 (initialize_node_lattices): Likewise.
20768 (estimate_local_effects): Likewise.
20769 (update_profiling_info): Likewise.
20770 (create_specialized_node): Likewise.
20771 (perhaps_add_new_callers): Likewise.
20772 (decide_about_value): Likewise.
20773 (decide_whether_version_node): Likewise.
20774 (identify_dead_nodes): Likewise.
20775 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
20776 (dump_inline_summary): Likewise.
20777 (estimate_node_size_and_time): Likewise.
20778 (inline_analyze_function): Likewise.
20779 * ipa-inline.c (report_inline_failed_reason): Likewise.
20780 (want_early_inline_function_p): Likewise.
20781 (edge_badness): Likewise.
20782 (update_edge_key): Likewise.
20783 (inline_small_functions): Likewise. Add dumping of order to two other
20785 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
20786 instead of node uids.
20787 (propagate_pure_const): Likewise.
20788 (propagate_pure_const): Likewise.
20789 * ipa-utils.c (dump_cgraph_node_set): Likewise.
20790 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
20791 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
20793 * tree-pretty-print.c (dump_function_header): Likewise.
20794 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
20795 Print symbol order instead of node uids.
20797 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20799 * config/s390/s390.c (s390_register_move_cost): Don't impose the
20800 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
20802 2013-05-15 Richard Biener <rguenther@suse.de>
20804 PR tree-optimization/57275
20805 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
20806 return value for fail to do runtime alias checks for gather loads.
20808 2013-05-15 Jan Hubicka <jh@suse.cz>
20812 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
20813 weakrefs are not external.
20814 (lto_symtab_merge_decls): Fix thinko when dealing with
20815 non-lto_symtab decls.
20816 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
20817 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
20818 * varpool.c (dump_varpool_node): Dump more flags.
20820 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20822 * config/i386/i386.c (processor_alias_table): Add instruction
20823 FSGSBASE for AMD bdver3 architecture.
20825 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20827 * tree.c (warn_deprecated_use): Print file:line using locus color.
20828 * diagnostic.c (diagnostic_report_current_module): Print file:line
20829 and file:line:column using locus color.
20831 2013-05-14 Mike Stump <mikestump@comcast.net>
20833 * gdbinit.in: Add __null.
20835 2013-05-14 Mike Stump <mikestump@comcast.net>
20837 * recog.h: Rename struct recog_data to Recog_data.
20838 * recog.c: Likewise.
20839 * reload.c (can_reload_into): Likewise.
20840 * config/picochip/picochip.c: Likewise.
20842 2013-05-14 Mike Stump <mikestump@comcast.net>
20844 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
20846 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
20848 * resource.h (struct resources): Remove unch_memory member.
20849 (CLEAR_RESOURCE): Don't clear unch_memory.
20850 * resource.c (mark_referenced_resources): Don't set it.
20851 (mark_set_resources): Likewise.
20852 (mark_target_live_regs): Don't clear it.
20853 (init_resource_info): Likewise.
20854 * reorg.c (resource_conflicts_p): Don't compare it.
20855 (redundant_insn): Don't set it.
20857 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20859 * emit-rtl.c (next_label): Remove unused function.
20860 (skip_consecutive_labels, link_cc0_insns): Move to ...
20861 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20862 only place where these functions are used, and make them static.
20864 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20866 * fold-const.c (fold_negate_expr): Handle vectors.
20867 (fold_truth_not_expr): Make it static.
20868 (fold_invert_truthvalue): New static function.
20869 (invert_truthvalue_loc): Handle vectors. Do not call
20870 fold_truth_not_expr directly.
20871 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20872 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20873 (fold_comparison): Handle vector constants.
20874 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20875 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20876 * tree.h (fold_truth_not_expr): Remove declaration.
20878 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20880 * config/aarch64/aarch64-simd.md
20881 (aarch64_vcond_internal<mode>): Rename to...
20882 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20883 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20884 float modes. Clarify all iterator modes.
20885 (vcond<mode><mode>): Use new name for vcond expanders.
20886 (vcond<v_cmp_result><mode>): Likewise.
20887 (vcondu<mode><mode>: Likewise.
20888 * config/aarch64/iterators.md (VDQF_COND): New.
20890 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20893 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20894 variable for the shift amount. Check that we shift by non-negative
20897 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20900 * config/arm/arm.h (EPILOGUE_USES): Only return true
20901 for LR_REGNUM after epilogue_completed.
20903 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20905 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20906 is error_mark_node.
20908 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20911 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20912 and Solaris 11+/x86 with gld.
20913 * configure: Regenerate.
20915 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20917 * expmed.c (expand_shift_1): Canonicalize rotates by
20918 constant bitsize / 2 to bitsize - 1.
20919 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20920 case ROTATERT>: Likewise.
20923 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20925 * config/i386/i386.md (rotateinv): New code attr.
20926 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20927 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20928 roll $31, %eax, etc.
20930 2013-05-14 Richard Biener <rguenther@suse.de>
20932 PR middle-end/57235
20933 * tree-eh.c (sink_clobbers): Give up for successors with
20934 multiple predecessors and no virtual uses.
20936 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20938 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20939 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20941 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20943 PR middle-end/57251
20944 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20945 the case when both op0 and op1 have VOIDmode.
20947 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20949 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20950 in multiply-accumulate mode.
20952 2013-05-13 Guozhi Wei <carrot@google.com>
20954 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20956 2013-05-13 Kai Tietz <ktietz@redhat.com>
20959 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20960 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20961 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20962 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20963 (ix86_expand_prologue): Likewise.
20964 (ix86_expand_split_stack_prologue): Likewise.
20965 (legitimate_pic_address_disp_p): Likewise.
20966 (legitimize_pic_address): Likewise.
20967 (legitimize_tls_address): Likewise.
20968 (legitimize_pe_coff_symbol): Likewise.
20969 (output_pic_addr_const): Likewise.
20970 (construct_plt_address): Likewise.
20971 (ix86_expand_call): Likewise.
20972 (x86_output_mi_thunk): Likewise.
20973 (x86_function_profiler): Likewise.
20975 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20977 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20978 similar switch cases.
20979 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20980 (aarch64_simd_mov_to_<mode>low): Delete.
20981 (aarch64_simd_mov_to_<mode>high): Delete.
20982 (move_lo_quad_<mode>): Add w<-r alternative.
20983 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20984 (aarch64_simd_mov_from_*): Update type attribute.
20985 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20988 2013-05-13 Jan Hubicka <jh@suse.cz>
20990 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20991 * config/i386/i386.c (ix86_compute_frame_layout,
20992 ix86_expand_epilogue, emit_i387_cw_initialization,
20993 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20994 ix86_local_alignment): Fix use of size/speed predicates.
20996 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20998 PR tree-optimization/45216
20999 PR tree-optimization/57157
21000 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
21001 the (-Y) & (B - 1) variant if OP is |.
21002 * expmed.c (expand_shift_1): For rotations by const0_rtx just
21003 return shifted. Use (-op1) & (prec - 1) as other_amount
21004 instead of prec - op1.
21006 2013-05-13 Martin Jambor <mjambor@suse.cz>
21008 PR middle-end/42371
21009 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
21010 (ipa_constant_data): New type.
21011 (ipa_jump_func): Use ipa_constant_data to hold information about
21012 constant jump functions.
21013 (ipa_get_jf_constant): Adjust to jump function type changes.
21014 (ipa_get_jf_constant_rdesc): New function.
21015 (ipa_param_descriptor): New field controlled_uses.
21016 (ipa_get_controlled_uses): New function.
21017 (ipa_set_controlled_uses): Likewise.
21018 * ipa-ref.h (ipa_find_reference): Declare.
21019 * ipa-prop.c (ipa_cst_ref_desc): New type.
21020 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
21022 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
21023 New parameter cs. Adjust all callers.
21024 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
21025 (remove_described_reference): New function.
21026 (jfunc_rdesc_usable): Likewise.
21027 (try_make_edge_direct_simple_call): Decrement controlled use count,
21028 attempt to remove reference if it hits zero.
21029 (combine_controlled_uses_counters): New function.
21030 (propagate_controlled_uses): Likewise.
21031 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
21032 (ipa_edge_duplication_hook): Duplicate reference descriptions.
21033 (ipa_print_node_params): Print described use counter.
21034 (ipa_write_jump_function): Adjust to jump function type changes.
21035 (ipa_read_jump_function): New parameter CS, pass it to
21036 ipa_set_jf_constant. Adjust caller.
21037 (ipa_write_node_info): Stream controlled use count
21038 (ipa_read_node_info): Likewise.
21039 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
21041 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
21042 count. Remove cloning-added reference if it reaches zero.
21043 * ipa-ref.c (ipa_find_reference): New function.
21045 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
21047 * config/i386/i386.c (processor_target_table): Modified default
21048 alignment values for AMD BD and BT architectures.
21050 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
21052 * tree-vect-generic.c (uniform_vector_p): Move ...
21053 * tree.c (uniform_vector_p): ... here.
21054 * tree.h (uniform_vector_p): Declare it.
21055 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
21058 2013-05-13 Jakub Jelinek <jakub@redhat.com>
21060 PR tree-optimization/57230
21061 * tree-ssa-strlen.c (handle_char_store): Record length for
21062 array store from STRING_CST.
21064 PR tree-optimization/57230
21065 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
21068 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
21070 * config/epiphany/epiphany.c (epiphany_init): Check size of
21071 NUM_MODES_FOR_MODE_SWITCHING.
21072 (epiphany_expand_prologue):
21073 Remove CONFIG_REGNUM initial value handling code.
21074 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
21075 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
21076 (emit_set_fp_mode, epiphany_mode_after): Likewise.
21077 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
21078 Don't return 1 for FP_MODE_NONE.
21079 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
21080 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
21081 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
21082 * config/epiphany/epiphany.md (save_config): New pattern.
21084 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
21086 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
21088 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
21090 * config/i386/i386.md (memory): Handle sseishft1.
21091 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
21092 (*vec_extractv2di_1): Ditto.
21094 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
21096 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
21099 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
21101 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
21102 Add mthumb/march=armv7-a multilib.
21103 Add mthumb/march=armv7-r multilib.
21104 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
21106 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
21108 * config/v850/t-rtems: Add more multilibs.
21110 2013-05-10 Richard Biener <rguenther@suse.de>
21112 PR tree-optimization/57214
21113 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
21114 not propagate from SSA names that occur in abnormal PHI nodes.
21116 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
21118 * stor-layout.c (element_precision): New function.
21119 * machmode.h (element_precision): Declare it.
21120 * tree.c (build_minus_one_cst): New function.
21121 (element_precision): Likewise.
21122 * tree.h (build_minus_one_cst): Declare new function.
21123 (element_precision): Likewise.
21124 * fold-const.c (operand_equal_p): Use element_precision.
21125 (fold_binary_loc): Handle vector types.
21126 * convert.c (convert_to_integer): Use element_precision.
21127 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
21130 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
21132 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
21133 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
21134 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
21135 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
21136 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
21137 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
21138 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
21139 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
21140 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
21141 (Uuw8): New constraints.
21142 (Usb4): Move into alphabetical order.
21143 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
21144 (sd8_operand, ub8_operand, uw8_operand): New predicates.
21145 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
21146 previously unnamed patterns.
21147 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
21148 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
21149 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
21150 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
21151 of set_attr_alternative/if_then_else. Use extended_mips16 instead
21152 of specific lengths.
21154 2013-05-10 Jakub Jelinek <jakub@redhat.com>
21156 * config/i386/i386.md (rotateinv): New code attr.
21157 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
21158 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
21159 roll $31, %eax, etc.
21161 PR tree-optimization/45216
21162 PR tree-optimization/57157
21163 * tree-ssa-forwprop.c (simplify_rotate): New function.
21164 (ssa_forward_propagate_and_combine): Call it.
21166 2013-05-10 Richard Biener <rguenther@suse.de>
21168 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
21169 disable peeling when we version for aliasing.
21170 (vector_alignment_reachable_p): Honor explicit user alignment.
21171 (vect_supportable_dr_alignment): Likewise.
21172 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
21173 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
21174 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
21175 then peeling to arrange for the cost-model check to come first.
21177 2013-05-10 Alan Modra <amodra@gmail.com>
21179 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
21180 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
21181 * configure: Regenerate.
21183 2013-05-10 Alan Modra <amodra@gmail.com>
21186 * varasm.c (default_elf_select_section): Move !DECL_P check..
21187 (get_named_section): ..to here before calling get_section_name.
21189 (default_section_type_flags): Add DECL_P check.
21190 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
21191 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
21193 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
21195 * config/epiphany/epiphany.c (epiphany_expand_prologue):
21196 When using gen_stack_adjust_str with a register offset, add a
21197 REG_FRAME_RELATED_EXPR note.
21199 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
21201 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
21202 (*vec_extractv4si_zext_mem): Ditto.
21203 (*vec_extractv2di): Add 0->x and x->x alternatives.
21204 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
21205 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
21207 2013-05-09 Jason Merrill <jason@redhat.com>
21209 N3639 C++1y VLA support
21210 * gimplify.c (gimplify_vla_decl): Don't touch an existing
21213 * tree.c (build_constructor_va): New.
21214 * tree.h: Declare it.
21216 2013-05-09 Martin Jambor <mjambor@suse.cz>
21219 * gimple-fold.c (canonicalize_constructor_val): Call
21220 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
21222 2013-05-09 Jan Hubicka <jh@suse.cz>
21223 Richard Biener <rguenther@suse.de>
21226 * symtab.c (symtab_make_decl_local): Do not add private names.
21228 2013-05-09 Jan Hubicka <jh@suse.cz>
21231 * symtab.c (insert_to_assembler_name_hash): Handle clones.
21232 (unlink_from_assembler_name_hash): Likewise.
21233 (symtab_prevail_in_asm_name_hash, symtab_register_node,
21234 symtab_unregister_node, symtab_initialize_asm_name_hash,
21235 change_decl_assembler_name): Update.
21237 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
21239 * config/aarch64/aarch64.md: New movtf split.
21240 (*movtf_aarch64): Update.
21241 (aarch64_movdi_tilow): Handle TF modes and rename to
21242 aarch64_movdi_<mode>low.
21243 (aarch64_movdi_tihigh): Handle TF modes and rename to
21244 aarch64_movdi_<mode>high
21245 (aarch64_movtihigh_di): Handle TF modes and rename to
21246 aarch64_mov<mode>high_di
21247 (aarch64_movtilow_di): Handle TF modes and rename to
21248 aarch64_mov<mode>low_di
21249 (aarch64_movtilow_tilow): Remove spurious whitespace.
21250 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
21252 (aarch64_print_operand): Update.
21254 2013-05-09 Alan Modra <amodra@gmail.com>
21256 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
21258 * configure: Regenerate.
21260 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
21262 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
21263 splitter preparation statements.
21264 * config/i386/sse.md (*vec_extract* splitters): Ditto.
21265 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
21268 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21270 * gimple-ssa-strength-reduction.c (count_candidates): Change
21271 return value to int.
21272 (analyze_candidates_and_replace): Change type of length to int.
21274 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
21276 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
21277 (*vec_extract<mode>): Use VI12_128 mode iterator.
21278 (*vec_extract<mode>_mem): Ditto.
21279 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
21282 2013-05-08 Diego Novillo <dnovillo@google.com>
21287 2012-08-17 Diego Novillo <dnovillo@google.com>
21290 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
21291 * config.in: Regenerate.
21292 * configure: Regenerate.
21293 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
21295 2013-05-08 Jan Hubicka <jh@suse.cz>
21298 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
21299 * cgraph.h (symtab_node_base): Add unique_name.
21300 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
21301 input_overwrite_node, input_varpool_node): Stream unique_name.
21302 * cgraphclones.c (cgraph_create_virtual_clone,
21303 cgraph_function_versioning): Set unique_name.
21304 * ipa.c (function_and_variable_visibility): Set unique_name.
21306 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21308 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
21309 (alloc_cand_and_find_basis): Restrict conditional candidate
21310 processing to CAND_MULTs.
21312 2013-05-08 Jan Hubicka <jh@suse.cz>
21315 lto-symtab.c (lto_symtab_symbol_p): New function.
21316 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
21317 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
21318 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
21319 Skip static symbols.
21321 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
21323 PR tree-optimization/57200
21324 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
21325 Only call inform if the preceding warning_at returns true.
21327 2013-05-07 Han Shen <shenhan@google.com>
21329 * cfgexpand.c (record_or_union_type_has_array_p): New function.
21330 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
21331 * common.opt (fstack-protector-strong): New option.
21332 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
21333 * doc/invoke.texi (Optimization Options): Document
21334 "-fstack-protector-strong".
21335 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
21337 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
21339 * config/mips/mips.c (mips_machine_reorg2): Return 0.
21341 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
21343 * ira.c (update_equiv_regs): Add insn having equiv memory even if
21344 it is not lhs of the insn.
21345 (setup_reg_equiv): Remove insn having equiv memory which it is not
21347 * lra-constraints.c (process_address): Try to improve generation
21348 code for address base + disp.
21349 (lra_constraints): Make correct the code for checking insn setting
21350 up backward equivalence. Remove insn only if it is in the init
21352 * lra-eliminations.c (update_reg_eliminate): Change return value.
21353 (lra_eliminate): Use the result.
21355 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
21357 * config/i386/sse.md (ssescalarnummask): New mode attribute.
21358 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
21359 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
21360 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
21361 register target operands.
21362 (*vec_extractv8hi_sse2): New pattern.
21363 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
21364 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
21365 (*vec_extract<mode>_mem): New insn and split pattern.
21367 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
21369 * config/arm/arm.c (arm_asan_shadow_offset): New function.
21370 (TARGET_ASAN_SHADOW_OFFSET): Define.
21371 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
21372 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
21374 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21376 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
21377 (incr_vec_index): Return -1 if increment not found.
21378 (create_add_on_incoming_edge): Assert if increment not found.
21379 (record_increment): Limit number of increments recorded.
21380 (all_phi_incrs_profitable): Return false if an increment not found.
21381 (replace_profitable_candidates): Don't process increments that were
21383 (analyze_candidates_and_replace): Limit size of incr_vec.
21385 2013-05-07 Richard Biener <rguenther@suse.de>
21387 * calls.c (special_function_p): setjmp-like functions are leaf.
21388 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
21389 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
21391 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
21393 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
21394 (aarch64_simd_mov<mode>): New expander.
21395 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
21396 (aarch64_simd_mov_to_<mode>high): Likewise.
21397 (aarch64_simd_mov_from_<mode>low): Likewise.
21398 (aarch64_simd_mov_from_<mode>high): Likewise.
21399 (aarch64_dup_lane<mode>): Update.
21400 (aarch64_dup_lanedi): New instruction pattern.
21401 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
21402 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
21404 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21406 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
21407 (replace_mult_candidate): Remove unnecessary argument; remove
21408 unnecessary parameter from call to introduce_cast_before_cand.
21409 (replace_unconditional_candidate): Remove unnecessary parameter
21410 from call to replace_mult_candidate.
21411 (replace_conditional_candidate): Likewise.
21412 (insert_initializers): Use make_temp_ssa_name.
21413 (introduce_cast_before_cand): Remove unnecessary argument; use
21414 make_temp_ssa_name.
21415 (replace_one_candidate): Remove unnecessary argument; remove
21416 unnecessary parameter from calls to introduce_cast_before_cand.
21417 (replace_profitable_candidates): Remove unnecessary parameters
21418 from calls to replace_one_candidate.
21420 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21422 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
21423 phi def as possibly hiding a basis for a CAND_ADD whose operands
21424 have been commuted in the analysis.
21425 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
21427 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21429 * config/aarch64/aarch64.md
21430 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
21431 shift value between 0-4.
21433 2013-05-07 Richard Biener <rguenther@suse.de>
21435 * double-int.h (rshift): New overload.
21436 * double-int.c (rshift): New function.
21437 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
21438 (create_reference_ops_from_ref): Remove.
21439 (vn_reference_insert): Use shared ops for constructing the
21440 reference and copy it.
21442 2013-05-07 Richard Biener <rguenther@suse.de>
21444 PR middle-end/57190
21445 * tree-eh.c (sink_clobbers): Properly propagate
21446 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
21448 2013-05-07 Jakub Jelinek <jakub@redhat.com>
21450 PR tree-optimization/57149
21451 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
21452 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
21453 collect_phi_def_edges, execute_late_warn_uninitialized): Use
21454 uninit_undefined_value_p instead of ssa_undefined_value_p.
21457 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
21458 for modifier == EXPAND_INITIALIZER.
21460 2013-05-07 Anton Blanchard <anton@samba.org>
21462 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
21463 for powerpc64 little endian.
21464 * configure: Regenerate.
21466 2013-05-06 Graham Stott <grahams@btinternet.com>
21468 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
21469 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
21470 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
21471 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
21473 2013-05-06 Graham Stott <grahams@btinternet.com>
21475 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
21476 codes which allow non-lvalues.
21478 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
21480 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
21481 components are all 1s.
21482 (integer_minus_onep): New function.
21483 * tree.h (integer_minus_onep): Declare it.
21484 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
21485 integer_minus_onep instead of integer_all_onesp.
21487 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21490 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
21491 variations of these patterns.
21493 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21495 * config/i386/i386.md (isa): Add x64_sse4 member.
21496 (enabled): Handle x64_sse4.
21497 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
21498 instruction for 64bit SSE4_1 targets. Update insn attributes.
21499 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
21500 instruction for SSE4_1 targets. Update insn attributes.
21501 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21502 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
21503 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
21505 (*vec_extractv4si): Rename from *sse4_1_pextrd.
21506 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
21507 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
21509 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21512 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
21514 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
21516 * final.c (do_assembler_dialects): Don't handle curly braces and
21517 vertical bar escaped by % as dialect delimiters.
21518 (output_asm_insn): Print curly braces and vertical bar if escaped
21519 by % and ASSEMBLER_DIALECT defined.
21520 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
21521 * doc/tm.texi: Regenerated.
21523 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
21525 * config/mips/mips.c: Include tree-pass.h.
21526 (mips_reorg): Split in pre- and post-dbr_schedule parts.
21527 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
21528 (pass_mips_machine_reorg2): New machine specific pass.
21529 (insert_pass_mips_machine_reorg2): New pass plugin definition.
21530 (mips_option_override): Register the new pass.
21531 * rtl.h (cleanup_barriers): Remove prototype.
21532 (dbr_schedule): Likewise.
21533 * jump.c (cleanup_barriers): Make static.
21534 * reorg.c (dbr_schedule): Likewise.
21536 2013-05-06 Richard Biener <rguenther@suse.de>
21538 PR tree-optimization/57185
21539 * tree-parloops.c (add_field_for_reduction): Handle anonymous
21540 SSA names properly.
21542 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21545 * config/i386/i386.c (add_parameter_dependencies): Add dependence
21546 between "first_arg" and "insn", not "last" and "insn".
21548 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21550 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
21551 (find_candidates_in_block): Re-enable slsr_process_phi.
21552 (create_phi_basis): Fix double counting of candidate adjustment.
21554 2013-05-06 Richard Biener <rguenther@suse.de>
21556 PR middle-end/57147
21557 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
21558 the edge is also fallthru, preserve it and just clear the
21560 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
21561 also complex, preserve that and just clear the fallthru flag.
21562 * tree-inline.c (update_ssa_across_abnormal_edges): Also
21563 update virtual operands.
21565 2013-05-06 Alan Modra <amodra@gmail.com>
21567 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
21568 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
21569 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
21570 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
21571 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21572 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
21574 2013-05-06 Alan Modra <amodra@gmail.com>
21576 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
21577 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
21578 (DEFAULT_ASM_ENDIAN): Define.
21579 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21580 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
21581 Update -K PIC clause from sysv4.h.
21582 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
21583 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
21585 2013-05-06 Alan Modra <amodra@gmail.com>
21587 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
21588 twice for little-endian.
21589 (ashrdi3_no_power, ashrdi3): Support little-endian.
21591 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21594 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
21595 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
21597 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
21598 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
21599 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
21602 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
21603 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21605 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
21606 * config/pa/pa.opt: Make mbig-switch a no-op.
21607 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
21608 (CASE_VECTOR_MODE): Always return SImode.
21609 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
21610 for the !TARGET_BIG_SWITCH case.
21611 * config/pa/pa-linux.h: Likewise.
21612 * config/pa/pa-openbsd.h: Likewise.
21613 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
21614 * config/pa/pa.md (short_jump): Remove define_insn.
21615 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
21616 (casesi0): Remove define_insn.
21617 (type): Remove btable_branch.
21618 (pa_combine_type): Likewise.
21619 (in_nullified_branch_delay): Likewise.
21620 (in_call_delay): Likewise.
21621 (define_delay): Likewise.
21622 (define_insn_reservation "Z3"): Likewise.
21623 (define_insn_reservation "Z4"): Likewise.
21624 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
21625 (pa_adjust_insn_length): Remove adjustment for btable branches.
21626 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
21629 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
21631 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21632 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
21633 Add m->r,x alternatives.
21634 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
21635 splitters using SWI48x mode iterator.
21636 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
21637 TARGET_64BIT. Add m->x alternative.
21638 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
21639 Add o->x alternative. Enable for TARGET_SSE.
21640 (sse_storeq): Remove expander.
21641 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
21642 with memory input operand.
21643 (*vec_extractv2di_1 splitter): New.
21644 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
21645 * config/i386/i386.md (ssevecmodelower): New mode attribute.
21647 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21649 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
21650 (INT_LOWPART): Delete.
21651 (extract_MB): Adjust.
21652 (extract_ME): Adjust.
21653 (print_operand): Adjust.
21655 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21657 * config/rs6000/predicates.md (reg_or_add_cint_operand,
21658 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21659 (reg_or_logical_cint_operand, easy_fp_constant,
21660 logical_const_operand): Delete "CONST_DOUBLE" case.
21661 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
21662 "HOST_BITS_PER_WIDE_INT == 64" test.
21663 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
21664 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
21665 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
21666 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
21668 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
21669 CONST_DOUBLE DImode/VOIDmode case.
21670 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
21671 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21672 CONST_DOUBLE VOIDmode case.
21673 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21674 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
21675 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
21676 Delete CONST_DOUBLE case.
21677 (splitters for mov FMOVE64 const_double): Delete
21678 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21679 "HOST_BITS_PER_WIDE_INT >= 64" test.
21680 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
21682 (mov DI const_double): Delete.
21684 2013-05-04 Jakub Jelinek <jakub@redhat.com>
21686 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
21687 on op shows all bits zero in mode of a lowpart subreg, return zero.
21689 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
21692 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
21693 to save TFmode registers and DImode to save TImode registers for
21694 caller save operations.
21695 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
21696 mark being partially clobbered since they only use the first
21699 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
21700 and TDmode only use the upper 64-bits of each VSX register.
21702 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21704 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
21705 (find_candidates_in_block): Disable slsr_process_phi.
21707 2013-05-03 Guozhi Wei <carrot@google.com>
21709 * coverage.c (coverage_obj_init): Move the construction of gcov
21711 (build_init_ctor): ... here.
21713 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21715 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
21716 (slsr_cand_d): Redefine def_phi.
21717 (stride_status, phi_adjust_status, count_phis_status): New enums.
21718 (find_phi_def): New.
21719 (find_basis_for_base_expr): New.
21720 (find_basis_for_candidate): Handle hidden bases.
21721 (alloc_cand_and_find_basis): Handle phi candidates.
21722 (slsr_process_phi): New.
21723 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
21724 (create_mul_imm_cand): Likewise.
21725 (create_add_ssa_cand): Exclude phi base candidates.
21726 (create_add_imm_cand): Likewise.
21727 (slsr_process_cast): Likewise.
21728 (slsr_process_copy): Likewise.
21729 (find_candidates_in_block): Handle phi candidates.
21730 (dump_candidate): Likewise.
21731 (unconditional_cands): Delete.
21732 (unconditional_cands_with_known_stride_p): Delete.
21733 (phi_dependent_cand_p): New.
21734 (cand_increment): Handle phi-dependent candidates.
21735 (replace_dependent): Delete.
21736 (replace_mult_candidate): New.
21737 (replace_unconditional_candidate): New.
21738 (incr_vec_index): Move to avoid forward reference.
21739 (create_add_on_incoming_edge): New.
21740 (create_phi_basis): New.
21741 (replace_dependents): Delete.
21742 (replace_conditional_candidate): New.
21743 (phi_add_costs): New.
21744 (replace_uncond_cands_and_profitable_phis): New.
21745 (record_increment): Handle phi adjustments.
21746 (record_phi_increments): New.
21747 (record_increments): Handle phi adjustments.
21748 (phi_incr_cost): New.
21749 (lowest_cost_path): Handle phis.
21750 (total_savings): Likewise.
21751 (analyze_increments): Likewise.
21752 (ncd_with_phi): New.
21753 (ncd_of_cand_and_phis): New.
21754 (nearest_common_dominator_for_cands): Handle phi increments.
21755 (all_phi_incrs_profitable): New.
21756 (replace_profitable_candidates): Handle phi-dependent candidates.
21757 (analyze_candidates_and_replace): Likewise.
21759 2013-05-03 Teresa Johnson <tejohnson@google.com>
21762 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
21763 do not exceed REG_BR_PROB_BASE.
21765 2013-05-03 Jeff Law <law@redhat.com>
21767 PR tree-optimization/57144
21768 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
21769 operand of the condition will bit into the new type when eliminating
21770 a cast feeding a condition.
21772 2013-05-03 Jakub Jelinek <jakub@redhat.com>
21774 PR rtl-optimization/57130
21775 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
21776 of COMPARE as in_code to the recursive call if needed.
21778 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
21780 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
21781 (enabled): Handle new members.
21782 * config/i386/sse.md (*vec_concatv2si): Merge from
21783 *vec_concatv2si_sse2 and vec_concatv2si_sse.
21784 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
21786 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
21788 PR tree-optimization/57027
21789 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
21790 for fnms opportunity, check we got the prerequisite kind
21791 of tree / gimple before using accessor functions.
21793 2013-05-03 Richard Biener <rguenther@suse.de>
21795 * double-int.h (lshift): New overload without precision
21796 and arith argument.
21797 (operator *=, operator +=, operator -=): Move ...
21798 * double-int.c (operator *=, operator +=, operator -=): ... here
21799 and implement more efficiently.
21800 (mul_double_with_sign): Remove.
21801 (lshift_double): Adjust to take unsinged shift argument, push
21802 dispatching code to callers.
21803 (mul_double_wide_with_sign): Add early out for callers that
21804 are not interested in high parts or overflow.
21805 (lshift): New function.
21806 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
21807 dispatch code here.
21808 (lrotate, rrotate): Use logical shifts.
21809 * expr.c (get_inner_reference): Use lshift.
21810 * fixed-value.c (do_fixed_divide): Likewise.
21811 * tree-dfa.c (get_ref_base_and_extent): Likewise.
21812 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
21813 (indirect_refs_may_alias_p): Likewise.
21814 (stmt_kills_ref_p_1): Likewise.
21816 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21818 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
21820 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21822 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
21823 scalar form of FABD instruction.
21825 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
21827 * lra-constraints.c (process_alt_operands): Add checking alt
21828 number to choose the best alternative.
21830 2013-05-02 Richard Biener <rguenther@suse.de>
21832 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
21833 bitmap and its handling.
21834 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
21836 2013-05-02 Richard Biener <rguenther@suse.de>
21838 PR middle-end/57140
21839 * tree-inline.c (copy_loops): Properly handle removed loops.
21840 (copy_cfg_body): Mark destination loops for fixup if source
21841 loops needed fixup.
21843 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
21846 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
21847 generating simple_return for naked functions.
21849 2013-05-02 Martin Jambor <mjambor@suse.cz>
21851 PR middle-end/56988
21852 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
21853 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
21855 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21856 ipa_agg_replacement_value structures.
21857 (known_aggs_to_agg_replacement_list): Likewise.
21858 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21859 (read_agg_replacement_chain): Likewise.
21860 (ipcp_transform_function): Also check that by_ref flags match.
21862 2013-05-02 Richard Biener <rguenther@suse.de>
21864 * graphds.h (struct graph): Add obstack member.
21865 * graphds.c (new_graph): Initialize obstack and allocate
21867 (add_edge): Allocate edge from the obstack.
21868 (free_graph): Free the obstack instead of all edges and vertices.
21870 2013-05-02 Teresa Johnson <tejohnson@google.com>
21872 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21874 * cfg.c (update_bb_profile_for_threading): Ditto.
21875 * tree-inline.c (copy_bb): Ditto.
21876 (copy_edges_for_bb): Ditto.
21877 (initialize_cfun): Ditto.
21878 (copy_cfg_body): Ditto.
21879 (expand_call_inline): Ditto.
21880 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21881 (estimate_node_size_and_time): Ditto.
21882 (inline_merge_summary): Ditto.
21883 * cgraphclones.c (cgraph_clone_edge): Ditto.
21884 (cgraph_clone_node): Ditto.
21885 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21886 (compute_trg_info): Ditto.
21888 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21890 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21891 S reg when fp attribute set.
21892 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21894 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21896 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21898 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21899 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21900 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21902 2013-05-02 Richard Biener <rguenther@suse.de>
21904 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21905 (struct instantiate_cache_entry): New type.
21906 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21907 (struct instantiate_cache_type): New type.
21908 (set_instantiated_value, get_instantiated_value): Remove.
21909 (get_instantiated_value_entry): New function.
21910 (instantiate_scev_name): Use the new cache and adjust.
21911 (instantiate_scev_poly): Adjust.
21912 (instantiate_scev_binary): Likewise.
21913 (instantiate_array_ref): Likewise.
21914 (instantiate_scev_convert): Likewise.
21915 (instantiate_scev_not): Likewise.
21916 (instantiate_scev_3): Likewise.
21917 (instantiate_scev_2): Likewise.
21918 (instantiate_scev_r): Likewise.
21919 (instantiate_scev): Likewise.
21920 (resolve_mixers): Likewise.
21922 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21925 * lra-constraints.c (best_small_class_operands_num): Remove.
21926 (process_alt_operands): Remove small_class_operands_num. Take
21927 small classes operands into losers and only if the operand is not
21928 matched. Modify debugging output.
21929 (curr_insn_transform): Remove best_small_class_operands_num.
21932 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21934 * config/aarch64/aarch64-builtins.c
21935 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21936 * config/aarch64/aarch64-simd-builtins.def
21937 (reduc_splus_): Add new modes.
21938 (reduc_uplus_): New.
21939 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21940 (reduc_uplus_v4sf): Likewise.
21941 (reduc_splus_v4sf): Likewise.
21942 (aarch64_addv<mode>): Likewise.
21943 (reduc_uplus_<mode>): Likewise.
21944 (reduc_splus_<mode>): Likewise.
21945 (aarch64_addvv2di): Likewise.
21946 (reduc_uplus_v2di): Likewise.
21947 (reduc_splus_v2di): Likewise.
21948 (aarch64_addvv2si): Likewise.
21949 (reduc_uplus_v2si): Likewise.
21950 (reduc_splus_v2si): Likewise.
21951 (reduc_<sur>plus_<mode>): New.
21952 (reduc_<sur>plus_v2di): Likewise.
21953 (reduc_<sur>plus_v2si): Likewise.
21954 (reduc_<sur>plus_v4sf): Likewise.
21955 (aarch64_addpv4sf): Likewise.
21956 * config/aarch64/arm_neon.h
21957 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21958 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21959 add UNSPEC_SADDV, UNSPEC_UADDV.
21961 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21963 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21965 * config/aarch64/arm_neon.h
21966 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21968 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21970 * config/aarch64/aarch64-builtins
21971 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21973 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21975 * config/aarch64/aarch64-simd-builtins.def
21976 (reduc_smax_): New.
21977 (reduc_smin_): Likewise.
21978 (reduc_umax_): Likewise.
21979 (reduc_umin_): Likewise.
21980 (reduc_smax_nan_): Likewise.
21981 (reduc_smin_nan_): Likewise.
21984 (smax): Update for V2SF, V4SF and V2DF modes.
21987 (smin_nan): Likewise.
21988 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21989 (<su><maxmin><mode>3): ...This, refactor.
21990 (s<maxmin><mode>3): New.
21991 (<maxmin_uns><mode>3): Likewise.
21992 (reduc_<maxmin_uns>_<mode>): Refactor.
21993 (reduc_<maxmin_uns>_v4sf): Likewise.
21994 (reduc_<maxmin_uns>_v2si): Likewise.
21995 (aarch64_<fmaxmin><mode>: Remove.
21996 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21998 (vmin<q>_f<32,64>): Likewise.
21999 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
22001 (su): Add mappings for smax, smin, umax, umin.
22003 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
22004 (FMAXMIN): Rename as...
22005 (FMAXMIN_UNS): ...This.
22007 (fmaxminv): Likewise.
22008 (fmaxmin): Likewise.
22010 (maxmin_uns_op): Likewise.
22012 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22014 * config/aarch64/arm_neon.h
22015 (vac<ge, gt><sd>_f<32, 64>): Rename to...
22016 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
22017 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
22019 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22021 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
22022 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
22024 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22026 * config/aarch64/aarch64-simd.md
22027 (vcond<mode>_internal): Handle special cases for constant masks.
22028 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
22029 (vcondu<mode><mode>): Likewise.
22030 (vcond<v_cmp_result><mode>): New.
22032 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22034 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
22035 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
22036 * config/aarch64/aarch64-simd-builtins.def
22037 (cmeq): Update to BUILTIN_VALLDI.
22042 * config/aarch64/arm_neon.h
22043 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
22044 to builtins or C as appropriate.
22046 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22048 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
22050 (cmhi): Rename to...
22052 * config/aarch64/aarch64-simd.md
22053 (simd_mode): Add SF.
22054 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
22055 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
22056 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
22057 (cstore<mode>_neg): ...This.
22058 * config/aarch64/iterators.md
22060 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
22061 (COMPARISONS): New.
22062 (UCOMPARISONS): Likewise.
22063 (optab): Add missing comparisons.
22069 (VCMP_S): Likewise.
22070 (VCMP_U): Likewise.
22071 (V_cmp_result): Add DF, SF modes.
22072 (v_cmp_result): Likewise.
22074 (vmtype): Likewise.
22075 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
22077 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
22079 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
22080 define_insn to define_insn_and_split.
22081 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
22082 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
22083 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
22084 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
22085 (thumb2_negscc): Likewise.
22087 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
22089 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
22091 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
22093 * config/arm/thumb2.md: Remove trailing whitespaces.
22095 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
22097 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
22098 Use gen_int_mode rather than GEN_INT.
22100 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
22102 * value-prof.c (stream_in_histogram_value): Remove the strayed
22105 2013-04-30 Richard Biener <rguenther@suse.de>
22107 PR middle-end/57122
22108 * cfghooks.c (split_edge): Properly check for the loop latch edge.
22110 2013-04-30 Richard Biener <rguenther@suse.de>
22112 PR middle-end/57107
22113 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
22115 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
22117 PR rtl-optimization/56957
22118 PR rtl-optimization/57105
22119 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
22120 variable. Use just INSN_UID for determining whether an insn
22121 should be only disconnected from the insn stream.
22122 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
22124 2013-04-30 Jakub Jelinek <jakub@redhat.com>
22126 PR tree-optimization/57104
22127 * tsan.c (instrument_expr): Don't instrument accesses to
22128 DECL_HARD_REGISTER VAR_DECLs.
22130 2013-04-30 Richard Biener <rguenther@suse.de>
22132 * function.h (loops_for_fn): New inline function.
22133 (set_loops_for_fn): Likewise.
22134 * cfgloop.h (place_new_loop): Add struct function parameter.
22135 (get_loop): Likewise.
22136 (get_loops): Likewise.
22137 (number_of_loops): Likewise.
22138 (fel_next): Adjust.
22139 (fel_init): Likewise.
22140 * cfg.c (get_loop_copy): Adjust.
22141 * cfgloop.c (flow_loops_dump): Likewise.
22142 (record_loop_exits): Likewise.
22143 (verify_loop_structure): Likewise.
22144 * cfgloopanal.c (mark_irreducible_loops): Likewise.
22145 (estimate_reg_pressure_cost): Likewise.
22146 (mark_loop_exit_edges): Likewise.
22147 * cfgloopmanip.c (place_new_loop): Likewise.
22148 (add_loop): Likewise.
22149 (duplicate_loop): Likewise.
22150 * graph.c (draw_cfg_nodes): Likewise.
22151 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
22152 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
22153 (extract_affine_chrec): Likewise.
22154 (build_scop_iteration_domain): Likewise.
22155 * graphite.c (graphite_initialize): Likewise.
22156 * ira-build.c (create_loop_tree_nodes): Likewise.
22157 (more_one_region_p): Likewise.
22158 (rebuild_regno_allocno_maps): Likewise.
22159 (mark_loops_for_removal): Likewise.
22160 (mark_all_loops_for_removal): Likewise.
22161 (remove_unnecessary_regions): Likewise.
22162 (ira_build): Likewise.
22163 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
22164 * loop-init.c (fix_loop_structure): Likewise.
22165 (gate_rtl_move_loop_invariants): Likewise.
22166 (gate_rtl_unswitch): Likewise.
22167 (gate_rtl_unroll_and_peel_loops): Likewise.
22168 (rtl_doloop): Likewise.
22169 * lto-streamer-in.c (input_cfg): Likewise.
22170 * lto-streamer-out.c (output_cfg): Likewise.
22171 * modulo-sched.c (sms_schedule): Likewise.
22172 * predict.c (tree_estimate_probability): Likewise.
22173 (tree_estimate_probability_driver): Likewise.
22174 (estimate_loops): Likewise.
22175 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
22176 (move_sese_region_to_fn): Likewise.
22177 (debug_loop_num): Likewise.
22178 * tree-chrec.c (chrec_evaluate): Likewise.
22179 (hide_evolution_in_other_loops_than_loop): Likewise.
22180 (chrec_component_in_loop_num): Likewise.
22181 (reset_evolution_in_loop): Likewise.
22182 (evolution_function_is_invariant_rec_p): Likewise.
22183 * tree-if-conv.c (main_tree_if_conversion): Likewise.
22184 * tree-inline.c (copy_loops): Likewise.
22185 (copy_cfg_body): Likewise.
22186 (tree_function_versioning): Likewise.
22187 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
22188 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
22190 (add_to_evolution_1): Likewise.
22191 (scev_const_prop): Likewise.
22192 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
22193 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
22194 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
22195 (tree_ssa_lim_initialize): Likewise.
22196 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
22197 (verify_loop_closed_ssa): Likewise.
22198 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
22199 (tree_ssa_loop_im): Likewise.
22200 (tree_ssa_loop_unswitch): Likewise.
22201 (tree_vectorize): Likewise.
22202 (check_data_deps): Likewise.
22203 (tree_ssa_loop_ivcanon): Likewise.
22204 (tree_ssa_loop_bounds): Likewise.
22205 (tree_complete_unroll): Likewise.
22206 (tree_complete_unroll_inner): Likewise.
22207 (tree_parallelize_loops): Likewise.
22208 (tree_ssa_loop_prefetch): Likewise.
22209 (tree_ssa_loop_ivopts): Likewise.
22210 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
22211 * tree-vectorizer.c (vectorize_loops): Likewise.
22213 2013-04-29 Mike Frysinger <vapier@gentoo.org>
22215 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
22216 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
22217 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
22218 with EABI_LINK_SPEC.
22220 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22223 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
22226 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
22229 * lra-constraints.c (process_alt_operands): Discourage a bit more
22230 using memory for pseudos. Print cost dump for alternatives.
22231 Modify cost values for conflicts with early clobbers.
22232 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
22234 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22237 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
22239 2013-04-29 Ian Bolton <ian.bolton@arm.com>
22241 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
22242 from/to S register.
22243 (movdi_aarch64): Support LDR/STR from/to D register.
22245 2013-04-29 Ian Bolton <ian.bolton@arm.com>
22247 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
22248 or higher by default.
22250 2013-04-29 Richard Biener <rguenther@suse.de>
22252 PR middle-end/57075
22253 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
22254 even if not adding abnormal edges for calls that can make
22257 2013-04-29 Richard Biener <rguenther@suse.de>
22259 PR middle-end/57103
22260 * tree-cfg.c (move_stmt_op): Fix condition under which to update
22262 (move_stmt_r): Remove redundant checking.
22264 2013-04-29 Teresa Johnson <tejohnson@google.com>
22267 * basic-block.h (apply_scale): New function.
22268 (apply_probability): Use apply_scale.
22269 * gimple-streamer-in.c (input_bb): Ditto.
22270 * lto-streamer-in.c (input_cfg): Ditto.
22271 * lto-cgraph.c (merge_profile_summaries): Ditto.
22272 * tree-optimize.c (execute_fixup_cfg): Ditto.
22273 * tree-inline.c (copy_bb): Update comment to use apply_scale.
22274 (copy_edges_for_bb): Ditto.
22275 (copy_cfg_body): Ditto.
22277 2013-04-29 Tom de Vries <tom@codesourcery.com>
22279 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
22280 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
22281 (tail_merge_optimize): Handle current_loops == NULL.
22283 2013-04-26 Jeff Law <law@redhat.com>
22285 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
22286 (simplify_cond_using_ranges): Generalize code to simplify
22287 COND_EXPRs where one argument is a constant and the other
22288 is an SSA_NAME created by an integral type conversion.
22290 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22292 * config/arm/arm.md (store_minmaxsi): Use only when
22293 optimize_insn_for_size_p.
22295 2013-04-29 Christian Bruel <christian.bruel@st.com>
22298 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
22300 2013-04-29 Richard Biener <rguenther@suse.de>
22302 PR middle-end/57089
22303 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
22304 loop tree make sure to schedule a fixup for the child as well.
22305 (expand_omp_for_generic): Properly add loops.
22306 (expand_omp_for_static_nochunk): Likewise.
22307 (expand_omp_for_static_chunk): Likewise.
22308 (expand_omp_for): For the degenerate case fixup loops.
22309 (expand_omp_sections): Fix default bb placement in loops.
22310 (expand_omp_atomic_pipeline): Properly add loops.
22312 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22314 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
22316 2013-04-29 Tom de Vries <tom@codesourcery.com>
22318 * tree-ssa-tail-merge.c: Update header comment.
22320 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22322 * config/aarch64/arm_neon.h
22323 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
22324 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
22325 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
22326 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
22327 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
22328 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
22329 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
22330 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
22332 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22334 * config/aarch64/aarch64-simd.md
22335 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
22336 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
22337 fix_trunc, fixuns_trunc.
22338 (ftrunc<VDQF:mode>2): New.
22339 * config/aarch64/iterators.md (optab): Add fix, fixuns.
22340 (fix_trunc_optab): New.
22342 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22344 * config/aarch64/aarch64-builtins.c
22345 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
22346 iceilf, lround, iroundf.
22348 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22351 * config/i386/i386.h (enum ix86_tune_indices)
22352 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
22353 New, split from X86_TUNE_INTER_UNIT_MOVES.
22354 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
22355 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
22356 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
22357 (TARGET_INTER_UNIT_MOVES): Remove.
22358 * config/i386/i386.c (initial_ix86_tune_features): Update.
22359 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
22360 (ix86_expand_convert_uns_didf_sse): Use
22361 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
22362 (ix86_expand_vector_init_one_nonzero): Ditto.
22363 (ix86_expand_vector_init_interleave): Ditto.
22364 (inline_secondary_memory_needed): Return true for moves from SSE class
22365 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
22366 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
22367 * config/i386/constraints.md (Yi, Ym): Depend on
22368 TARGET_INTER_UNIT_MOVES_TO_VEC.
22369 (Yj, Yn): New constraints.
22370 * config/i386/i386.md (*movdi_internal): Change constraints of
22371 operand 1 from Yi to Yj and from Ym to Yn.
22372 (*movsi_internal): Ditto.
22373 (*movdf_internal): Ditto.
22374 (*movsf_internal): Ditto.
22375 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
22376 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
22377 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
22378 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
22379 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
22380 * config/i386/sse.md (movdi_to_sse): Ditto.
22381 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
22382 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
22383 TARGET_INTER_UNIT_MOVES.
22384 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
22385 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
22386 instead of TARGET_INTER_UNIT_MOVES.
22387 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
22388 operand 1 from Yi to Yj and from Ym to Yn.
22390 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22392 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
22393 (float_truncate_hi_): Likewise.
22394 (float_extend_lo_): Likewise.
22395 (float_truncate_lo_): Likewise.
22396 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
22397 (aarch64_float_extend_lo_v2df): Likewise.
22398 (vec_unpacks_hi_v4sf): Likewise.
22399 (aarch64_float_truncate_lo_v2sf): Likewise.
22400 (aarch64_float_truncate_hi_v4sf): Likewise.
22401 (vec_pack_trunc_v2df): Likewise.
22402 (vec_pack_trunc_df): Likewise.
22404 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22406 * config/aarch64/aarch64-builtins.c
22407 (aarch64_fold_builtin): Fold float conversions.
22408 * config/aarch64/aarch64-simd-builtins.def
22409 (floatv2si, floatv4si, floatv2di): New.
22410 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
22411 * config/aarch64/aarch64-simd.md
22412 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
22413 * config/aarch64/iterators.md (FLOATUORS): New.
22414 (optab): Add float, floatuns.
22415 (su_optab): Likewise.
22417 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22419 * config/aarch64/aarch64-builtins.c
22420 (aarch64_builtin_vectorized_function): Use new names for
22422 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
22423 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
22424 (fcvtzu): Split as...
22425 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
22426 (fcvtas): Split as...
22427 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
22428 (fcvtau): Split as...
22429 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
22430 (fcvtps): Split as...
22431 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
22432 (fcvtpu): Split as...
22433 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
22434 (fcvtms): Split as...
22435 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
22436 (fcvtmu): Split as...
22437 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
22438 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
22439 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
22440 (lfrintnusf, lfrintnudf): Likewise.
22441 * config/aarch64/aarch64-simd.md
22442 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
22444 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
22445 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
22446 (fcvt_pattern): Likewise.
22448 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22450 * config/aarch64/aarch64-simd.md
22451 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
22452 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
22454 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22456 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
22457 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
22458 (vrnd<a,m,n,p>_f32): Implement using builtins.
22459 (vrnd<i,x><q>_f<32, 64>): New.
22461 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22463 * config/aarch64/aarch64-builtins.c
22464 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
22465 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
22466 (frintz): Rename to...
22468 (frintp): Rename to...
22470 (frintm): Rename to...
22472 (frinti): Rename to...
22473 (nearbyint): ...this.
22474 (frintx): Rename to...
22476 (frinta): Rename to...
22478 * config/aarch64/aarch64-simd.md
22479 (aarch64_frint<frint_suffix><mode>): Delete.
22480 (<frint_pattern><mode>2): Convert to insn.
22481 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
22482 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
22483 (frint_pattern): Likewise.
22484 (frint_suffix): Likewise.
22486 2013-04-29 Richard Biener <rguenther@suse.de>
22488 PR tree-optimization/57081
22489 * loop-init.c: Include tree-flow.h.
22490 (loop_optimizer_finalize): Free number of iteration estimates.
22491 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
22493 2013-04-29 Jakub Jelinek <jakub@redhat.com>
22495 PR tree-optimization/57083
22496 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
22497 non-singleton shift count range, zero extend low_bound for uns case.
22499 * config/i386/predicates.md (general_vector_operand): New predicate.
22500 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
22501 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
22502 if they aren't nonimmediate operands. If their original values
22503 satisfy const_vector_equal_evenodd_p, don't shift them.
22504 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
22505 predicates. For the SSE4.1 case force operands[{1,2}] into registers
22506 if not nonimmediate_operand.
22507 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
22508 instead of register_operand.
22509 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
22511 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22513 * stor-layout.c (finalize_size_functions): Allocate a structure and
22514 reset cfun before dumping the functions.
22516 2013-04-27 Jakub Jelinek <jakub@redhat.com>
22518 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
22521 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
22522 use xop_pmacsdqh if uns_p.
22523 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
22524 the immediate rotate count.
22526 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
22528 * rtl.h (struct rtx_def): Add comment for field jump.
22529 (LRA_SUBREG_P): New macro.
22530 * recog.c (register_operand): Check LRA_SUBREG_P.
22531 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22533 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22534 (lra_final_code_change): Skip subreg change for operators.
22535 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22536 if there are no operand changes.
22537 * lra-constraints.c (curr_insn_set): New.
22538 (match_reload): Set LRA_SUBREG_P.
22539 (emit_spill_move): Ditto.
22540 (check_and_process_move): Use curr_insn_set. Process only single
22541 set insns. Don't initialize sec_mem_p and change_p.
22542 (simplify_operand_subreg): Use LRA_SUBREG_P.
22543 (reg_in_class_p): New function.
22544 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22545 of #ifdef. Add code to remove cycling.
22546 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22547 non-null disp. Reload inner instead of disp when base and index
22548 are null. Try to put lo_sum into register.
22549 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22550 (check_and_process_move): Move code for move cost check to
22551 simple_move_p. Remove equiv_substitution.
22552 (simple_move_p): New function.
22553 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22554 curr_insn_set. Call check_and_process_move only for single set
22555 insns. Use the new function. Move call of check_and_process_move
22556 after operand equiv substitution and address process.
22558 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22561 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
22562 with nonlocal goto receivers or returns twice calls, ignore
22563 unininitialized values from abnormal edges to nl goto receiver
22564 or returns twice call.
22566 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22568 PR tree-optimization/57051
22569 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
22570 and VEC_RSHIFT_EXPR if shift count is a multiple of element
22573 2013-04-26 Richard Biener <rguenther@suse.de>
22575 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
22576 (expand_omp_taskreg): Likewise. Mark loops for fixup.
22577 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
22578 (fixup_loop_arrays_after_move): New function.
22579 (move_sese_region_to_fn): Properly outline the loop tree parts
22580 of the SESE region.
22582 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
22584 * config/i386/i386.md (type, unit): Fix long lines.
22586 2013-04-26 Richard Biener <rguenther@suse.de>
22588 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
22589 (lto-streamer-out.o): Likewise.
22590 * cfgloop.c (init_loops_structure): Export, add struct function
22591 argument and adjust.
22592 (flow_loops_find): Adjust.
22593 * cfgloop.h (enum loop_estimation): Add EST_LAST.
22594 (init_loops_structure): Declare.
22595 * lto-streamer-in.c: Include cfgloop.h.
22596 (input_cfg): Input the loop tree.
22597 * lto-streamer-out.c: Include cfgloop.h.
22598 (output_cfg): Output the loop tree.
22599 (output_struct_function_base): Do not drop PROP_loops.
22601 2013-03-26 Richard Biener <rguenther@suse.de>
22603 * tree-cfg.c (execute_build_cfg): Build the loop tree.
22604 (pass_build_cfg): Provide PROP_loops.
22605 (move_sese_region_to_fn): Remove loops that are outlined into fn
22607 * tree-inline.c: Include cfgloop.h.
22608 (initialize_cfun): Do not drop PROP_loops.
22609 (copy_loops): New function.
22610 (copy_cfg_body): Copy loop structure.
22611 (tree_function_versioning): Initialize destination loop tree.
22612 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
22613 (pass_parallelize_loops): Do IL verification.
22614 * loop-init.c (loop_optimizer_init): Fixup loops if required.
22615 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
22616 the CFG make sure we fixup loops as well.
22617 * tree-ssa-tail-merge.c: Include cfgloop.h.
22618 (replace_block_by): When merging loop latches mark loops for fixup.
22619 * lto-streamer-out.c (output_struct_function_base): Drop
22620 PROP_loops for now.
22621 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
22622 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
22623 * ipa-split.c: Include cfgloop.h.
22624 (split_function): Add the new return block to the loop tree root.
22625 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
22626 whether we have removed the forwarder block.
22627 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
22628 * cfgloop.h (place_new_loop): Declare.
22629 * cfgloopmanip.c (place_new_loop): Export.
22630 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
22631 (tree-switch-conversion.o): Likewise.
22632 (tree-complex.o): Likewise.
22633 (tree-inline.o): Likewise.
22634 (tree-ssa-tailmerge.o): Likewise.
22635 (ipa-split.o): Likewise.
22636 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
22637 (tree-ssa-copy.o): Likewise.
22638 * tree-switch-conversion.c: Include cfgloop.h
22639 (process_switch): If we emit a bit-test cascade, schedule loops
22641 * tree-complex.c: Include cfgloop.h.
22642 (expand_complex_div_wide): Properly add new basic-blocks to loops.
22643 * asan.c: Include cfgloop.h.
22644 (create_cond_insert_point): Properly add new basic-blocks to
22645 loops, schedule loop fixup.
22646 * cfgloop.c (verify_loop_structure): Check that looks are not
22648 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
22650 (expand_omp_for_generic): Likewise.
22651 (expand_omp_sections): Likewise.
22652 (expand_omp_atomic_pipeline): Schedule loops for fixup.
22653 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
22654 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
22655 is initialized, not when loops are present.
22656 * tree-parloops.c (parallelize_loops): Remove checking here.
22657 * passes.c (init_optimization_passes): Schedule a copy-propagation
22658 pass before complete unrolling of inner loops.
22660 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22662 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
22663 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
22664 (colorize_init): Add argument to _WIN32 version.
22665 * toplev.c: Include diagnostic-color.h.
22666 (process_options): Default to -fdiagnostics-color=auto if
22667 GCC_COLORS env var is in the environment.
22668 * common.opt (fdiagnostics-color=): Add Var and Init.
22669 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
22670 env var is in the environment, the default is auto rather than never.
22672 * diagnostic.h (file_name_as_prefix): Add context argument.
22673 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
22674 the string as locus.
22675 * langhooks.c (lhd_print_error_function): Adjust caller.
22677 2013-04-25 Lawrence Crowl <crowl@google.com>
22679 * var-tracking.c (shared_hash_def::htab):
22680 Change type to hash_table. Update dependent calls and types.
22682 2013-04-25 Lawrence Crowl <crowl@google.com>
22684 * Makefile.in: Update as needed below.
22686 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
22687 Move declaration to after the type's method definitons.
22689 * attribs.c (htab_t scoped_attributes::attribute_hash):
22690 Change type to hash_table. Update dependent calls and types.
22692 * bitmap.c (htab_t bitmap_desc_hash):
22693 Change type to hash_table. Update dependent calls and types.
22695 * cselib.c (htab_t cselib_hash_table):
22696 Change type to hash_table. Update dependent calls and types.
22698 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
22699 (hash_string_slot_node): Move implementation into lto-streamer.h
22700 struct string_slot_hasher.
22701 (eq_string_slot_node): Likewise.
22703 * data-streamer-out.c: Update output_block::string_hash_table
22704 dependent calls and types.
22706 * dwarf2cfi.c (htab_t trace_index):
22707 Change type to hash_table. Update dependent calls and types.
22709 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
22710 Change type to hash_table. Update dependent calls and types.
22711 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
22712 (htab_t optimize_external_refs::map): Likewise.
22713 (htab_t output_comp_unit::extern_map): Likewise.
22714 (htab_t output_comdat_type_unit::extern_map): Likewise.
22715 (htab_t output_macinfo::macinfo_htab): Likewise.
22716 (htab_t optimize_location_lists::htab): Likewise.
22717 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
22719 * except.c (htab_t ehspec_hash_type):
22720 Change type to hash_table. Update dependent calls and types.
22721 (assign_filter_values::ttypes): Likewise.
22722 (assign_filter_values::ehspec): Likewise.
22723 (sjlj_assign_call_site_values::ar_hash): Likewise.
22724 (convert_to_eh_region_ranges::ar_hash): Likewise.
22726 * gcse.c (htab_t pre_ldst_table):
22727 Change type to hash_table. Update dependent calls and types.
22729 * ggc-common.c (htab_t saving_htab):
22730 Change type to hash_table. Update dependent calls and types.
22731 (htab_t loc_hash): Likewise.
22732 (htab_t ptr_hash): Likewise.
22733 (call_count): Rename ggc_call_count.
22734 (call_alloc): Rename ggc_call_alloc.
22735 (loc_descriptor): Rename make_loc_descriptor.
22736 (add_statistics): Rename ggc_add_statistics.
22738 * ggc-common.c (saving_htab):
22739 Change type to hash_table. Update dependent calls and types.
22741 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
22742 (push_gimplify_context): Likewise.
22743 (pop_gimplify_context): Likewise.
22744 (struct gimple_temp_hash_elt): Added.
22745 (struct gimplify_hasher): Likewise.
22746 (struct gimplify_ctx.temp_htab):
22747 Change type to hash_table. Update dependent calls and types.
22749 * gimple-fold.c: Include gimplify-ctx.h.
22751 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
22752 Change type to hash_table. Update dependent calls and types.
22753 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
22754 avoid potential global name collision.
22756 * gimplify.c: Include gimplify-ctx.h.
22757 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
22758 (htab_t gimplify_ctx::temp_htab):
22759 Update dependent calls and types for new type hash_table.
22760 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
22761 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
22763 * gimplify-ctx.h: New.
22764 (struct gimple_temp_hash_elt): Move from gimplify.c.
22765 (class gimplify_hasher): New.
22766 (struct gimplify_ctx): Move from gimple.h.
22767 (htab_t gimplify_ctx::temp_htab):
22768 Change type to hash_table. Update dependent calls and types.
22770 * graphite-clast-to-gimple.c: Include graphite-htab.h.
22771 (htab_t ivs_params::newivs_index):
22772 Change type to hash_table. Update dependent calls and types.
22773 (htab_t ivs_params::params_index): Likewise.
22774 (htab_t print_generated_program::params_index): Likewise.
22775 (htab_t gloog::newivs_index): Likewise.
22776 (htab_t gloog::params_index): Likewise.
22778 * graphite.c: Include graphite-htab.h.
22779 4htab_t graphite_transform_loops::bb_pbb_mapping):
22780 Change type to hash_table. Update dependent calls and types.
22782 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
22783 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
22784 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
22786 * graphite-dependences.c: Include graphite-htab.h.
22787 (loop_is_parallel_p): Change hash table type of parameter.
22789 * graphite-htab.h: New.
22790 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
22791 (extern find_pbb_via_hash): Move from graphite-poly.h.
22792 (extern loop_is_parallel_p): Move from graphite-poly.h.
22793 (extern get_loop_body_pbbs): Move from graphite-poly.h.
22795 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
22796 (extern loop_is_parallel_p): Move to graphite-htab.h.
22797 (extern get_loop_body_pbbs): Move to graphite-htab.h.
22799 * haifa-sched.c (htab_t delay_htab):
22800 Change type to hash_table. Update dependent calls and types.
22801 (htab_t delay_htab_i2): Likewise.
22803 * ira-color.c (htab_t allocno_hard_regs_htab):
22804 Change type to hash_table. Update dependent calls and types.
22806 * ira-costs.c (htab_t cost_classes_htab):
22807 Change type to hash_table. Update dependent calls and types.
22809 * loop-invariant.c (htab_t merge_identical_invariants::eq):
22810 Change type to hash_table. Update dependent calls and types.
22812 * loop-iv.c (htab_t bivs):
22813 Change type to hash_table. Update dependent calls and types.
22815 * loop-unroll.c (htab_t opt_info::insns_to_split):
22816 Change type to hash_table. Update dependent calls and types.
22817 (htab_t opt_info::insns_with_var_to_expand): Likewise.
22819 * lto-streamer.h (struct string_slot): Move from data-streamer.h
22820 (struct string_slot_hasher): New.
22821 (htab_t output_block::string_hash_table):
22822 Change type to hash_table. Update dependent calls and types.
22824 * lto-streamer-in.c (freeing_string_slot_hasher): New.
22825 (htab_t file_name_hash_table):
22826 Change type to hash_table. Update dependent calls and types.
22828 * lto-streamer-out.c: Update output_block::string_hash_table dependent
22831 * lto-streamer.c (htab_t tree_htab):
22832 Change type to hash_table. Update dependent calls and types.
22834 * omp-low.c: Include gimplify-ctx.h.
22836 * passes.c (htab_t name_to_pass_map):
22837 Change type to hash_table. Update dependent calls and types.
22838 (pass_traverse): Rename to passes_pass_traverse.
22840 * plugin.c (htab_t event_tab):
22841 Change type to hash_table. Update dependent calls and types.
22843 * postreload-gcse.c (htab_t expr_table):
22844 Change type to hash_table. Update dependent calls and types.
22845 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
22847 * sese.c (debug_rename_map_1): Make extern.
22848 (htab_t copy_bb_and_scalar_dependences::rename_map):
22849 Change type to hash_table. Update dependent calls and types.
22851 * sese.h (extern debug_rename_map): Move to .c file.
22853 * store-motion.c (htab_t store_motion_mems_table):
22854 Change type to hash_table. Update dependent calls and types.
22856 * trans-mem.c (htab_t tm_new_mem_hash):
22857 Change type to hash_table. Update dependent calls and types.
22859 * tree-browser.c (htab_t TB_up_ht):
22860 Change type to hash_table. Update dependent calls and types.
22862 * tree-cfg.c (htab_t discriminator_per_locus):
22863 Change type to hash_table. Update dependent calls and types.
22865 * tree-complex.c: Include tree-hasher.h
22866 (htab_t complex_variable_components):
22867 Change type to hash_table. Update dependent calls and types.
22869 * tree-eh.c (htab_t finally_tree):
22870 Change type to hash_table. Update dependent calls and types.
22872 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22873 struct int_tree_hasher.
22874 (extern int_tree_map_eq): Likewise.
22875 (uid_decl_map_hash): Removed.
22876 (extern decl_tree_map_eq): Likewise.
22878 * tree-hasher.h: New.
22879 (struct int_tree_hasher): New.
22880 (typedef int_tree_htab_type): New.
22882 * tree-inline.c: Include gimplify-ctx.h.
22884 * tree-mudflap.c: Include gimplify-ctx.h.
22886 * tree-parloops.c: Include tree-hasher.h.
22887 (htab_t eliminate_local_variables_stmt::decl_address):
22888 Change type to hash_table. Update dependent calls and types.
22889 (htab_t separate_decls_in_region::decl_copies): Likewise.
22891 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22892 Change type to hash_table. Update dependent calls and types.
22894 * tree-sra.c (candidates):
22895 Change type to hash_table. Update dependent calls and types.
22897 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22899 (int_tree_map_hash): Likewise.
22901 * tree-ssa-dom.c (htab_t avail_exprs):
22902 Change type to hash_table. Update dependent calls and types.
22904 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22905 Change type to hash_table. Update dependent calls and types.
22907 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22908 Change type to hash_table. Update dependent calls and types.
22910 * tree-ssa-phiopt.c (seen_ssa_names):
22911 Change type to hash_table. Update dependent calls and types.
22913 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22914 Change type to hash_table. Update dependent calls and types.
22916 * tree-ssa-uncprop.c (equiv):
22917 Change type to hash_table. Update dependent calls and types.
22919 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22921 PR rtl-optimization/57003
22922 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22923 call note_stores with kill_clobbered_value callback again after
22924 killing regs_invalidated_by_call.
22926 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22928 * config/aarch64/aarch64-simd.md
22929 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22930 (aarch64_simd_bsl<mode>): Likewise.
22931 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22933 2013-04-25 Marek Polacek <polacek@redhat.com>
22935 PR tree-optimization/57066
22936 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22938 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22940 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22942 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22944 * config/aarch64/aarch64-builtins.c
22945 (aarch64_fold_builtin): New.
22946 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22947 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22948 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22949 * config/aarch64/arm_neon.h
22950 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22952 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22953 Tejas Belagod <tejas.belagod@arm.com>
22955 * config/aarch64/aarch64-builtins.c
22956 (aarch64_gimple_fold_builtin): New.
22957 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22958 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22959 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22960 (addvv4sf): Update.
22961 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22963 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22965 * config/aarch64/aarch64.md
22966 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22968 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22970 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22971 (*ngcsi_uxtw): New pattern.
22973 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22974 Julian Brown <julian@codesourcery.com>
22976 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22977 (TB_DREG): Add T_V4HF.
22978 (v4hf_UP): New macro.
22979 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22980 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22981 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22983 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22984 (arm_vector_mode_supported_p): Handle V4HF.
22985 (arm_mangle_map): Handle V4HFmode.
22986 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22987 * config/arm/arm_neon_builtins.def: Add entries for
22988 vcvtv4hfv4sf, vcvtv4sfv4hf.
22989 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22990 (neon_vcvtv4hfv4sf): Likewise.
22991 * config/arm/neon-gen.ml: Handle half-precision floating point
22993 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22994 * config/arm/arm_neon.h: Regenerate.
22995 * config/arm/neon.ml (type elts): Add F16.
22996 (type vectype): Add T_float16x4, T_floatHF.
22997 (type vecmode): Add V4HF.
22998 (type features): Add Requires_FP_bit feature.
22999 (elt_width): Handle F16.
23000 (elt_class): Likewise.
23001 (elt_of_class_width): Likewise.
23002 (mode_of_elt): Refactor.
23003 (type_for_elt): Handle F16, fix error messages.
23004 (vectype_size): Handle T_float16x4.
23005 (vcvt_sh): New function.
23006 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
23007 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
23008 (string_of_mode): Handle V4HF.
23009 * doc/arm-neon-intrinsics.texi: Regenerate.
23011 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
23013 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
23014 format specifier in 'X' case.
23016 2013-04-25 Alan Modra <amodra@gmail.com>
23019 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
23020 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
23021 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
23022 Repeat for many other rotate/shift and mask patterns using subregs.
23023 Name lshiftrt insns.
23024 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
23025 on WORDS_BIG_ENDIAN.
23027 2013-04-25 Alan Modra <amodra@gmail.com>
23029 * config.gcc: Support little-endian powerpc-linux targets.
23030 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
23031 (LINK_OS_LINUX_SPEC): Define.
23032 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
23033 Preserve MASK_LITTLE_ENDIAN.
23034 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
23035 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
23036 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
23037 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
23038 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
23039 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
23040 Correct fp word order for little-endian. Don't shift toc entries
23041 smaller than a word for little-endian.
23042 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
23043 (bswapdi2 splits): Correct low-part subreg for little-endian.
23044 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
23045 low/high where such is correct only for be.
23046 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
23047 little-endian for -mcall-aixdesc.
23049 2013-04-25 Alan Modra <amodra@gmail.com>
23051 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
23052 replace_equiv_address_nv.
23054 2013-04-25 Alan Modra <amodra@gmail.com>
23056 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
23058 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23061 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23062 * rtl.h (struct rtx_def): ...
23064 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23066 PR rtl-optimizations/57046
23067 * lra-constraints (split_reg): Set up lra_risky_transformations_p
23068 for multi-reg splits.
23070 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
23072 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
23074 2013-04-24 Sterling Augustine <saugustine@google.com>
23076 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
23077 (comp_dir_string, debug_str_dwo_section): New.
23078 (DEBUG_STR_DWO_SECTION): Rename to ...
23079 (DEBUG_DWO_STR_SECTION): ... this.
23080 (DEBUG_NORM_STR_SECTION): Delete.
23081 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
23082 (DEBUG_STR_DWO_SECTION_FLAGS): New.
23083 (find_AT_string): Move most logic to ...
23084 (find_AT_string_in_table): ... here. New.
23085 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
23086 add_skeleton_AT_string. Delete logic.
23087 (output_skeleton_debug_sections): Remove call to
23088 add_top_level_skeleton_die_attrs.
23089 (add_comp_dir_attribute): Move logic to comp_dir_string.
23090 (dwarf2out_init): Initialize debug_str_dwo_section.
23091 (output_indirect_string): Call find_string_form.
23092 (output_indirect_strings): Rewrite.
23093 (prune_unused_types): Empty skeleton_debug_str_hash.
23094 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
23095 (dwarf2out_finish): Call output_indirect_strings.
23097 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
23099 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
23101 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23103 * rtl.h (struct rtx_def): Add comment for field jump.
23104 (LRA_SUBREG_P): New macro.
23105 * recog.c (register_operand): Check LRA_SUBREG_P.
23106 * lra.c (lra): Add note at the end of RTL code. Align non-empty
23108 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
23109 (lra_final_code_change): Skip subreg change for operators.
23110 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
23111 if there are no operand changes.
23112 * lra-constraints.c (curr_insn_set): New.
23113 (match_reload): Set LRA_SUBREG_P.
23114 (emit_spill_move): Ditto.
23115 (check_and_process_move): Use curr_insn_set. Process only single
23116 set insns. Don't initialize sec_mem_p and change_p.
23117 (simplify_operand_subreg): Use LRA_SUBREG_P.
23118 (reg_in_class_p): New function.
23119 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
23120 of #ifdef. Add code to remove cycling.
23121 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
23122 non-null disp. Reload inner instead of disp when base and index
23123 are null. Try to put lo_sum into register.
23124 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
23125 (check_and_process_move): Move code for move cost check to
23126 simple_move_p. Remove equiv_substitution.
23127 (simple_move_p): New function.
23128 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
23129 curr_insn_set. Call check_and_process_move only for single set
23130 insns. Use the new function. Move call of check_and_process_move
23131 after operand equiv substitution and address process.
23133 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
23135 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
23136 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
23137 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
23139 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
23141 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
23143 2013-04-24 Marek Polacek <polacek@redhat.com>
23145 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
23146 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
23147 (select_loops_exit_conditions): Likewise.
23148 (number_of_iterations_for_all_loops): Likewise.
23149 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
23150 (scev_analysis): Likewise.
23152 2013-04-02 Catherine Moore <clm@codesourcery.com>
23153 Chao-ying Fu <fu@mips.com>
23155 * config/mips/micromips.md (jraddiusp): New pattern.
23156 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
23157 instruction if possible.
23159 2013-04-24 Alan Modra <amodra@gmail.com>
23161 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
23163 2013-04-24 Julian Brown <julian@codesourcery.com>
23164 Chung-Lin Tang <cltang@codesourcery.com>
23166 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
23167 dependency behavior in enumeration type DIE generation. Add TODO note
23168 to comments about future DW_FORM_sdata/udata re-work of related code.
23170 2013-04-23 Lawrence Crowl <crowl@google.com>
23172 * Makefile.in: Update as needed below.
23174 * hash-table.h (class hash_table):
23175 Correct many methods with parameter types compare_type to the correct
23176 value_type. (Correct code was unlikely to notice the change.)
23177 (hash_table::elements_with_deleted) New.
23178 (class hashtable::iterator): New.
23179 (hashtable::begin()): New.
23180 (hashtable::end()): New.
23181 (FOR_EACH_HASH_TABLE_ELEMENT): New.
23183 * statistics.c (statistics_hashes):
23184 Change type to hash_table. Update dependent calls and types.
23186 * tree-into-ssa.c (var_infos):
23187 Change type to hash_table. Update dependent calls and types.
23189 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
23190 Change type to hash_table. Update dependent calls and types.
23192 * tree-ssa-loop-im.c (struct mem_ref.refs):
23193 Change type to hash_table. Update dependent calls and types.
23195 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
23196 Change type to hash_table. Update dependent calls and types.
23198 * tree-ssa-sccvn.c (vn_tables_s::nary):
23199 Change type to hash_table. Update dependent calls and types.
23200 (vn_tables_s::phis): Likewise.
23201 (vn_tables_s::references): Likewise.
23203 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
23204 (vn_reference_eq): Update parameter and return types.
23206 * tree-ssa-structalias.c (pointer_equiv_class_table):
23207 Change type to hash_table. Update dependent calls and types.
23208 (location_equiv_class_table): Likewise.
23210 * tree-vect-data-refs.c: Consequential changes for making
23211 peeling a hash_table.
23213 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
23214 (destroy_loop_vec_info): Dependent hash_table update.
23216 * tree-vectorizer.h (peeling_htab):
23217 Change type to hash_table. Update dependent calls and types.
23219 2013-04-23 Shiva Chen <shiva0217@gmail.com>
23221 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
23222 to check the register content is equal or not.
23223 * lra-constraints.c (match_reload): Use lra_assign_reg_val
23224 to assign register content record.
23225 * lra-eliminations.c (update_reg_eliminate): Use
23226 lra_update_reg_val_offset to update register content offset.
23227 * lra-int.h (struct lra_reg): Add offset member.
23228 (lra_reg_val_equal_p): New static inline function.
23229 (lra_update_reg_val_offset): New static inline function.
23230 (lra_assign_reg_val): New static inline function.
23231 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
23232 to assign register content record.
23233 (initialize_lra_reg_info_element): Initial offset to zero.
23235 2013-04-23 Catherine Moore <clm@codesourcery.com>
23237 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
23238 operands. Record compression.
23240 2013-04-23 Xinliang David Li <davidxl@google.com>
23242 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
23244 2013-04-23 Richard Biener <rguenther@suse.de>
23246 PR middle-end/57036
23247 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
23248 parameter, only add abnormal goto edges from the copied body
23249 if the call could perform abnormal gotos.
23250 (copy_cfg_body): Adjust.
23252 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
23254 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
23256 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
23258 * coretypes.h (gimple_stmt_iterator): Add struct to make
23261 2013-04-23 Richard Biener <rguenther@suse.de>
23263 PR tree-optimization/57026
23264 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
23265 from SSA names occuring in abnormal PHI nodes.
23267 2013-04-22 Andi Kleen <ak@linux.intel.com>
23269 * lto/lto.c (print_lto_report_1): Fix LTO report names.
23271 2013-04-22 Andi Kleen <ak@linux.intel.com>
23273 * lto/lto.c (print_lto_report_1): Declare early.
23274 (read_cgraph_and_symbols): Call print_lto_report_1 early.
23276 2013-04-22 Andi Kleen <ak@linux.intel.com>
23278 * common.opt (-flto-report-wpa): Add.
23279 * doc/invoke.texi (-flto-report-wpa): Add.
23280 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
23283 2013-04-22 Xinliang David Li <davidxl@google.com>
23285 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
23286 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
23287 * Makefile.in: New dependency
23289 David Daney <ddaney.cavm@gmail.com>
23291 * configure.ac (gcc_cv_as_micromips_support): Use the
23292 --fatal-warnings option.
23293 * configure: Regenerate.
23295 2013-04-22 Marek Polacek <polacek@redhat.com>
23298 * tsan.c (instrument_expr): Don't instrument expression
23299 in case its size is zero.
23301 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
23305 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23307 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23309 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23311 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
23312 (gimple_stmt_iterator): New typedef.
23313 * gimple.h (gimple_stmt_iterator): Rename to...
23314 (gimple_stmt_iterator_d): ... This.
23315 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
23316 trees be valid for GIMPLE and GENERIC.
23317 (TARGET_GIMPLE_FOLD_BUILTIN): New.
23318 * gimple-fold.c (gimple_fold_call): Call target hook
23319 gimple_fold_builtin.
23320 * hooks.c (hook_bool_gsiptr_false): New.
23321 * hooks.h (hook_bool_gsiptr_false): New.
23322 * target.def (fold_stmt): New.
23323 * doc/tm.texi: Regenerate.
23325 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
23328 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
23329 a set sp if no stack realignment.
23331 2013-04-22 Nick Clifton <nickc@redhat.com>
23333 * config.gcc (tilegx-linux): Extend extra_objs rather than
23335 (tilepro-linux): Likewise.
23337 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23339 * config/aarch64/aarch64-builtins.c
23341 (CF0, CF1, CF2, CF3, CF4, CF10): New.
23342 (VAR<1-12>): Add MAP parameter.
23343 (BUILTIN_*): Likewise.
23344 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
23345 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
23346 (aarch64_ushl_n<mode>): Likewise.
23347 (aarch64_sshr_n<mode>): Likewise.
23348 (aarch64_ushr_n<mode>): Likewise.
23349 (aarch64_<maxmin><mode>): Likewise.
23350 (aarch64_sqrt<mode>): Likewise.
23351 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
23352 (vshr<q>_n_*): Likewise.
23354 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23356 * config/aarch64/aarch64-builtins.c
23357 (aarch64_simd_builtin_type_mode): Handle SF types.
23359 (BUILTIN_GPF): Define.
23360 (aarch64_init_simd_builtins): Handle SF types.
23361 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
23362 (frecps): Likewise.
23363 (frecpx): Likewise.
23364 * config/aarch64/aarch64-simd.md
23365 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
23366 (aarch64_frecpe<mode>): New.
23367 (aarch64_frecps<mode>): Likewise.
23368 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
23369 (v8type): Add frecp<esx>.
23370 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
23371 (aarch64_frecps<mode>): Likewise.
23372 * config/aarch64/iterators.md (FRECP): New.
23373 (frecp_suffix): Likewise.
23374 * config/aarch64/arm_neon.h
23375 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
23377 2013-04-22 Christian Bruel <christian.bruel@st.com>
23380 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
23381 (REG_CLASS_NAMES): Idem.
23382 (REG_CLASS_CONTENTS): Idem.
23383 (REGCLASS_HAS_FP_REG): Idem.
23384 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
23385 (sh_conditional_register_usage): Idem.
23387 2013-04-21 Jeff Law <law@redhat.com>
23389 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
23390 (ssa_forward_propagate_and_combine): Use it.
23392 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
23394 * lra.c: Update the flow chart diagram.
23396 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
23398 PR rtl-optimization/56847
23399 * lra-constraints.c (process_alt_operands): Discourage alternative
23400 with non-matche doffsettable memory constraint fro memory with
23403 2013-04-19 Richard Biener <rguenther@suse.de>
23405 PR tree-optimization/56982
23406 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
23408 * gimplify.c (gimplify_call_expr): Notice special calls.
23409 (gimplify_modify_expr): Likewise.
23410 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
23411 abnormal control flow receivers.
23412 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
23413 in the same way as cfun->has_nonlocal_labels.
23414 (gimple_purge_dead_abnormal_call_edges): Likewise.
23415 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
23416 receivers start a basic-block.
23418 2013-04-19 Richard Biener <rguenther@suse.de>
23420 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
23422 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
23423 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
23424 (SLP_TREE_LOAD_PERMUTATION): Add.
23425 (vect_transform_slp_perm_load): Adjust prototype.
23426 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
23427 (vect_free_slp_instance): Likewise.
23428 (vect_create_new_slp_node): Likewise.
23429 (vect_supported_slp_permutation_p): Remove.
23430 (vect_slp_rearrange_stmts): Adjust.
23431 (vect_supported_load_permutation_p): Likewise. Inline
23432 vect_supported_slp_permutation_p here.
23433 (vect_analyze_slp_instance): Compute load permutations per
23434 slp node instead of per instance.
23435 (vect_get_slp_defs): Adjust.
23436 (vect_transform_slp_perm_load): Likewise.
23437 (vect_schedule_slp_instance): Remove redundant code.
23438 (vect_schedule_slp): Remove hack for PR56270, add it ...
23439 * tree-vect-stmts.c (vectorizable_load): ... here, do not
23440 CSE loads for SLP. Adjust.
23442 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
23444 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
23445 spelling in two comments.
23447 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
23450 * config/arm/arm.c (load_multiple_sequence): Require SP
23451 as base register for loads if SP is in the register list.
23453 2013-04-19 Martin Jambor <mjambor@suse.cz>
23455 PR tree-optimization/56718
23456 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
23457 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
23458 and made public. Adjusted all callers.
23459 (ipa_intraprocedural_devirtualization): New function.
23460 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
23461 (ipa_intraprocedural_devirtualization): Likewise.
23462 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
23464 2013-04-19 Richard Biener <rguenther@suse.de>
23466 PR tree-optimization/57000
23467 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
23469 2013-04-19 Terry Guo <terry.guo@arm.com>
23471 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
23473 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
23474 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
23475 (cortex_m4_fmacs): Use new reservations.
23476 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
23478 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
23480 PR rtl-optimization/56999
23481 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
23483 (lra_coalesce): Remove split_origin_bitmap and related code.
23484 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
23485 ranges if necessary.
23487 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
23489 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
23491 (ix86_expand_call): Remove clobbered_registers array and use
23492 x86_64_ms_sysv_extra_clobbered_registers instead.
23493 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
23495 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
23497 * config/i386/i386.md (*call_rex64_ms_sysv): Use
23498 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
23499 (*call_value_rex64_ms_sysv): Ditto.
23501 2013-04-18 Cary Coutant <ccoutant@google.com>
23503 * dwarf2out.c (output_pubnames): Check die_perennial_p of
23504 parent instead of die_mark.
23506 2013-04-18 Diego Novillo <dnovillo@google.com>
23508 * gimple.c (create_gimple_tmp): New.
23509 (get_expr_type): New.
23510 (build_assign): New.
23511 (build_type_cast): New.
23512 * gimple.h (enum ssa_mode): Define.
23513 (gimple_seq_set_location): New.
23514 * asan.c (build_check_stmt): Change some gimple_build_* calls
23515 to use build_assign and build_type_cast.
23517 2013-04-18 Richard Biener <rguenther@suse.de>
23519 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
23520 handle negative step. Remove redundant checks.
23521 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
23522 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
23523 for negative step and grouped loads fail to vectorize.
23525 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23527 * emit-rtl.c (reset_insn_used_flags): New function.
23528 (reset_all_used_flags): Use it.
23529 (verify_insn_sharing): New function.
23530 (verify_rtl_sharing): Fix verification for SEQUENCEs.
23532 2013-04-18 Jakub Jelinek <jakub@redhat.com>
23534 PR tree-optimization/56984
23535 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
23536 and (x >> M) >= N don't register any assertion if N << M is the
23539 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23541 * lower-subreg.c (resolve_simple_move): If called self-recursive,
23542 do not delete_insn insns that have not yet been emitted, only
23543 unlink them with remove_insn.
23544 * df-scan.c (df_insn_delete): Revert r197492.
23546 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
23548 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
23549 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
23551 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23553 * config/arm/arm.md (movsicc_insn): Convert define_insn into
23554 define_insn_and_split.
23555 (and_scc,ior_scc,negscc): Likewise.
23556 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
23558 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23560 * config/arm/arm.c (use_return_insn): Return 0 for targets that
23561 can benefit from using a sequence of LDRD instructions in epilogue
23562 instead of a single LDM instruction.
23564 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
23567 * doc/extend.texi: Fix typo.
23569 2013-04-17 Richard Biener <rguenther@suse.de>
23571 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
23572 (vect_build_slp_tree): ... here.
23573 (vect_build_slp_tree_1): Compute which stmts of the SLP group
23574 match. Remove special-casing of mismatched complex loads.
23575 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
23576 re-try the match with swapped commutative operands.
23577 (vect_supported_load_permutation_p): Remove special-casing of
23578 mismatched complex loads.
23579 (vect_analyze_slp_instance): Adjust.
23581 2013-04-17 Richard Biener <rguenther@suse.de>
23583 PR rtl-optimization/56921
23584 * cfgloop.h (struct loop): Add simple_loop_desc member.
23585 (struct niter_desc): Mark with GTY(()).
23586 (simple_loop_desc): Do not use aux field but simple_loop_desc.
23587 * loop-iv.c (get_simple_loop_desc): Likewise.
23588 (free_simple_loop_desc): Likewise.
23591 2013-04-16 Richard Biener <rguenther@suse.de>
23593 PR rtl-optimization/56921
23594 * loop-init.c (pass_rtl_move_loop_invariants): Add
23595 TODO_do_not_ggc_collect to todo_flags_finish.
23596 (pass_rtl_unswitch): Same.
23597 (pass_rtl_unroll_and_peel_loops): Same.
23598 (pass_rtl_doloop): Same.
23600 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
23602 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
23603 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
23604 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
23606 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
23607 * tree-streamer.c (record_common_node): Adjust reference in comment.
23609 2013-04-17 Terry Guo <terry.guo@arm.com>
23611 * config/arm/cortex-m4.md: Add a new bypass.
23613 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23615 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
23617 (*subs_<optab><mode>_multp2): New pattern.
23618 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23619 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23621 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23623 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
23624 (*subs_mul_imm_<mode>): New pattern.
23626 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
23629 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
23630 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
23631 (vsx_movti_32bit): Change j->wa to O->wa.
23633 2013-04-16 Richard Biener <rguenther@suse.de>
23635 PR rtl-optimization/56921
23636 * loop-init.c (pass_rtl_move_loop_invariants): Add
23637 TODO_do_not_ggc_collect to todo_flags_finish.
23638 (pass_rtl_unswitch): Same.
23639 (pass_rtl_unroll_and_peel_loops): Same.
23640 (pass_rtl_doloop): Same.
23642 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
23644 * config/arm/arm.c (emit_multi_reg_push): New declaration
23645 for an existing function.
23646 (arm_emit_strd_push): New function.
23647 (arm_expand_prologue): Used here.
23648 (arm_emit_ldrd_pop): New function.
23649 (arm_expand_epilogue): Used here.
23650 (arm_get_frame_offsets): Update condition.
23651 (arm_emit_multi_reg_pop): Add a special case for load of a single
23652 register with writeback.
23654 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
23656 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
23659 2013-04-16 Richard Biener <rguenther@suse.de>
23661 PR tree-optimization/56756
23662 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
23663 (first_mem_ref_loc): New.
23664 (execute_sm): Place the load temporarily before a previous
23665 access instead of in the latch edge to ensure its SSA dependencies
23666 are defined at points dominating the load.
23668 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
23670 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
23671 correct fix by moving header and footer insn to the footer of
23672 the merged basic block. Clear BB_END of the merged-away block.
23674 PR middle-end/43631
23675 * emit-rtl.c (make_note_raw): New function.
23676 (link_insn_into_chain): New static inline function.
23677 (add_insn): Use it.
23678 (add_insn_before, add_insn_after): Factor insn chain linking code...
23679 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
23680 using link_insn_into_chain.
23681 (note_outside_basic_block_p): New helper function for emit_note_after
23682 and emit_note_before.
23683 (emit_note_after): Use nobb variant of add_insn_after if the note
23684 should not be contained in a basic block.
23685 (emit_note_before): Use nobb variant of add_insn_before if the note
23686 should not be contained in a basic block.
23687 (emit_note_copy): Use make_note_raw.
23688 (emit_note): Likewise.
23689 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
23690 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
23691 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
23692 the moved barrier the tail of the basic block it follows.
23693 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
23695 2013-04-15 Jakub Jelinek <jakub@redhat.com>
23697 PR tree-optimization/56962
23698 * gimple-ssa-strength-reduction.c (record_increment): Only set
23699 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
23700 either rhs1 or rhs2 is equal to c->base_expr.
23702 2013-04-15 Richard Biener <rguenther@suse.de>
23704 PR tree-optimization/56933
23705 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
23707 (GROUP_READ_WRITE_DEPENDENCE): Remove.
23708 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
23709 * tree-vect-data-refs.c (vect_analyze_group_access): Move
23710 dependence check ...
23711 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
23713 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
23714 GROUP_READ_WRITE_DEPENDENCE.
23716 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23718 * emit-rtl.c (reset_all_used_flags): New function.
23719 (verify_rtl_sharing): Call reset_all_used_flags before and after
23720 performing the checks.
23722 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23724 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
23725 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
23726 * config/arm/constraints.md (De): New constraint.
23727 * config/arm/neon.md (anddi3_neon): Delete.
23728 (neon_vand<mode>): Expand to standard anddi3 pattern.
23729 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
23730 Move earlier in the file.
23731 (neon_inv_logic_op2): Likewise.
23732 (arm_anddi_operand_neon): New predicate.
23734 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23736 * configure.ac (gcc_cv_ld_as_needed): Set
23737 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
23738 Use -z ignore, -z record on *-*-solaris2*.
23739 (HAVE_LD_AS_NEEDED): Update comment.
23740 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
23741 * configure: Regenerate.
23742 * config.in: Regenerate.
23743 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
23744 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
23745 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
23746 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
23747 equivalents. Fix markup.
23748 * doc/tm.texi: Regenerate.
23750 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
23752 * config/i386/i386.opt: New option mstack-protector-guard=.
23753 * config/i386/i386-opts.h: Add enum stack_protector_guard.
23754 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
23755 TARGET_SSP_TLS_GUARD.
23756 * config/i386/i386.c (ix86_option_override_internal): Set
23757 ix86_stack_protector_guard.
23758 * config/i386/i386.md (stack_protect_set): Enable for
23759 TARGET_SSP_TLS_GUARD only.
23760 (stack_protect_set_<mode>): Ditto.
23761 (stack_protect_test): Ditto.
23762 (stack_protect_test_<mode>): Ditto.
23763 * doc/invoke.texi (i386 Option): Document.
23765 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
23768 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
23769 (S_MODES): Set H_MODE bit.
23770 (SF_MODES): Set only S_MODE and SF_MODE bits.
23771 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
23772 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
23773 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
23774 <MODE_FLOAT>: Likewise.
23776 2013-04-15 Joey Ye <joey.ye@arm.com>
23778 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
23780 2013-04-15 Joey Ye <joey.ye@arm.com>
23782 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
23784 (thumb_far_jump_used_p): Count instruction size and set
23787 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
23789 * reorg.c (fill_simple_delay_slots): Reindent block of code.
23790 * resource.c (mark_target_live_regs): Reformat conditional block.
23792 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23794 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
23795 notes, they are emitted only just before final.
23796 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
23798 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23800 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
23801 * cfgrtl.c (delete_insn): Call it here instead.
23802 * lra-spills.c (lra_final_code_change): Use delete_insn.
23803 * haifa-sched.c (sched_remove_insn): Likewise.
23804 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
23805 returning to the nop pool.
23806 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
23807 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
23809 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23811 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
23812 * doc/tm.texi: Regenerated.
23814 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
23816 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
23819 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23821 * df-core.c (df_find_def): Compare register numbers.
23822 (df_find_use): Likewise.
23824 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
23827 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
23828 lra_in_progress for return.
23830 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23832 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
23833 define_insn into define_insn_and_split and emit movsicc patterns.
23835 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23837 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
23839 2013-04-12 Richard Biener <rguenther@suse.de>
23841 * tree-pass.h (TODO_do_not_ggc_collect): New.
23842 * passes.c (execute_one_ipa_transform_pass): Honor
23843 TODO_do_not_ggc_collect.
23844 (execute_one_pass): Likewise.
23847 2013-04-10 Richard Biener <rguenther@suse.de>
23849 * passes.c (init_optimization_passes): Remove reload pass.
23850 * ira.c (do_reload): Merge into ...
23852 (rest_of_handle_reload): Remove.
23853 (pass_reload): Likewise.
23854 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23855 of reload for vzeroupper pass placement.
23857 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23859 PR tree-optimization/56918
23860 PR tree-optimization/56920
23861 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23862 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23863 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23864 use wide_mul_with_sign method.
23866 2013-04-12 Richard Biener <rguenther@suse.de>
23868 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23869 not be considered a gimple constant.
23871 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23873 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23874 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23877 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23878 Jakub Jelinek <jakub@redhat.com>
23880 * opts.c: Include diagnostic-color.h.
23881 (common_handle_option): Handle OPT_fdiagnostics_color_.
23882 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23883 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23884 (diagnostic-color.o): New.
23885 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23886 (diagnostic_color_rule): New enum.
23887 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23888 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23889 the location string.
23890 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23891 either NULL, or color kind.
23892 * diagnostic-color.c: New file.
23893 * diagnostic-color.h: New file.
23894 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23896 * doc/invoke.texi (-fdiagnostics-color): Document.
23897 * pretty-print.h (pp_show_color): Define.
23898 (struct pretty_print_info): Add show_color field.
23899 * diagnostic.c: Include diagnostic-color.h.
23900 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23901 macros. Colorize error:, warning: etc. strings and also the location
23903 (diagnostic_show_locus): Colorize the caret line.
23904 * pretty-print.c: Include diagnostic-color.h.
23905 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23906 inside of %< %> quotes or quoted through q format modifier.
23908 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23910 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23912 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23914 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23915 code in CC_NZ mode.
23916 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23919 2013-04-11 Marek Polacek <polacek@redhat.com>
23921 PR tree-optimization/48184
23922 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23924 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23926 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23927 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23928 (skip_simple_arithmetic): Tidy up.
23929 * tree.h (skip_simple_constant_arithmetic): Declare.
23931 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23933 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23935 2013-04-11 Richard Biener <rguenther@suse.de>
23937 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23938 generate vector constants.
23940 2013-04-11 Richard Biener <rguenther@suse.de>
23942 PR tree-optimization/56878
23943 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23944 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23946 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23947 Prefer to align the DR with the most invariant base address.
23949 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23951 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23954 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23956 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23957 floating-point vector comparisons against 0.
23959 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23961 PR tree-optimization/56899
23962 * fold-const.c (extract_muldiv_1): Apply distributive law
23963 only if TYPE_OVERFLOW_WRAPS (ctype).
23965 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23968 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23969 loading has side effect.
23971 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23973 * config/sparc/sparc.c: Include tree-pass.h.
23974 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23975 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23976 head of file. Change return type. Split off gate function.
23977 (sparc_gate_work_around_errata): New function.
23978 (pass_work_around_errata): New pass definition.
23979 (insert_pass_work_around_errata) New pass insert definition to
23980 insert pass_work_around_errata just after delayed-branch scheduling.
23981 (sparc_option_override): Insert the pass.
23982 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23984 2013-04-10 David S. Miller <davem@davemloft.net>
23986 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23987 or -mcpu=hypersparc.
23989 * target.def (cstore_mode): New hook.
23990 * target.h: Include insn-codes.h
23991 * targhooks.c: Likewise.
23992 (default_cstore_mode): New function.
23993 * targhooks.h: Declare it.
23994 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23995 * doc/tm.texi: Rebuild.
23996 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23997 target hook, rather than inspecting the insn_data.
23998 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23999 (TARGET_CSTORE_MODE): Redefine.
24000 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
24002 * config/sparc/predicates.md (cstore_result_operand): New special
24004 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
24005 Use it for operand 0.
24006 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
24007 (*snesi_special): Likewise.
24008 (*snesi_zero): Likewise.
24009 (*seqsi_zero): Likewise.
24010 (*sltu_insn): Likewise.
24011 (*sgeu_insn): Likewise.
24012 (*seqdi_special): Make operand 0 and comparison operation be of
24014 (*snedi_special): Likewise.
24015 (*snedi_special_vis3): Likewise.
24016 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
24017 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
24018 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
24019 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
24020 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
24021 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
24022 (*sltu_extend_sp64): Likewise.
24023 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
24024 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
24025 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
24026 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
24027 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
24029 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
24031 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
24032 (aarch64_start_file): Use the new function.
24034 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
24035 Jason Merrill <jason@redhat.com>
24037 * common.opt: Add -gdwarf.
24038 * opts.c (common_handle_option): Handle it.
24039 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
24041 2013-04-10 Richard Biener <rguenther@suse.de>
24043 * passes.c (execute_todo): Do not call ggc_collect conditional here.
24044 (execute_one_ipa_transform_pass): But unconditionally here.
24045 (execute_one_pass): And here.
24046 (init_optimization_passes): Remove reload pass.
24047 * tree-pass.h (TODO_ggc_collect): Remove.
24048 (pass_reload): Likewise.
24049 * ira.c (do_reload): Merge into ...
24051 (rest_of_handle_reload): Remove.
24052 (pass_reload): Likewise.
24053 * config/i386/i386.c (ix86_option_override): Refer to ira instead
24054 of reload for vzeroupper pass placement.
24055 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
24056 and todo_flags_finish of all passes.
24058 2013-04-10 Richard Biener <rguenther@suse.de>
24060 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
24061 first_const_oprnd field, rename first_def_type to first_op_type.
24062 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
24063 (vect_get_and_check_slp_defs): Always use the type of the
24064 operand. Allow mixed vect_external_def, vect_constant_def types.
24065 (vect_get_constant_vectors): Handle mixed vect_external_def,
24066 vect_constant_def types.
24068 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
24070 PR tree-optimization/55524
24071 * tree-ssa-math-opts.c
24072 (convert_mult_to_fma): Don't use an fms construct
24073 when we don't have an fms operation, but fnma, and it looks
24074 likely that we'll be able to use the latter.
24076 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
24078 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
24080 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
24081 inline fail caused by overwritable functions.
24083 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
24085 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
24086 unnecessary bits in the constant power of two case.
24088 2013-04-10 Richard Biener <rguenther@suse.de>
24090 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
24091 broken code swapping operands.
24092 (vect_build_slp_tree): Do not compute load permutations here.
24093 (vect_analyze_slp_instance): Compute load permutations here,
24094 after building the SLP tree.
24096 2013-04-09 Christian Bruel <christian.bruel@st.com>
24098 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
24099 of next/prev_real_insn.
24101 2013-04-09 Jan Hubicka <jh@suse.cz>
24103 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
24104 Drop aliased parameter.
24105 (function_and_variable_visibility): Do not handle alias pairs.
24106 * cgraph.c (varpool_externally_visible_p): Update prototype.
24107 * varpool.c (varpool_add_new_variable): Update.
24109 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24111 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
24113 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
24115 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
24117 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
24119 2013-04-09 Marek Polacek <polacek@redhat.com>
24121 PR tree-optimization/48762
24122 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
24124 2013-04-09 Richard Biener <rguenther@suse.de>
24126 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
24128 (vect_build_slp_tree): Likewise.
24129 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
24130 calculating the cost of a SLP instance.
24131 (vect_analyze_slp_instance): Use it from here, after building
24134 2013-04-09 Jakub Jelinek <jakub@redhat.com>
24136 PR middle-end/56883
24137 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
24138 expand_omp_for_static_chunk): Use simple_p = true in
24139 force_gimple_operand_gsi calls when assigning to addressable decls.
24141 2013-04-09 Jeff Law <law@redhat.com>
24143 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
24144 when the boolean was created by converting a wider object which
24145 had a boolean range.
24147 2013-04-09 Richard Biener <rguenther@suse.de>
24149 * tree-vectorizer.h (slp_void_p): Remove.
24150 (slp_tree): Typedef before _slp_tree declaration.
24151 (struct _slp_tree): Use a vector of slp_tree as children.
24152 (vect_get_place_in_interleaving_chain): Remove.
24153 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
24155 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
24157 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
24158 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
24159 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
24160 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
24161 Use slp_node instead of slp_void_p and adjust.
24163 2013-04-09 Richard Biener <rguenther@suse.de>
24165 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
24166 work that is not necessary.
24168 2013-04-09 Jakub Jelinek <jakub@redhat.com>
24170 PR tree-optimization/56854
24171 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
24172 forward into clobber stmts if it would change MEM_REF lhs into
24175 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
24177 * tree.c (type_hash_lookup, type_hash_add): Make static.
24178 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
24180 2013-04-09 Richard Biener <rguenther@suse.de>
24182 * tree.h (unsave_expr_now): Remove.
24183 * tree-inline.c (mark_local_for_remap_r): Remove.
24184 (unsave_expr_1): Likewise.
24185 (unsave_r): Likewise.
24186 (unsave_expr_now): Likewise.
24187 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
24188 (propagate_tree_value): Likewise.
24190 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
24192 * doc/rtl.texi (sequence): Rewrite documentation to match the
24193 current use of SEQUENCE rtl objects.
24194 * rtl.def (SEQUENCE): Likewise.
24196 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
24197 Update documentation.
24198 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
24199 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
24201 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
24203 2013-04-08 Teresa Johnson <tejohnson@google.com>
24205 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
24206 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
24208 (estimate_edge_size_and_time): Add comment to suggest using rounding
24210 (estimate_node_size_and_time): Ditto.
24211 (remap_edge_change_prob): Use helper rounding divide methods.
24212 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
24213 (gimple_mod_pow2_value_transform): Ditto.
24214 (gimple_mod_subtract_transform): Ditto.
24215 (gimple_ic_transform): Ditto.
24216 (gimple_stringops_transform): Ditto.
24217 * stmt.c (conditional_probability): Ditto.
24218 (emit_case_dispatch_table): Ditto.
24219 * lto-cgraph.c (merge_profile_summaries): Ditto.
24220 * tree-optimize.c (execute_fixup_cfg): Ditto.
24221 * cfgcleanup.c (try_forward_edges): Ditto.
24222 * cfgloopmanip.c (scale_loop_profile): Ditto.
24224 (duplicate_loop_to_header_edge): Ditto.
24225 (lv_adjust_loop_entry_edge): Ditto.
24226 * tree-vect-loop.c (vect_transform_loop): Ditto.
24227 * profile.c (compute_branch_probabilities): Ditto.
24228 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
24229 * lto-streamer-in.c (input_cfg): Ditto.
24230 * gimple-streamer-in.c (input_bb): Ditto.
24231 * ipa-cp.c (update_profiling_info): Ditto.
24232 (update_specialized_profile): Ditto.
24233 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
24234 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
24236 * sched-rgn.c (compute_dom_prob_ps): Ditto.
24237 (compute_trg_info): Ditto.
24238 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
24239 (purge_dead_edges): Ditto.
24240 * loop-unswitch.c (unswitch_loop): Ditto.
24241 * cgraphclones.c (cgraph_clone_edge): Ditto.
24242 (cgraph_clone_node): Ditto.
24243 * tree-inline.c (copy_bb): Ditto.
24244 (copy_edges_for_bb): Ditto.
24245 (initialize_cfun): Ditto.
24246 (copy_cfg_body): Ditto.
24247 (expand_call_inline): Ditto.
24249 2013-04-08 Kai Tietz <ktietz@redhat.com>
24251 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
24252 TARGET_CYGWIN64 by TARGET_64BIT.
24254 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
24256 * config/epiphany/epiphany.md (GPR_1): New constant.
24257 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
24258 * config/epiphany/epiphany.c (gen_compare_reg):
24259 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
24260 is already in place.
24261 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
24262 Don't require being called during rtl expansion; If y operlaps r0,
24264 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
24265 (epiphany_expand_epilogue): Likewise.
24267 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
24268 Don't use CC_FPmode for ORDERED / UNORDERED.
24269 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
24271 * config/epiphany/constraints.md (CnL): New constraint.
24272 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
24273 * config/epiphany/predicates.md (add_operand): Allow 1024.
24275 * config/epiphany/epiphany.md (logical_op): New code iterator.
24276 (op_mnc): New code attribute.
24277 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
24278 (mov_f+1, mov_f+2): New peephole2 patterns.
24280 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
24281 (cstoresi4): Also allow re-use of zero result when doing a NE
24282 comparison to a non-zero operand.
24283 Use (clobber (scratch)) for first insn if the gpr output is not needed.
24285 * config/epiphany/epiphany.md (<insn_opname>v2si3):
24286 Use gen_addsi3_i / gen_subsi3_i.
24288 2013-04-08 Jakub Jelinek <jakub@redhat.com>
24292 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
24293 contain anything but clobbers, at most one __builtin_stack_restore,
24294 optionally debug stmts and final resx, and if it has at least one
24295 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
24296 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
24297 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
24298 which isn't defaut definition, remove them.
24299 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
24300 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
24301 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
24302 with MEM_REF LHS with SSA_NAME address.
24304 2013-04-08 Jeff Law <law@redhat.com>
24306 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
24308 2013-04-08 Richard Biener <rguenther@suse.de>
24310 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
24312 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
24313 determined vector type.
24314 (vect_analyze_data_refs): Likewise.
24315 (vect_get_new_vect_var): Adjust.
24316 (vect_create_destination_var): Preserve SSA name versions.
24317 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
24318 not dump anything here.
24320 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
24322 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
24323 Add member lr_slot_known.
24324 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
24326 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
24327 Remove code that sets lr_slot_offset according to what a previous
24328 version of epiphany_emit_save_restore used to do.
24329 (epiphany_emit_save_restore): When doing an lr save or restore,
24330 set/verify lr_slot_known and lr_slot_offset.
24332 2013-04-08 Xinyu Qi <xyqi@marvell.com>
24335 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
24338 2013-04-08 Richard Biener <rguenther@suse.de>
24340 * alias.c (find_base_term): Fix thinko in previous change.
24342 2013-04-08 Jakub Jelinek <jakub@redhat.com>
24344 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
24345 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
24346 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
24347 if possible to compute val.
24348 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
24349 For QImode integers don't require anything about precision. Use
24350 const_with_all_bytes_same to find out if the constant doesn't have
24351 repeated bytes in it.
24353 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24355 * config/s390/s390.c (s390_expand_insv): Only accept insertions
24358 2013-04-08 Marek Polacek <polacek@redhat.com>
24360 PR rtl-optimization/48182
24361 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
24364 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24367 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
24368 nuses, make sure we have a label.
24370 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24373 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
24374 (rs6000_emit_swdiv_low_precision): Remove.
24375 (rs6000_emit_swdiv): Rewrite to handle between one and four
24376 iterations of Newton-Raphson generally; modify required number of
24377 iterations for some cases.
24378 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
24380 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
24382 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
24383 set-but-unused variable.
24385 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
24386 basic blocks of released function bodies garbage-collectable.
24388 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
24389 (struct rtl_opt_pass): Add TODO_df_finish.
24391 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
24393 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24395 * config/arm/constraints.md (q): New constraint.
24396 * config/arm/ldrdstrd.md: New file.
24397 * config/arm/arm.md (ldrdstrd.md) New include.
24398 (arm_movdi): Use "q" instead of "r" constraint
24399 for double-word memory access.
24400 (movdf_soft_insn): Likewise.
24401 * config/arm/vfp.md (movdi_vfp): Likewise.
24402 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
24403 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
24404 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
24405 (mem_ok_for_ldrd_strd): Likewise.
24406 (output_move_double): Update assertion.
24408 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24410 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
24412 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24414 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
24415 define_insn_and_split.
24416 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
24418 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24420 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
24421 define_insn_and_split.
24422 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
24423 (shiftsi3_compare): New pattern.
24424 (rrx): New pattern.
24425 * config/arm/unspecs.md (UNSPEC_RRX): New.
24427 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24429 * config/arm/arm.md (negdi_extendsidi): New pattern.
24430 (negdi_zero_extendsidi): Likewise.
24432 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24434 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
24435 define_insn_and_split.
24436 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
24437 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
24439 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24441 * config/arm/arm.md (arm_subdi3): Convert define_insn into
24442 define_insn_and_split.
24443 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
24444 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
24446 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24448 * config/arm/arm.md (subsi3_carryin): New pattern.
24449 (subsi3_carryin_const): Likewise.
24450 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
24451 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
24453 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24455 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
24457 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24459 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
24460 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
24462 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24464 * config/arm/arm.c (arm_expand_builtin): Change fcode
24465 type to unsigned int.
24467 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24469 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
24471 2013-04-04 Ian Lance Taylor <iant@google.com>
24473 * doc/standards.texi (Standards): The Go frontend supports the Go 1
24476 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
24478 PR middle-end/56729
24479 * df-scan.c (df_insn_delete): Disable failing assert.
24481 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24483 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
24484 New function prototype.
24485 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
24486 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
24487 (arm_builtin_vectorized_function): New function.
24489 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24491 * config/arm/arm_neon_builtins.def: New file.
24492 * config/arm/arm.c (neon_builtin_data): Move contents to
24493 arm_neon_builtins.def.
24494 (enum arm_builtins): Include neon builtin definitions.
24495 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
24496 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
24498 2013-04-04 Marek Polacek <polacek@redhat.com>
24500 PR tree-optimization/48186
24501 * predict.c (maybe_hot_frequency_p): Return false if
24502 HOT_BB_FREQUENCY_FRACTION is 0.
24503 (cgraph_maybe_hot_edge_p): Likewise.
24505 2013-04-04 Richard Biener <rguenther@suse.de>
24507 PR tree-optimization/56826
24508 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
24511 2013-04-04 Richard Biener <rguenther@suse.de>
24513 PR tree-optimization/56213
24514 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
24515 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
24517 2013-04-04 Richard Biener <rguenther@suse.de>
24519 PR tree-optimization/56837
24520 * tree-loop-distribution.c (classify_partition): For non-zero
24521 values require that the value has the same precision as its
24522 mode to be useful as memset value.
24524 2013-04-03 Nick Clifton <nickc@redhat.com>
24526 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
24527 (fmssf4): Use fmsf.s on E3V5 architectures.
24528 (fnmasf4): Use fnmaf.s on E3V5 architectures.
24529 (fnmssf4): Use fnmsf.s on E3V5 architectures.
24531 2013-04-03 Jeff Law <law@redhat.com>
24533 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
24534 (lra-eliminations.o): Likewise.
24536 2013-04-03 Teresa Johnson <tejohnson@google.com>
24538 * gcov-io.c (compute_working_sets): Moved most of body of old
24539 compute_working_sets here from profile.c.
24540 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
24541 (gcov_working_set_t): Moved typedef here from basic-block.h
24542 (compute_working_set): Declare.
24543 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
24544 (get_working_sets): Renamed from compute_working_set,
24545 replace most of body with call to new compute_working_sets.
24546 (get_exec_counts): Replace call to compute_working_sets
24547 to get_working_sets.
24548 * profile.h (get_working_sets): Renamed from compute_working_set.
24549 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
24550 to get_working_sets.
24551 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
24552 * gcov-dump.c (dump_working_sets): New function.
24554 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
24556 * hwint.c (sext_hwi, zext_hwi): New functions.
24557 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
24558 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
24559 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
24560 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
24561 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
24562 (sext_hwi, zext_hwi): New functions.
24564 2013-04-03 Jeff Law <law@redhat.com>
24566 PR tree-optimization/56799
24567 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
24568 back test for widening conversion erroneously dropped in prior change.
24570 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24573 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
24574 instead of next_real_insn.
24576 2013-04-03 Marek Polacek <polacek@redhat.com>
24579 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
24581 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24584 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
24586 (thumb1_output_casesi): Likewise.
24587 (thumb2_output_casesi): Likewise.
24589 2013-04-03 Richard Biener <rguenther@suse.de>
24591 PR tree-optimization/56817
24592 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
24594 (tree_unroll_loops_completely_1): ... new function to manually
24595 walk the loop tree, properly defering outer loops of unrolled
24596 loops to later iterations.
24598 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24600 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
24601 (vectorizable_load): Likewise.
24602 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
24603 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
24605 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24607 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
24610 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
24612 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
24614 2013-04-03 Bin Cheng <bin.cheng@arm.com>
24616 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
24618 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24620 PR tree-optimization/56790
24621 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
24624 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24626 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
24628 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
24629 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
24632 2013-04-03 Jakub Jelinek <jakub@redhat.com>
24635 * tree.h (force_folding_builtin_constant_p): New decl.
24636 * builtins.c (force_folding_builtin_constant_p): New variable.
24637 (fold_builtin_constant_p): Fold immediately also if
24638 force_folding_builtin_constant_p.
24640 2013-04-03 Richard Biener <rguenther@suse.de>
24642 PR tree-optimization/56812
24643 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
24644 DRs of the same interleaving chain are independent.
24646 2013-04-02 Jason Merrill <jason@redhat.com>
24648 * gdbinit.in (pbb): Use debug fn.
24650 2013-04-02 Lawrence Crowl <crowl@google.com>
24652 * sese.h (struct ivtype_map_elt_s): Remove unused.
24653 (extern debug_ivtype_map): Remove unused.
24654 (extern eq_ivtype_map_elts): Remove unused.
24655 * sese.c (debug_ivtype_map): Removed unused.
24656 (debug_ivtype_map_1): Removed unused.
24657 (debug_ivtype_elt): Remove unused.
24658 (eq_ivtype_map_elts): Remove unused.
24660 2013-04-02 Kai Tietz <ktietz@redhat.com>
24663 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
24664 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
24665 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
24667 (legitimize_pe_coff_symbol): Likewise.
24668 (is_imported_p): New helper-function.
24669 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
24670 for Windows x64 targets.
24671 (ix86_expand_prologue): Optimize for pe-coff targets.
24672 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
24673 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
24674 medium/large code-model.
24675 (legitimize_pic_address): Likewise.
24676 (legitimize_tls_address): Likewise.
24677 (ix86_expand_call): Likewise.
24678 (x86_output_mi_thunk): Likewise.
24679 (get_dllimport_decl): Add new beimport argument.
24680 (construct_plt_address): Don't assert for x64 pe-coff targets.
24681 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
24683 (SYMBOL_FLAG_STUBVAR): New macro.
24684 (SYMBOL_REF_STUBVAR_P): Likewise.
24685 * config/i386/winnt.c (stub_list): New structure.
24686 (stub_head): New local variable.
24687 (i386_pe_record_stub): New function.
24688 (i386_pe_file_end): Emit refptr-stubs.
24690 2013-04-02 Jakub Jelinek <jakub@redhat.com>
24692 PR rtl-optimization/56745
24693 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
24694 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
24697 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
24698 and both of them are MEM_REFs, just compare first argument for
24699 equality and attempt to deal even with differing offsets.
24702 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
24703 of gimple_clobber_p to be MEM_REF.
24704 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
24705 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
24706 after gimplification.
24707 * asan.c (get_mem_ref_of_assignment): Don't instrument
24708 gimple_clobber_p stmts.
24709 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
24710 gimple_clobber_p stmt if they have MEM_REF lhs and
24711 are dead because of another gimple_clobber_p stmt.
24712 * tree-ssa-live.c (clear_unused_block_pointer): Treat
24713 gimple_clobber_p stmts like debug stmts.
24714 (remove_unused_locals): Remove clobbers with MEM_REF lhs
24715 that refer to unused VAR_DECLs or uninitialized values.
24716 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
24717 gimple_clobber_p stmts if they refer to removed parameters.
24718 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
24721 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
24723 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
24724 using SWI48 mode attribute.
24726 2013-04-02 Wei Mi <wmi@google.com>
24728 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
24729 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
24730 *<rotate_insn><mode>3_mask in i386.md.
24732 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
24734 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
24736 2013-04-02 Richard Biener <rguenther@suse.de>
24738 PR tree-optimization/56778
24739 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
24740 Runtime alias tests are not supported for gather loads.
24741 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
24742 stmts referenced from SSA operands before updating SSA form.
24744 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
24745 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24747 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
24748 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
24749 * config/arm/cortex-a53.md: New file.
24750 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
24751 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
24752 * config/arm/arm.c (arm_issue_rate): Likewise.
24753 * config/arm/arm-tune.md: Regenerate
24754 * config/arm/arm-tables.opt: Regenerate.
24755 * config/arm/arm-cores.def: Add cortex-a53.
24757 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
24759 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
24762 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24764 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
24765 scalar load/store operations using B/H registers.
24766 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
24768 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24770 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
24772 * config/aarch64/aarch64.c
24773 (aarch64_simd_scalar_immediate_valid_for_move): New.
24774 * config/aarch64/aarch64-protos.h
24775 (aarch64_simd_scalar_immediate_valid_for_move): New.
24776 * config/aarch64/constraints.md (Dh, Dq): New.
24777 * config/aarch64/iterators.md (hq): New.
24779 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
24781 * reorg.c (get_branch_condition): Deal with conditional returns.
24782 (fill_simple_delay_slots): Remove dead code dealing with jumps.
24784 2013-04-01 Wei Mi <wmi@google.com>
24786 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
24787 Truncate operand 2 using %b asm operand modifier.
24788 (*<shift_insn><mode>3_mask): Ditto.
24789 (*<rotate_insn><mode>3_mask): Ditto.
24791 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
24793 PR middle-end/56798
24794 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
24796 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
24798 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
24800 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
24802 2013-03-30 Lawrence Crowl <crowl@google.com>
24804 * dse.c (clear_alias_sets): Remove never set.
24805 (disqualified_clear_alias_sets): Remove never set.
24806 (clear_alias_mode_pool): Remove never set.
24807 (dse_step0): Remove condition that is never true.
24808 (canon_address): Remove condition that is never true.
24809 (dse_step7): Remove condition that is never true.
24810 (rest_of_handle_dse): Remove condition that is never true.
24811 (rest_of_handle_dse::did_global): Remove never read from above.
24812 (dse_step2_spill): Remove never called from above.
24813 (dse_step5_spill): Remove never called from above.
24815 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
24817 * doc/md.texi (Standard Names) <casesi>: Update documentation for
24818 JUMP_TABLE_DATA changes.
24819 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
24820 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
24821 (Insns) <jump_table_data>: New entry.
24822 * doc/tm.texi: Regenerate.
24824 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
24826 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
24827 for table jump at the end of a basic block using tablejump_p.
24828 * targhooks.c (default_invalid_within_doloop): Likewise.
24829 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
24830 target hook implementation that is identical to the default hook.
24831 (rs6000_invalid_within_doloop): Remove.
24833 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
24834 unused variable from tablejump_p call.
24836 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
24837 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
24838 (INSN_DELETED_P): Likewise.
24839 (emit_jump_table_data): New prototype.
24840 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
24841 after 4th as unused.
24842 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
24843 * sched-vis.c (print_insn): Likewise.
24844 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
24845 insn for compatibility with back ends that use next_active_insn to
24846 identify jump table data.
24847 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
24848 (remove_insn): Likewise.
24849 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
24851 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
24852 (emit_jump_table_data): New function.
24854 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24855 basic block, a JUMP_TABLE_DATA never is.
24856 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24857 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24858 off from code handling real insns.
24859 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24860 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24861 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24862 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24863 is not a NONDEBUG_INSN_P.
24864 * ira-costs.c (scan_one_insn): Likewise.
24865 * jump.c (mark_all_labels): Likewise.
24866 (mark_jump_label_1): Likewise.
24867 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24868 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24870 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24871 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24872 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24873 JUMP_TABLE_DATA_P insns.
24874 (calculate_elim_costs_all_insns): Likewise.
24875 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24876 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24877 (delete_output_reload): Code style fixups.
24878 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24879 insn flags on this non-insn.
24880 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24881 as scheduling barriers, for pre-change compatibility.
24882 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24883 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24885 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24886 redundant JUMP_TABLE_DATA_P test.
24887 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24888 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24889 (frv_for_each_packet): Likewise.
24890 * config/i386/i386.c (min_insn_size): Likewise.
24891 (ix86_avoid_jump_mispredicts): Likewise.
24892 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24893 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24894 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24895 (mips16_insn_length): Robustify.
24896 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24897 (mips16_split_long_branches): Likewise.
24898 * config/pa/pa.c (pa_combine_instructions): Likewise.
24899 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24900 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24901 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24902 as contributing to pool range lengths.
24903 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24904 Remove redundant JUMP_TABLE_DATA_P test.
24905 (sh_loop_align): Likewise.
24906 (split_branches): Likewise.
24907 (sh_insn_length_adjustment): Likewise.
24908 * config/spu/spu.c (get_branch_target): Likewise.
24910 2013-03-29 Jan Hubicka <jh@suse.cz>
24912 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24913 gcov streaming; stream hot bb threshold to ltrans.
24914 * predict.c (get_hot_bb_threshold): Break out from ....
24915 (maybe_hot_count_p): ... here.
24916 (set_hot_bb_threshold): New function.
24917 * lto-section-in.c (lto_section_name): Add profile.
24918 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24919 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24920 and data-streamer.h
24921 (histogram_entry): New structure.
24922 (histogram, histogram_pool): New global vars.
24923 (histogram_hash): New structure.
24924 (histogram_hash::hash): New method.
24925 (histogram_hash::equal): Likewise.
24926 (account_time_size): New function.
24927 (cmp_counts): New function.
24928 (dump_histogram): New function.
24929 (ipa_profile_generate_summary): New function.
24930 (ipa_profile_write_summary): New function.
24931 (ipa_profile_read_summary): New function.
24932 (ipa_profile): Decide on threshold.
24933 (pass_ipa_profile): Add ipa_profile_write_summary and
24934 ipa_profile_read_summary.
24935 * Makefile.in (ipa.o): Update dependencies.
24936 * lto-streamer.h (LTO_section_ipa_profile): New section.
24938 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24940 * tree.h (VAR_P): New.
24942 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24945 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24947 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24949 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24950 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24951 (control_flow_insn_p): Likewise.
24952 * cfgrtl.c (duplicate_insn_chain): Likewise.
24953 * final.c (get_attr_length_1): Likewise.
24954 (shorten_branches): Likewise.
24955 (final_scan_insn): Likewise.
24956 * function.c (instantiate_virtual_regs): Likewise.
24957 * gcse.c (insert_insn_end_basic_block): Likewise.
24958 * ira-costs.c (scan_one_insn): Likewise.
24959 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24960 * lra.c (check_rtl): Likewise.
24961 * reload1.c (elimination_costs_in_insn): Likewise.
24962 * reorg.c (follow_jumps): Likewise.
24964 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24965 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24966 (thumb_far_jump_used_p): Likewise.
24967 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24968 (workaround_speculation): Likewise.
24969 (add_sched_insns_for_speculation): Likewise.
24970 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24971 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24972 (frv_for_each_packet): Likewise.
24973 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24974 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24975 (final_emit_insn_group_barriers): Likewise.
24976 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24977 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24978 (mips16_insn_length): Likewise.
24979 * config/pa/pa.c (pa_reorg): Likewise.
24980 (pa_combine_instructions): Likewise.
24981 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24982 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24983 (sh_reorg): Likewise.
24984 (split_branches): Likewise.
24985 * config/spu/spu.c (get_branch_target): Likewise.
24987 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24990 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24992 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24993 Fix declaration name.
24995 2013-03-28 Lawrence Crowl <crowl@google.com>
24997 * graphds.h (struct graph.indicies): Remove unused.
24998 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24999 (SCOP_ORIGINAL_PDDRS): Remove unused.
25000 * sese.h (extern insert_loop_close_phis): Removed unused.
25001 (extern insert_guard_phis): Removed unused.
25002 (extern ivtype_map_elt_info): Removed unused.
25003 (new_ivtype_map_elt): Removed unused.
25004 * sese.c (ivtype_map_elt_info): Removed unused.
25006 2013-03-28 Lawrence Crowl <crowl@google.com>
25008 * Makefile.in: Add several missing include dependences.
25010 (test-dump.o): New. This object is not added to any executable,
25011 but is present for ad-hoc testing.
25013 (debug (const bitmap_head_def &)): New.
25014 (debug (const bitmap_head_def *)): New.
25016 (extern debug (const bitmap_head_def &)): New.
25017 (extern debug (const bitmap_head_def *)): New.
25019 (debug (edge_def &)): New.
25020 (debug (edge_def *)): New.
25022 (debug (basic_block_def &)): New.
25023 (debug (basic_block_def *)): New.
25025 (dump_node (const_tree, int, FILE *)): Correct source file.
25027 (debug (die_struct &)): New.
25028 (debug (die_struct *)): New.
25030 (extern debug (die_struct &)): New.
25031 (extern debug (die_struct *)): New.
25032 * gimple-pretty-print.c
25033 (debug (gimple_statement_d &)): New.
25034 (debug (gimple_statement_d *)): New.
25035 * gimple-pretty-print.h
25036 (extern debug (gimple_statement_d &)): New.
25037 (extern debug (gimple_statement_d *)): New.
25039 (debug (ira_allocno_copy &)): New.
25040 (debug (ira_allocno_copy *)): New.
25041 (debug (ira_allocno &)): New.
25042 (debug (ira_allocno *)): New.
25044 (extern debug (ira_allocno_copy &)): New.
25045 (extern debug (ira_allocno_copy *)): New.
25046 (extern debug (ira_allocno &)): New.
25047 (extern debug (ira_allocno *)): New.
25049 (debug (live_range &)): New.
25050 (debug (live_range *)): New.
25052 (debug (lra_live_range &)): New.
25053 (debug (lra_live_range *)): New.
25055 (debug (lra_live_range &)): New.
25056 (debug (lra_live_range *)): New.
25058 (debug (omega_pb_d &)): New.
25059 (debug (omega_pb_d *)): New.
25061 (extern debug (omega_pb_d &)): New.
25062 (extern debug (omega_pb_d *)): New.
25064 (debug (const rtx_def &)): New.
25065 (debug (const rtx_def *)): New.
25067 (debug_tree (tree): Move within file.
25068 (debug_raw (const tree_node &)): New.
25069 (debug_raw (const tree_node *)): New.
25070 (dump_tree_via_hooks (const tree_node *, int)): New.
25071 (debug (const tree_node &)): New.
25072 (debug (const tree_node *)): New.
25073 (debug_verbose (const tree_node &)): New.
25074 (debug_verbose (const tree_node *)): New.
25075 (debug_head (const tree_node &)): New.
25076 (debug_head (const tree_node *)): New.
25077 (debug_body (const tree_node &)): New.
25078 (debug_body (const tree_node *)): New.
25079 (debug_vec_tree (tree): Move and reimplement in terms of dump.
25080 (debug (vec<tree, va_gc> &)): New.
25081 (debug (vec<tree, va_gc> *)): New.
25083 (extern debug (const rtx_def &)): New.
25084 (extern debug (const rtx_def *)): New.
25086 (debug_raw (simple_bitmap_def &)): New.
25087 (debug_raw (simple_bitmap_def *)): New.
25088 (debug (simple_bitmap_def &)): New.
25089 (debug (simple_bitmap_def *)): New.
25091 (extern debug (simple_bitmap_def &)): New.
25092 (extern debug (simple_bitmap_def *)): New.
25093 (extern debug_raw (simple_bitmap_def &)): New.
25094 (extern debug_raw (simple_bitmap_def *)): New.
25096 (debug (vinsn_def &)): New.
25097 (debug (vinsn_def *)): New.
25098 (debug_verbose (vinsn_def &)): New.
25099 (debug_verbose (vinsn_def *)): New.
25100 (debug (expr_def &)): New.
25101 (debug (expr_def *)): New.
25102 (debug_verbose (expr_def &)): New.
25103 (debug_verbose (expr_def *)): New.
25104 (debug (vec<rtx> &)): New.
25105 (debug (vec<rtx> *)): New.
25107 (extern debug (vinsn_def &)): New.
25108 (extern debug (vinsn_def *)): New.
25109 (extern debug_verbose (vinsn_def &)): New.
25110 (extern debug_verbose (vinsn_def *)): New.
25111 (extern debug (expr_def &)): New.
25112 (extern debug (expr_def *)): New.
25113 (extern debug_verbose (expr_def &)): New.
25114 (extern debug_verbose (expr_def *)): New.
25115 (extern debug (vec<rtx> &)): New.
25116 (extern debug (vec<rtx> *)): New.
25118 (_list_iter_cond_expr): Make inline instead of static.
25120 (debug (sreal &)): New.
25121 (debug (sreal *)): New.
25123 (extern debug (sreal &)): New.
25124 (extern debug (sreal *)): New.
25126 (extern debug_raw (const tree_node &)): New.
25127 (extern debug_raw (const tree_node *)): New.
25128 (extern debug (const tree_node &)): New.
25129 (extern debug (const tree_node *)): New.
25130 (extern debug_verbose (const tree_node &)): New.
25131 (extern debug_verbose (const tree_node *)): New.
25132 (extern debug_head (const tree_node &)): New.
25133 (extern debug_head (const tree_node *)): New.
25134 (extern debug_body (const tree_node &)): New.
25135 (extern debug_body (const tree_node *)): New.
25136 (extern debug (vec<tree, va_gc> &)): New.
25137 (extern debug (vec<tree, va_gc> *)): New.
25139 (debug (struct loop &)): New.
25140 (debug (struct loop *)): New.
25141 (debug_verbose (struct loop &)): New.
25142 (debug_verbose (struct loop *)): New.
25143 * tree-dump.c: Add header dependence.
25145 (extern debug (struct loop &)): New.
25146 (extern debug (struct loop *)): New.
25147 (extern debug_verbose (struct loop &)): New.
25148 (extern debug_verbose (struct loop *)): New.
25150 (debug (data_reference &)): New.
25151 (debug (data_reference *)): New.
25152 (debug (vec<data_reference_p> &)): New.
25153 (debug (vec<data_reference_p> *)): New.
25154 (debug (vec<ddr_p> &)): New.
25155 (debug (vec<ddr_p> *)): New.
25157 (extern debug (data_reference &)): New.
25158 (extern debug (data_reference *)): New.
25159 (extern debug (vec<data_reference_p> &)): New.
25160 (extern debug (vec<data_reference_p> *)): New.
25161 (extern debug (vec<ddr_p> &)): New.
25162 (extern debug (vec<ddr_p> *)): New.
25164 (debug (pt_solution &)): New.
25165 (debug (pt_solution *)): New.
25167 (extern debug (pt_solution &)): New.
25168 (extern debug (pt_solution *)): New.
25170 (debug (_var_map &)): New.
25171 (debug (_var_map *)): New.
25172 (debug (tree_live_info_d &)): New.
25173 (debug (tree_live_info_d *)): New.
25175 (extern debug (_var_map &)): New.
25176 (extern debug (_var_map *)): New.
25177 (extern debug (tree_live_info_d &)): New.
25178 (extern debug (tree_live_info_d *)): New.
25180 2013-03-28 Jan Hubicka <jh@suse.cz>
25182 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
25184 2013-03-28 Ian Bolton <ian.bolton@arm.com>
25186 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
25187 record only when desired or required.
25189 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
25191 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
25192 *vec_extractv2di_1_rex64. Use x64 isa attribute.
25194 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25196 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
25197 (*andsi3_compare0_uxtw): New pattern.
25198 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
25199 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
25201 2013-03-28 Jan Hubicka <jh@suse.cz>
25203 * data-streamer-in.c (streamer_read_gcov_count): New function.
25204 * gimple-streamer-out.c: Include value-prof.h.
25205 (output_gimple_stmt): Output histogram.
25206 (output_bb): Use streamer_write_gcov_count.
25207 * value-prof.c: Include data-streamer.h
25208 (dump_histogram_value): Add HIST_TYPE_MAX.
25209 (stream_out_histogram_value): New function.
25210 (stream_in_histogram_value): New function.
25211 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
25212 (stream_out_histogram_value, stream_in_histogram_value): Declare.
25213 * data-streamer-out.c (streamer_write_gcov_count): New function.
25214 (streamer_write_gcov_count_stream): New function.
25215 * lto-cgraph.c (lto_output_edge): Update counter streaming.
25216 (lto_output_node): Likewise.
25217 (input_node, input_edge): Likewise.
25218 * lto-streamer-out.c (output_cfg): Update streaming.
25219 * lto-streamer-in.c (input_cfg): Likewise.
25220 * data-streamer.h (streamer_write_gcov_count,
25221 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
25222 * gimple-streamer-in.c: Include value-prof.h
25223 (input_gimple_stmt): Input histograms.
25224 (input_bb): Update profile streaming.
25226 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
25228 * genmodes.c (emit_max_int): New function.
25229 (emit_insn_modes_h): Added call to emit_max_function.
25230 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
25232 * machmode.def: Fixed comment.
25234 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
25236 * combine.c (try_combine): Removed useless assert.
25237 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
25239 2013-03-28 Marek Polacek <polacek@redhat.com>
25240 Richard Biener <rguenther@suse.de>
25242 PR tree-optimization/56695
25243 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
25244 build signed result of a vector comparison.
25245 * tree-cfg.c (verify_gimple_comparison): Check that a result
25246 of a vector comparison has signed type.
25248 2013-03-28 Richard Biener <rguenther@suse.de>
25250 PR tree-optimization/37021
25251 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
25252 do not restrict gaps between groups.
25253 * tree-vect-stmts.c (vectorizable_load): Properly account for
25254 a gap between groups.
25256 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
25258 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
25259 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
25262 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
25264 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
25265 * doc/extend.texi (Named Address Spaces): Ditto.
25266 (Variable Attributes): Ditto.
25268 2013-03-27 Kai Tietz <ktietz@redhat.com>
25270 * config.build: Add support for cygwin x64 target.
25271 * config.gcc: Likewise.
25272 * config.host: Likewise.
25273 * configure.ac: Likewise
25274 * configure: Regenerated.
25276 2013-03-27 Kai Tietz <ktietz@redhat.com>
25278 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
25279 * config/i386/t-cygwin-w64: New file.
25280 * config/i386/cygwin-w64.h: New file.
25281 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
25282 and add support for x64-cygwin target.
25283 (CPP_SPEC): Likewise.
25284 (CXX_WRAP_SPEC_LIST): Undefine before define.
25285 (LIBGCJ_SONAME): Use 15 as version.
25287 2013-03-27 Richard Biener <rguenther@suse.de>
25289 PR tree-optimization/56716
25290 * tree-ssa-structalias.c (perform_var_substitution): Adjust
25291 dumping for ref nodes.
25293 2013-03-27 Martin Jambor <mjambor@suse.cz>
25295 PR tree-optimization/55334
25296 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
25297 restricted pointers to arrays.
25299 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
25301 * Makefile.in (.SUFFIXES): Add .cc.
25302 (.c.o): Apply same recipe for implicit rule .cc.o.
25304 2013-03-27 Richard Biener <rguenther@suse.de>
25306 PR tree-optimization/37021
25307 * tree-vect-data-refs.c (vect_check_strided_load): Allow
25308 REALPART/IMAGPART_EXPRs around the supported refs.
25309 * tree-ssa-structalias.c (find_func_aliases): Assume that
25310 floating-point values are not used to transfer pointers.
25312 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
25314 * target.def (TARGET_HAS_IFUNC_P): New target hook.
25315 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
25316 * doc/tm.texi: Regenerate.
25317 * targhooks.h (default_has_ifunc_p): New.
25318 * targhooks.c (default_has_ifunc_p): Ditto.
25319 * config/linux-protos.h: New file.
25320 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
25321 hook for linux which disables support of indirect functions in android.
25322 * config/linux-android.c: New file.
25323 * config/t-linux-android.c: Ditto.
25324 * config.gcc: Added new object file linux-android.o.
25325 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
25326 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
25327 * varasm.c (do_assemble_alias): Likewise.
25328 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
25329 doesn't support indirect functions.
25330 * configure: Regenerate.
25332 2013-03-27 Bin Cheng <bin.cheng@arm.com>
25335 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
25336 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
25339 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25341 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
25343 2013-03-27 Terry Guo <terry.guo@arm.com>
25345 * config/arm/arm-cores.def: Added core cortex-r7.
25346 * config/arm/arm-tune.md: Regenerated.
25347 * config/arm/arm-tables.opt: Regenerated.
25348 * doc/invoke.texi: Added entry for core cortex-r7.
25350 2013-03-27 Walter Lee <walt@tilera.com>
25352 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
25353 double-decrement of next_scratch_regno.
25355 2013-03-27 Walter Lee <walt@tilera.com>
25357 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
25359 (insn_v1mulus): Ditto.
25360 (insn_v2muls): Ditto.
25362 2013-03-27 Walter Lee <walt@tilera.com>
25364 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
25365 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
25367 2013-03-27 Walter Lee <walt@tilera.com>
25369 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
25370 (*sibcall_value): Ditto.
25372 2013-03-27 Walter Lee <walt@tilera.com>
25374 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
25375 (insn_mnz_v8qi): ... this ...
25376 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
25378 (insn_v<n>mnz): Replaced by ...
25379 (insn_v1mnz): ... this ...
25380 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
25382 (insn_mz_<mode>): Replaced by ...
25383 (insn_mz_v8qi): ... this ...
25384 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
25386 (insn_v<n>mz): Replaced by ...
25387 (insn_v1mz): ... this ...
25388 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
25391 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
25393 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
25395 2013-03-26 Roland McGrath <mcgrathr@google.com>
25397 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
25398 than fprintf with a non-constant, non-format string.
25400 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
25402 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
25403 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
25404 operand 0 predicate.
25405 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
25406 attribute. Use general_x64nomem_operand as operand 1 predicate.
25407 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
25408 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
25409 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
25410 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
25411 (mov<mode>_insv_1): Remove expander. Merge insn with
25412 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
25413 Use general_x64nomem_operand as operand 1 predicate.
25414 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
25415 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
25416 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
25417 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
25418 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
25419 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
25420 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
25421 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
25422 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
25423 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
25424 (general_x64nomem_operand): Ditto.
25426 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
25428 * config/rtems.opt: Add -pthread option.
25430 2013-03-26 Richard Biener <rguenther@suse.de>
25432 * alias.c (find_base_term): Avoid redundant and not used recursion.
25433 (base_alias_check): Get the initial base term from the caller.
25434 (true_dependence_1): Compute and pass base terms to base_alias_check.
25435 (write_dependence_p): Likewise.
25436 (may_alias_p): Likewise.
25438 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
25440 * config/aarch64/aarch64.c (aarch64_classify_address): Support
25441 PC-relative load in SI modes and above only.
25443 2013-03-26 Xinyu Qi <xyqi@marvell.com>
25445 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
25446 * config/arm/iwmmxt.md (WCGR0): Update.
25447 (WCGR1, WCGR2, WCGR3): Likewise.
25449 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
25451 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
25452 Use x64 and nox64 isa attributes.
25454 2013-03-26 Richard Biener <rguenther@suse.de>
25456 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
25457 alignment computations and rely on get_object_alignment_1
25458 for the !TYPE_P case.
25459 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
25461 2013-03-26 Walter Lee <walt@tilera.com>
25463 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
25464 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
25466 2013-03-25 Jeff Law <law@redhat.com>
25468 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
25469 check for INTEGRAL_TYPE_P that was missing due to checking in
25470 wrong version of prior patch.
25472 2013-03-25 Walter Lee <walt@tilera.com>
25474 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
25475 TILEGX_INSN_SHUFFLEBYTES1.
25476 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
25478 (tilegx_builtins): Ditto.
25479 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
25481 2013-03-25 Walter Lee <walt@tilera.com>
25483 * config/tilegx/tilegx.md (floatsisf2): New pattern.
25484 (floatunssisf2): New pattern.
25485 (floatsidf2): New pattern.
25486 (floatunssidf2): New pattern.
25488 2013-03-25 Walter Lee <walt@tilera.com>
25490 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
25491 tests for constraint J, K, N, P.
25493 2013-03-25 Walter Lee <walt@tilera.com>
25495 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
25496 Use indirect/pcrel encoding.
25497 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
25500 2013-03-25 Steve Ellcey <sellcey@mips.com>
25502 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
25503 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
25504 * config/mips/mips.c (mips_option_override): Set IMADD default.
25505 * config/mips/mips.h (PTF_AVOID_IMADD): New.
25506 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
25507 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
25508 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
25509 * doc/invoke.texi (-mimadd/-mno-imadd): New.
25511 2013-03-25 Jeff Law <law@redhat.com>
25513 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
25514 slightly to avoid creating and folding useless trees. Simplify
25515 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
25517 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
25519 * config/i386/i386.md (*zero_extendsidi2): Merge with
25520 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
25521 * config/i386/predicates.md (x86_64_zext_operand): Rename from
25522 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
25523 targets. Clarify comment.
25525 2013-03-25 Martin Jambor <mjambor@suse.cz>
25527 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
25528 pass-through jump functions differently.
25529 (ipa_read_jump_function): Likewise. Also use setter functions to set
25532 2013-03-25 Martin Jambor <mjambor@suse.cz>
25534 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
25535 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
25537 (ipa_get_indirect_edge_target): New function.
25538 (devirtualization_time_bonus): New parameter known_aggs, pass it to
25539 ipa_get_indirect_edge_target. Update all callers.
25540 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
25541 ipa_get_indirect_edge_target_1 instead of calling
25542 ipa_get_indirect_edge_target.
25543 (create_specialized_node): Pass aggvlas to
25544 ipcp_discover_new_direct_edges.
25546 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25548 * config/arm/arm.md (f_sels, f_seld): New types.
25549 (*cmov<mode>): New pattern.
25550 * config/arm/predicates.md (arm_vsel_comparison_operator): New
25553 2013-03-25 Kai Tietz <ktietz@redhat.com>
25555 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
25556 POSIX-printf for mingw-hosted builds.
25558 2013-03-25 Richard Biener <rguenther@suse.de>
25560 PR middle-end/56694
25561 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
25562 must-not-throw stmt location.
25564 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25566 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
25567 Emit load-acquire versions when acq is true.
25568 (arm_emit_store_exclusive): Add rel parameter.
25569 Emit store-release versions when rel is true.
25570 (arm_split_compare_and_swap): Use acquire-release instructions
25572 of barriers when appropriate.
25573 (arm_split_atomic_op): Likewise.
25574 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
25575 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
25576 (VUNSPEC_SLX): Likewise.
25577 (VUNSPEC_LDA): Likewise.
25578 (VUNSPEC_STL): Likewise.
25579 * config/arm/sync.md (atomic_load<mode>): New pattern.
25580 (atomic_store<mode>): Likewise.
25581 (arm_load_acquire_exclusive<mode>): Likewise.
25582 (arm_load_acquire_exclusivesi): Likewise.
25583 (arm_load_acquire_exclusivedi): Likewise.
25584 (arm_store_release_exclusive<mode>): Likewise.
25586 2013-03-25 Catherine Moore <clm@codesourcery.com>
25588 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
25589 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
25590 * config/mip/predicates.md (lwsp_swsp_operand,
25591 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
25592 sb16_operand, db4_operand, db7_operand, ib3_operand,
25593 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
25594 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
25595 andi16_operand): New predicates.
25596 * config/mips/mips.md (compression): New attribute.
25597 (enabled): New attribute.
25598 (length): Consider compression in computing length.
25599 (shift_compression): New code attribute.
25600 (*add<mode>3): New operands. Record compression.
25601 (sub<mode>3): Likewise.
25602 (one_cmpl<mode>2): Likewise.
25603 (*and<mode>3): Likewise.
25604 (*ior<mode>3): Likewise.
25605 (unnamed pattern for xor): Likewise.
25606 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
25607 (*<optab><mode>3): Likewise.
25608 (*mov<mode>_internal: Likewise.
25609 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
25610 (mips_unsigned_immediate_p): New.
25611 (umips_lwsp_swsp_address_p): New.
25612 (m16_based_address_p): New.
25613 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
25614 (mips_unsigned_immediate_p): New prototype.
25615 (lwsp_swsp_address_p): New prototype.
25616 (m16_based_address_p): New prototype.
25617 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
25618 (mips_signed_immediate_p): New function.
25619 (m16_based_address_p): New function.
25620 (lwsp_swsp_address_p): New function.
25621 (mips_print_operand_punctuation): Recognize short delay slot insns
25622 for microMIPS.add<mode>3"
25624 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25627 * config/arm/iterators.md (v_cmp_result): New mode attribute.
25628 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
25630 2013-03-25 Richard Biener <rguenther@suse.de>
25632 PR tree-optimization/56689
25633 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
25636 2013-03-25 Richard Biener <rguenther@suse.de>
25638 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
25640 (memory_references): Likewise.
25641 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
25642 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
25643 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
25644 (gather_mem_refs_in_loops): Fold into ...
25645 (analyze_memory_references): ... this. Move initialization
25646 to tree_ssa_lim_initialize.
25647 (fill_always_executed_in): Rename to ...
25648 (fill_always_executed_in_1): ... this.
25649 (fill_always_executed_in): Move contains_call computation to
25650 this new function from ...
25651 (tree_ssa_lim_initialize): ... here.
25652 (tree_ssa_lim): Call fill_always_executed_in.
25654 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25656 * postreload.c (reload_combine): Fix code detecting returns.
25658 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25660 * function.c (emit_use_return_register_into_block): On cc0 targets,
25661 do not emit the sequence between cc0 setter and user.
25663 2013-03-25 Kai Tietz <ktietz@redhat.com>
25665 * config/i386/predicates.md (local_symbolic_operand): Interpret
25666 dll-imported symbols as none-local.
25668 2013-03-25 Richard Biener <rguenther@suse.de>
25670 * tree-ssa-loop-im.c (struct depend): Remove.
25671 (struct lim_aux_data): Make depends a vec of gimples.
25672 (free_lim_aux_data): Adjust.
25673 (add_dependency): Likewise.
25674 (set_level): Likewise.
25676 2013-03-25 Richard Biener <rguenther@suse.de>
25678 PR middle-end/56434
25679 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
25680 the pointer returned by calls with ECF_MALLOC set.
25682 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25684 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
25686 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25688 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
25689 using MMXMODE mode iterator.
25690 (*move<mode>_internal): Merge with *movv2sf_internal and
25691 *movv2sf_internal_rex64 using MMXMODE mode iterator.
25693 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25695 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
25696 (record_last_mem_set_info): Likewise.
25698 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
25699 of XNEWVEC followed by memset.
25700 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
25702 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25704 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
25705 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
25706 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
25707 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
25708 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
25709 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
25710 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
25711 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
25712 BARRIER_P instead of GET_CODE.
25714 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25716 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
25717 inaccuracy in the probing code.
25719 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
25720 (ctrapdi4): Likewise.
25722 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25724 * calls.c (expand_call): Add missing guard to code handling return
25725 of non-BLKmode structures in MSB.
25726 * function.c (expand_function_end): Likewise.
25728 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25730 * combine.c (try_combine): Adjust comment. Do not add the set of
25731 insn #0 if the destination indirectly is set or dies in insn #2.
25732 Tidy up code to distribute a new note.
25734 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25736 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
25737 also for alternatives 16 and 17.
25739 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25741 * config/i386/sse.md (*mov<mode>_internal): Merge with
25742 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
25743 Emit insn template depending on type attribute. Use
25744 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25745 movd instead of movq mnemonic for interunit moves. Rewrite mode
25746 attribute calculation. Remove unit attribute calculation.
25747 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
25748 Set prefix_data16 attribute for DImode ssemov types.
25749 Use Ym instead of y for SSE-MMX conversion alternatives.
25750 Reorder operand constraints.
25752 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
25754 * df.h (df_insn_delete): Adjust prototype.
25755 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
25756 and let it decide whether mark the basic block dirty.
25757 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
25758 * df-scan.c (df_insn_info_delete): New helper function, split
25759 off from df_insn_delete.
25760 (df_scan_free_bb_info): Use it.
25761 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
25763 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
25764 that the insn is actually an insn and it has a non-NULL basic block.
25765 Do not mark basic block dirty if only deleting a DEBUG_INSN.
25767 2013-03-22 Richard Biener <rguenther@suse.de>
25769 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
25771 (mem_ref_alloc): Do not allocate them.
25772 (refs_independent_p): Do not query or maintain a cache.
25774 2013-03-22 Richard Biener <rguenther@suse.de>
25776 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
25777 (gather_mem_refs_in_loops): Do not compute it.
25778 (analyze_memory_references): Do not allocate it.
25779 (tree_ssa_lim_finalize): Do not free it.
25780 (for_all_locs_in_loop): Do not query all_refs_in_loop.
25782 2013-03-22 Richard Biener <rguenther@suse.de>
25784 * is-a.h (as_a): Use gcc_checking_assert.
25786 2013-03-22 Ian Bolton <ian.bolton@arm.com>
25788 * config/aarch64/aarch64.c (aarch64_print_operand): New
25789 format specifier for printing a constant in hex.
25790 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
25791 format specifier for printing second operand.
25793 2013-03-22 Richard Biener <rguenther@suse.de>
25795 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
25797 (gather_mem_refs_in_loops): Perform store accumulation here.
25798 (create_vop_ref_mapping_loop): Remove.
25799 (create_vop_ref_mapping): Likewise.
25800 (analyze_memory_references): Initialize refs_stored_in_loop.
25801 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
25802 (record_indep_loop): Remove.
25803 (record_dep_loop): New function.
25804 (ref_indep_loop_p_1): Adjust to only walk over references
25805 in the loop, not its subloops.
25806 (ref_indep_loop_p): Rename to ...
25807 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
25808 maintaining a more fine-grained cache.
25809 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
25810 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
25812 2013-03-22 Richard Biener <rguenther@suse.de>
25814 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
25815 (struct mem_ref): Make accesses_in_loop a vec of a vec of
25816 aggregate mem_ref_loc.
25817 (free_mem_ref_locs): Inline into ...
25818 (memref_free): ... this and adjust.
25819 (mem_ref_alloc): Adjust.
25820 (mem_ref_locs_alloc): Remove.
25821 (record_mem_ref_loc): Adjust.
25822 (get_all_locs_in_loop): Rewrite into ...
25823 (for_all_locs_in_loop): ... this iterator.
25824 (rewrite_mem_ref_loc): New functor.
25825 (rewrite_mem_refs): Use for_all_locs_in_loop.
25826 (sm_set_flag_if_changed): New functor.
25827 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
25828 (ref_always_accessed): New functor.
25829 (ref_always_accessed_p): Use for_all_locs_in_loop.
25831 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25833 * tree-pass.h (PROP_gimple_lvec): New.
25834 * passes.c (dump_properties): Handle PROP_gimple_lvec.
25835 (init_optimization_passes): Move pass_lower_vector.
25836 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
25838 (pass_lower_vector): Provide PROP_gimple_lvec.
25839 (pass_lower_vector_ssa): Likewise.
25840 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
25842 2013-03-21 Mark Wielaard <mjw@redhat.com>
25844 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
25846 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25848 * config/i386/i386.md (*movdi_internal): Disparage slightly
25849 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
25850 conversion alternatives.
25852 2013-03-21 Jakub Jelinek <jakub@redhat.com>
25854 PR middle-end/48087
25855 * diagnostic.def (DK_WERROR): New kind.
25856 * diagnostic.h (werrorcount): Define.
25857 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25858 promoted to DK_ERROR, increment DK_WERROR counter instead of
25860 * toplev.c (toplev_main): Call print_ignored_options even if
25861 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25862 even if just werrorcount is non-zero.
25865 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25867 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25868 (string_cst_pool_decl): New function.
25869 (optimize_one_addr_into_implicit_ptr): New function.
25870 (resolve_addr_in_expr): Optimize DWARF location expression
25871 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25872 which doesn't live in memory, but has DW_AT_location or
25873 DW_AT_const_value, or refers to a string literal, into
25874 DW_OP_GNU_implicit_pointer.
25875 (optimize_location_into_implicit_ptr): New function.
25876 (resolve_addr): If removing DW_AT_location of a variable because
25877 it was DW_OP_addr of address of the variable, but the variable doesn't
25878 live in memory, try to emit const value attribute for the initializer.
25880 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25882 * tree.h (VECTOR_TYPE_P): New macro.
25883 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25884 TYPE_MODE): Use it.
25885 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25886 VEC_COND_EXPR cannot be lvalues.
25887 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25889 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25891 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25892 Restrict the transformation to equal modes.
25894 2013-03-21 Richard Biener <rguenther@suse.de>
25896 PR tree-optimization/39326
25897 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25898 (MEM_ANALYZABLE): Adjust.
25899 (record_mem_ref_loc): Move bitmap ops ...
25900 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25901 unanalyzable refs, do not record locations for it.
25902 (analyze_memory_references): Allocate ref zero as shared
25904 (refs_independent_p): Do not test for unanalyzed mems here.
25905 (ref_indep_loop_p_1): Special-case disambiguation against
25906 the unanalyzed ref.
25907 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25909 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25911 * config/arm/arm-protos.h (tune_params): Add
25912 prefer_neon_for_64bits field.
25913 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25914 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25915 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25916 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25917 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25918 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25919 (arm_option_override): Handle -mneon-for-64bits new option.
25920 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25921 (prefer_neon_for_64bits): Declare new variable.
25922 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25923 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25924 (arch_enabled): Handle new arch types. Remove support for onlya8
25926 (one_cmpldi2): Use new arch names.
25927 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25928 * config/arm/arm.opt (mneon-for-64bits): Add option.
25929 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25930 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25931 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25933 * doc/invoke.texi (-mneon-for-64bits): Document.
25935 2013-03-21 Richard Biener <rguenther@suse.de>
25937 PR tree-optimization/39326
25938 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25939 (sort_bbs_in_loop_postorder_cmp): New function.
25940 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25942 2013-03-21 Richard Biener <rguenther@suse.de>
25944 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25945 (vect_insert_into_interleaving_chain): Likewise.
25946 (vect_drs_dependent_in_basic_block): Inline ...
25947 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25949 (vect_analyze_data_ref_dependence): ... here. Simplify.
25950 (vect_check_interleaving): Simplify.
25951 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25952 (vect_slp_analyze_data_ref_dependences): ... this new function.
25953 (dr_group_sort_cmp): New function.
25954 (vect_analyze_data_ref_accesses): Compute data-reference groups
25955 here instead of in vect_analyze_data_ref_dependence. Use
25956 a more efficient algorithm.
25957 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25958 vect_slp_analyze_data_ref_dependences. Call
25959 vect_analyze_data_ref_accesses earlier.
25960 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25961 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25962 (vect_slp_analyze_data_ref_dependences): New prototype.
25964 2013-03-21 Richard Biener <rguenther@suse.de>
25966 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25967 ref is stored in the loop.
25968 (find_refs_for_sm): Walk only over all stores.
25969 (store_motion_loop): Allocate from lim_bitmap_obstack.
25970 (store_motion): Likewise.
25972 2013-03-21 Richard Biener <rguenther@suse.de>
25974 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25975 Update virtual SSA form.
25977 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25979 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25980 * configure: Regenerate.
25981 * config.in: Regenerate.
25982 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25983 if !HAVE_LD_EH_FRAME_CIEV3.
25985 2013-03-21 Richard Biener <rguenther@suse.de>
25987 * tree-cfg.c (verify_expr_no_block): New function.
25988 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25989 nor DECL_VALUE_EXPR have locations with associated blocks.
25990 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25991 (clear_unused_block_pointer): Remove code dealing with
25992 blocks in DECL_DEBUG_EXPR locations.
25994 2013-03-21 Richard Biener <rguenther@suse.de>
25996 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25997 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25998 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25999 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
26000 instead of DECL_DEBUG_EXPR_IS_FROM.
26001 * gimplify.c (gimplify_modify_expr): Likewise.
26002 * tree-cfg.c (verify_expr_location_1): Likewise.
26003 * tree-complex.c (create_one_component_var): Likewise.
26004 * tree-sra.c (create_access_replacement): Likewise.
26005 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
26006 (clear_unused_block_pointer): Likewise.
26007 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
26008 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
26009 * var-tracking.c (var_debug_decl): Likewise.
26010 (track_expr_p): Likewise.
26011 * tree-inline.c (add_local_variables): Likewise. Set
26012 DECL_HAS_DEBUG_EXPR_P after copying it.
26013 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
26014 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
26016 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
26019 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
26020 * configure: Regenerate.
26021 * config.in: Regenerate.
26022 * config/i386/i386.md (*movdf_internal): Use
26023 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
26024 movd instead of movq mnemonic for interunit moves.
26025 (*movdi_internal): Ditto.
26027 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26029 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
26030 (abd<mode>_3): New pattern.
26031 (aba<mode>_3): New pattern.
26032 (fabd<mode>_3): New pattern.
26034 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26036 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
26037 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
26038 occurrence of REGISTER_PREFIX as its empty string.
26040 2013-03-20 Jeff Law <law@redhat.com>
26042 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
26043 addititional equivalences for equality comparisons between an SSA_NAME
26044 and a constant where the SSA_NAME was set from a widening conversion.
26046 2013-03-20 Walter Lee <walt@tilera.com>
26048 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
26050 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26052 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
26053 depending on type attribute.
26054 (*movti_internal): Ditto.
26055 (*movtf_internal): Ditto.
26056 (*movxf_internal): Ditto.
26057 (*movdf_internal): Ditto.
26058 (*movsf_internal): Ditto.
26060 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26062 * config/i386/i386.md (*movti_internal): Set prefix attribute to
26063 maybe_vex for sselog1 and ssemov types.
26064 (*movdi_internal): Reorder operand constraints.
26065 (*movsi_internal): Ditto. Set prefix attribute to
26066 maybe_vex for sselog1 and ssemov types.
26067 (*movtf_internal): Set prefix attribute to maybe_vex
26068 for sselog1 and ssemov types.
26069 (*movdf_internal): Ditto. Set prefix_data16 attribute for
26070 DImode ssemov types. Reorder operand constraints.
26071 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
26072 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
26073 attribute for SImode ssemov types. Reorder operand constraints.
26075 2013-03-20 Martin Jambor <mjambor@suse.cz>
26077 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
26078 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
26080 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
26082 * config/rs6000/predicates.md (indexed_address, update_address_mem
26083 update_indexed_address_mem): New predicates.
26084 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
26085 attribute for load/store instructions.
26086 * config/rs6000/dfp.md (movsd_store): Likewise.
26087 (movsd_load): Likewise.
26088 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
26089 (unnamed HI->DI extend define_insn): Likewise.
26090 (unnamed SI->DI extend define_insn): Likewise.
26091 (unnamed QI->SI extend define_insn): Likewise.
26092 (unnamed QI->HI extend define_insn): Likewise.
26093 (unnamed HI->SI extend define_insn): Likewise.
26094 (unnamed HI->SI extend define_insn): Likewise.
26095 (extendsfdf2_fpr): Likewise.
26096 (movsi_internal1): Likewise.
26097 (movsi_internal1_single): Likewise.
26098 (movhi_internal): Likewise.
26099 (movqi_internal): Likewise.
26100 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
26101 attribute for load/store instructions.
26102 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
26104 (mov<mode>_softfloat): Likewise.
26105 (mov<mode>_hardfloat32): Likewise.
26106 (mov<mode>_hardfloat64): Likewise.
26107 (mov<mode>_softfloat64): Likewise.
26108 (movdi_internal32): Likewise.
26109 (movdi_internal64): Likewise.
26110 (probe_stack_<mode>): Likewise.
26112 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
26114 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
26115 floating point, and decimal floating point to reload iterator.
26117 * config/rs6000/constraints.md (wl constraint): New constraints to
26118 return FLOAT_REGS if certain options are used to reduce the number
26119 of separate patterns that exist in the file.
26120 (wx constraint): Likewise.
26121 (wz constraint): Likewise.
26123 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
26124 -mdebug=reg, print wg, wl, wx, and wz constraints.
26125 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
26126 Initialize the reload functions for 64-bit binary/decimal floating
26128 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
26129 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
26130 create the buffer on the stack to overcome not having a 32-bit
26132 (rs6000_emit_move): Likewise.
26133 (rs6000_secondary_memory_needed_rtx): Likewise.
26134 (rs6000_alloc_sdmode_stack_slot): Likewise.
26135 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
26136 via xxlxor, just like DFmode 0.0.
26138 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
26139 define as 1 if we are running on a power7 or newer.
26140 (enum r6000_reg_class_enum): Add new constraints.
26142 * config/rs6000/dfp.md (movsd): Delete, combine with binary
26143 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
26144 with other moves by using conditional constraits (wg). Use LFIWZX
26145 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
26146 (movsd splitter): Likewise.
26147 (movsd_hardfloat): Likewise.
26148 (movsd_softfloat): Likewise.
26150 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
26151 binary and decimal floating point moves.
26152 (fmove_ok): New attributes to combine binary and decimal floating
26153 point moves, and to combine power6x (mfpgpr) moves along normal
26155 (real_value_to_target): Likewise.
26156 (f32_lr): Likewise.
26157 (f32_lm): Likewise.
26158 (f32_li): Likewise.
26159 (f32_sr): Likewise.
26160 (f32_sm): Likewise.
26161 (f32_si): Likewise.
26162 (movsf): Combine binary and decimal floating point moves. Combine
26163 power6x (mfpgpr) moves with other moves by using conditional
26164 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
26165 (mov<mode> for SFmode/SDmode); Likewise.
26166 (SFmode/SDmode splitters): Likewise.
26167 (movsf_hardfloat): Likewise.
26168 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
26169 (movsf_softfloat): Likewise.
26170 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
26172 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
26173 wx and wz constraints.
26175 * config/rs6000/constraints.md (wg constraint): New constraint to
26176 return FLOAT_REGS if -mmfpgpr (power6x) was used.
26178 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
26181 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
26182 -mdebug=reg, print wg, wl, wx, and wz constraints.
26183 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
26184 Initialize the reload functions for 64-bit binary/decimal floating
26186 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
26187 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
26188 create the buffer on the stack to overcome not having a 32-bit
26190 (rs6000_emit_move): Likewise.
26191 (rs6000_secondary_memory_needed_rtx): Likewise.
26192 (rs6000_alloc_sdmode_stack_slot): Likewise.
26193 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
26194 via xxlxor, just like DFmode 0.0.
26196 * config/rs6000/dfp.md (movdd): Delete, combine with binary
26197 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
26198 with other moves by using conditional constraits (wg). Use LFIWZX
26199 and STFIWX for loading SDmode on power7.
26200 (movdd splitters): Likewise.
26201 (movdd_hardfloat32): Likewise.
26202 (movdd_softfloat32): Likewise.
26203 (movdd_hardfloat64_mfpgpr): Likewise.
26204 (movdd_hardfloat64): Likewise.
26205 (movdd_softfloat64): Likewise.
26207 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
26208 64-bit binary and decimal floating point moves.
26209 (FMOVE64X): Likewise.
26210 (movdf): Combine 64-bit binary and decimal floating point moves.
26211 Combine power6x (mfpgpr) moves with other moves by using
26212 conditional constraits (wg).
26213 (mov<mode> for DFmode/DDmode): Likewise.
26214 (DFmode/DDmode splitters): Likewise.
26215 (movdf_hardfloat32): Likewise.
26216 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
26217 (movdf_softfloat32): Likewise.
26218 (movdf_hardfloat64_mfpgpr): Likewise.
26219 (movdf_hardfloat64): Likewise.
26220 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
26221 (movdf_softfloat64): Likewise.
26222 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
26223 (reload_<mode>_load): Move to later in the file so they aren't in
26224 the middle of the floating point move insns.
26225 (reload_<mode>_store): Likewise.
26227 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
26230 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
26231 constraint if -mdebug=reg.
26232 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
26233 Enable using dd reload support if needed.
26235 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
26236 binary and decimal floating point moves in rs6000.md.
26237 (movtd_internal): Likewise.
26239 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
26240 decimal floating point moves.
26242 (movtf_internal): Likewise.
26243 (mov<mode>_internal, TDmode/TFmode): Likewise.
26244 (movtf_softfloat): Likewise.
26245 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
26247 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
26248 movdi_internal64, using wg constraint for move direct operations.
26249 (movdi_internal64): Likewise.
26251 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
26252 MODES_TIEABLE_P for selected modes. Print the numerical value of
26253 the various virtual registers. Use GPR/FPR first/last values,
26254 instead of hard coding the register numbers. Print which modes
26255 have reload functions registered.
26256 (rs6000_option_override_internal): If -mdebug=reg, trace the options
26257 settings before/after setting cpu, target and subtarget settings.
26258 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
26259 and for secondary reload failures in rs6000_secondary_reload_inner.
26260 (rs6000_secondary_reload_fail): Likewise.
26261 (rs6000_secondary_reload_inner): Likewise.
26263 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
26264 macros for first/last GPR and FPR registers.
26265 (LAST_GPR_REGNO): Likewise.
26266 (FIRST_FPR_REGNO): Likewise.
26267 (LAST_FPR_REGNO): Likewise.
26269 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
26270 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
26271 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
26272 (vcond<mode><mode>): Likewise.
26273 (vcondu<mode><mode>): Likewise.
26274 (vector_gtu<mode>): Likewise.
26275 (vector_gte<mode>): Likewise.
26276 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
26277 to prevent the compiler from converting DImode operations to TImode.
26278 (ior<mode>3): Likewise.
26279 (and<mode>3): Likewise.
26280 (one_cmpl<mode>2): Likewise.
26281 (nor<mode>3): Likewise.
26282 (andc<mode>3): Likewise.
26284 * config/rs6000/constraints.md (wt constraint): New constraint
26285 that returns VSX_REGS if TImode is allowed in VSX registers.
26287 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
26288 constant under VSX.
26290 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
26291 similar to TImode, but it is restricted to being in the GPRs.
26293 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
26294 TImode to occupy a single VSX register.
26296 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
26297 -mvsx-timode for power7/power8.
26298 (power7 cpu): Likewise.
26299 (power8 cpu): Likewise.
26301 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
26302 sure that TFmode/TDmode take up two registers if they are ever
26303 allowed in the upper VSX registers.
26304 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
26306 (rs6000_init_hard_regno_mode_ok): Likewise.
26307 (rs6000_debug_reg_global): Add debugging for PTImode and wt
26308 constraint. Print if LRA is turned on.
26309 (rs6000_option_override_internal): Give an error if -mvsx-timode
26310 and VSX is not enabled.
26311 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
26312 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
26313 to reg+offset addressing. Use PTImode when checking offset
26314 addresses for validity.
26315 (reg_offset_addressing_ok_p): Likewise.
26316 (rs6000_legitimate_offset_address_p): Likewise.
26317 (rs6000_legitimize_address): Likewise.
26318 (rs6000_legitimize_reload_address): Likewise.
26319 (rs6000_legitimate_address_p): Likewise.
26320 (rs6000_eliminate_indexed_memrefs): Likewise.
26321 (rs6000_emit_move): Likewise.
26322 (rs6000_secondary_reload): Likewise.
26323 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
26324 reloads to fpr registers to continue to use reg+offset addressing,
26325 but 64-bit reloads to altivec registers need reg+reg addressing.
26326 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
26327 it. Treat LO_SUM like a PLUS operation.
26328 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
26329 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
26330 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
26331 registers to share a register with a smaller sized type, since VSX
26332 puts scalars in the upper 64-bits.
26333 (print_operand): Add support for PTImode.
26334 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
26335 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
26336 registers, but don't have arithmetic support.
26337 (rs6000_memory_move_cost): Add test for VSX.
26338 (rs6000_opt_masks): Add -mvsx-timode.
26340 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
26343 (VSr): Use wt constraint for TImode.
26344 (VSv): Drop TImode support.
26345 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
26346 (vsx_movti_64bit): Likewise.
26347 (vsx_movti_32bit): Likewise.
26348 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
26349 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
26350 one '?' on the appropriate output constraint. Do not allow TImode
26351 logical operations on 32-bit systems.
26352 (vsx_ior<mode>3): Likewise.
26353 (vsx_xor<mode>3): Likewise.
26354 (vsx_one_cmpl<mode>2): Likewise.
26355 (vsx_nor<mode>3): Likewise.
26356 (vsx_andc<mode>3): Likewise.
26357 (vsx_concat_<mode>): Likewise.
26358 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
26360 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
26361 OPTION_MASK_VSX_TIMODE.
26362 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
26363 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
26365 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
26366 (TI2 iterator): New iterator for TImode, PTImode.
26367 (wd mode attribute): Add values for vector types.
26368 (movti_string): Replace TI move operations with operations for TImode
26369 and PTImode. Add support for TImode being allowed in VSX registers.
26370 (mov<mode>_string, TImode/PTImode): Likewise.
26371 (movti_ppc64): Likewise.
26372 (mov<mode>_ppc64, TImode/PTImode): Likewise.
26373 (TI mode splitters): Likewise.
26375 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
26378 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
26380 PR tree-optimization/56355
26381 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
26382 Also handle integers with undefined overflow.
26384 2013-03-20 Catherine Moore <clm@codesourcery.com>
26385 Maciej W. Rozycki <macro@codesourcery.com>
26386 Tom de Vries <tom@codesourcery.com>
26387 Nathan Sidwell <nathan@codesourcery.com>
26388 Iain Sandoe <iain@codesourcery.com>
26389 Nathan Froyd <froydnj@codesourcery.com>
26390 Chao-ying Fu <fu@mips.com>
26392 * doc/extend.texi: (micromips, nomicromips, nocompression):
26393 Document new function attributes.
26394 * doc/invoke.texi (minterlink-compressed, mmicromips,
26395 m14k, m14ke, m14kec): Document new options.
26396 (minterlink-mips16): Update documentation.
26397 * doc/md.texi (ZC, ZD): Document new constraints.
26398 * configure.ac (gcc_cv_as_micromips): Check if linker
26399 supports the .set micromips directive.
26400 * configure: Regenerate.
26401 * config.in: Regenerate.
26402 * config/mips/mips-tables.opt: Regenerate.
26403 * config/mips/micromips.md: New file.
26404 * constraints.md (ZC, ZD): New constraints.
26405 * config/mips/predicates.md (movep_src_register): New predicate.
26406 (movep_src_operand): New predicate.
26407 (non_volatile_mem_operand): New predicate.
26408 * config/mips/mips.md (multimem): New type.
26409 (length): Differentiate between 17-bit and 18-bit branch offsets.
26410 (MOVEP1, MOVEP2): New mode iterator.
26411 (mov_<load>l): Use ZC constraint.
26412 (mov_<load>r): Likewise.
26413 (mov_<store>l): Likewise.
26414 (mov_<store>r): Likewise.
26415 (*branch_equality<mode>_inverted): Add microMIPS support.
26416 (*branch_equality<mode>): Likewise.
26417 (*jump_absolute): Likewise.
26418 (indirect_jump_<mode>): Likewise.
26419 (tablejump_<mode>): Likewise.
26420 (<optab>_internal): Likewise.
26421 (sibcall_internal): Likewise.
26422 (sibcall_value_internal): Likewise.
26423 (prefetch): Use constraint ZD.
26424 * config/mips/mips.opt (minterlink-compressed): New option.
26425 (minterlink-mips16): Now an alias for minterlink-compressed.
26426 (mmicromips): New option.
26427 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26428 (compare_and_swap_12): Likewise.
26429 (sync_add<mode>): Likewise.
26430 (sync_<optab>_12): Likewise.
26431 (sync_old_<optab>_12): Likewise.
26432 (sync_new_<optab>_12): Likewise.
26433 (sync_nand_12): Likewise.
26434 (sync_old_nand_12): Likewise.
26435 (sync_new_nand_12): Likewise.
26436 (sync_sub<mode>): Likewise.
26437 (sync_old_add<mode>): Likewise.
26438 (sync_old_sub<mode>): Likewise.
26439 (sync_new_add<mode>): Likewise.
26440 (sync_new_sub<mode>): Likewise.
26441 (sync_<optab><mode>): Likewise.
26442 (sync_old_<optab><mode>): Likewise.
26443 (sync_new_<optab><mode>): Likewise.
26444 (sync_nand<mode>): Likewise.
26445 (sync_old_nand<mode>): Likewise.
26446 (sync_new_nand<mode>): Likewise.
26447 (sync_lock_test_and_set<mode>): Likewise.
26448 (test_and_set_12): Likewise.
26449 (atomic_compare_and_swap<mode>): Likewise.
26450 (atomic_exchange<mode>_llsc): Likewise.
26451 (atomic_fetch_add<mode>_llsc): Likewise.
26452 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26453 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26454 (umips_save_restore_pattern_p): Likewise.
26455 (umips_load_store_pair_p): Likewise.
26456 (umips_output_load_store_pair): Likewise.
26457 (umips_movep_target_p): Likewise.
26458 (umips_12bit_offset_address_p): Likewise.
26459 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26460 (mips_base_mips16): Rename this...
26461 (mips_base_compression_flags): ...to this. Update all uses.
26462 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26463 (mips_mips16_decl_p): Delete.
26464 (mips_nomips16_decl_p): Delete.
26465 (mips_get_compress_on_flags): New function.
26466 (mips_get_compress_off_flags): New function.
26467 (mips_get_compress_mode): New function.
26468 (mips_get_compress_on_name): New function.
26469 (mips_get_compress_off_name): New function.
26470 (mips_insert_attributes): Support multiple compression types.
26471 (mips_merge_decl_attributes): Likewise.
26472 (umips_12bit_offset_address_p): New function.
26473 (mips_start_function_definition): Emit .set micromips directive.
26474 (mips_call_may_need_jalx_p): New function.
26475 (mips_function_ok_for_sibcall): Add microMIPS support.
26476 (mips_print_operand_punctuation): Support short delay slots and
26478 (umips_swm_mask, umips_swm_encoding): New.
26479 (umips_build_save_restore): New function.
26480 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26481 (was_mips16_p): Remove.
26482 (old_compression_mode): New.
26483 (mips_set_compression_mode): New function.
26484 (mips_set_current_function): Add microMIPS support.
26485 (mips_option_override): Likewise.
26486 (umips_save_restore_pattern_p): New function.
26487 (umips_output_save_restore): New function.
26488 (umips_load_store_pair_p_1): New function.
26489 (umips_load_store_pair_p): New function.
26490 (umips_output_load_store_pair_1): New function.
26491 (umips_output_load_store_pair): New function.
26492 (umips_movep_target_p) New function.
26493 (mips_prepare_pch_save): Add microMIPS support.
26494 * config/mips/mips.h (TARGET_COMPRESSION): New.
26495 (TARGET_CPU_CPP_BUILTINS): Update macro
26496 to use new compression flags and to support microMIPS.
26497 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26498 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26499 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26500 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26501 (ASM_SPEC): Support mmicromips and mno-micromips.
26502 (M16STORE_REG_P): New macro.
26503 (MIPS_CALL): Support TARGET_MICROMIPS.
26504 (MICROMIPS_J): New macro.
26505 (mips_base_mips16): Rename this...
26506 (mips_base_compression_flags): ...to this.
26507 (UMIPS_12BIT_OFFSET_P): New macro.
26508 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26509 (MULTILIB_DIRNAMES): Likewise.
26510 2013-03-20 Richard Biener <rguenther@suse.de>
26512 PR tree-optimization/56661
26513 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
26514 the result does not have to be distinct.
26516 2013-03-20 Richard Biener <rguenther@suse.de>
26518 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
26521 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
26522 Steven Bosscher <steven@gcc.gnu.org>
26524 PR rtl-optimization/56605
26525 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
26527 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26530 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
26531 that require movd instead of movq.
26533 2013-03-20 Richard Biener <rguenther@suse.de>
26535 * tree-ssa-structalias.c (struct variable_info): Add pointer
26536 to the first field of an aggregate with sub-vars. Make
26537 this and the pointer to the next subfield its ID.
26538 (vi_next): New function.
26539 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
26540 storedanything_id, integer_id): Increment by one.
26541 (new_var_info, get_call_vi, lookup_call_clobber_vi,
26542 get_call_clobber_vi): Adjust.
26543 (solution_set_expand): Simplify and speedup.
26544 (solution_set_add): Inline into ...
26545 (set_union_with_increment): ... this. Adjust accordingly.
26546 (do_sd_constraint): Likewise.
26547 (do_ds_constraint): Likewise.
26548 (do_complex_constraint): Simplify.
26549 (build_pred_graph): Adjust.
26550 (solve_graph): Likewise. Simplify and speedup.
26551 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
26552 get_constraint_for_component_ref, get_constraint_for_1,
26553 first_vi_for_offset, first_or_preceding_vi_for_offset,
26554 create_function_info_for, create_variable_info_for_1,
26555 create_variable_info_for, intra_create_variable_infos): Adjust.
26556 (init_base_vars): Push NULL for ID zero.
26557 (compute_points_to_sets): Adjust.
26559 2013-03-20 Richard Biener <rguenther@suse.de>
26561 * cfgloop.c (verify_loop_structure): Streamline and avoid
26562 ICEing on corrupt loop tree.
26563 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
26566 2013-03-20 Richard Biener <rguenther@suse.de>
26568 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
26569 check whether an SSA update is needed.
26571 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
26573 * config/mips/constraints.md (T): Rename to...
26577 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
26578 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
26580 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26582 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
26583 (*subsi3_carryin_uxtw): Likewise.
26585 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26587 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
26588 (*rorsi3_insn_uxtw): Likewise.
26590 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26592 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
26593 (*extrsi5_insn_uxtw): Likewise.
26595 2013-03-19 Richard Biener <rguenther@suse.de>
26597 PR tree-optimization/56273
26598 * passes.c (init_optimization_passes): Move second VRP after DOM.
26600 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
26602 * config/i386/i386.md (*movti_internal): Merge from
26603 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
26604 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
26605 nox64 isa attributes.
26607 2013-03-18 Richard Biener <rguenther@suse.de>
26609 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
26611 (merge_node_constraints): Likewise.
26612 (build_succ_graph): Likewise.
26613 (valid_graph_edge): Inline into single caller.
26614 (unify_nodes): Likewise. Use bitmap_set_bit return value
26616 (scc_visit): Fix formatting and variable use.
26617 (do_sd_constraint): Use gcc_checking_assert.
26618 (do_ds_constraint): Likewise.
26619 (do_complex_constraint): Likewise.
26620 (condense_visit): Likewise. Cleanup.
26621 (dump_pred_graph): New function.
26622 (perform_var_substitution): Dump the pred-graph before
26623 variable substitution.
26624 (find_equivalent_node): Use gcc_checking_assert.
26625 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
26627 2013-03-18 Richard Biener <rguenther@suse.de>
26629 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
26630 Remove cond_expr_stmt_list argument and do not gimplify the
26632 (vect_loop_versioning): Adjust.
26633 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
26634 Cleanup to use less temporaries.
26635 (vect_create_data_ref_ptr): Cleanup.
26637 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26639 PR tree-optimization/56635
26640 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
26641 require types_compatible_p types.
26643 2013-03-18 Nick Clifton <nickc@redhat.com>
26645 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
26646 spurious backslash.
26648 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
26649 Add missing line to comment describing function.
26651 2013-03-18 Richard Biener <rguenther@suse.de>
26653 PR tree-optimization/56210
26654 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26655 Handle string / character search functions.
26656 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
26658 2013-03-18 Richard Biener <rguenther@suse.de>
26660 PR middle-end/56483
26661 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
26662 and implement properly.
26663 * gimple.h (gimple_cond_single_var_p): Remove.
26665 2013-03-18 Richard Biener <rguenther@suse.de>
26667 * tree-data-ref.h (find_data_references_in_loop): Declare.
26668 * tree-data-ref.c (get_references_in_stmt): Use a stack
26669 vector pre-allocated in the callers.
26670 (find_data_references_in_stmt): Adjust.
26671 (graphite_find_data_references_in_stmt): Likewise.
26672 (create_rdg_vertices): Likewise.
26673 (find_data_references_in_loop): Export.
26674 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
26675 Compute dependences here...
26676 (vect_analyze_data_refs): ...not here. When we encounter
26677 a non-vectorizable data reference in basic-block vectorization
26678 truncate the data reference vector. Do not bother to
26679 fixup data-dependence information for gather loads.
26680 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
26681 of data references, as reported.
26683 2013-03-18 Richard Biener <rguenther@suse.de>
26685 PR tree-optimization/3713
26686 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
26687 has_constants and expr.
26688 (stmt_has_constants): Properly valueize SSA names when deciding
26689 whether the stmt has constants.
26691 2013-03-18 Richard Biener <rguenther@suse.de>
26693 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
26694 whole function when there is nothing to do.
26695 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
26696 * tree-vectorizer.c (vectorize_loops): Update virtual and
26697 loop-closed SSA once.
26698 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
26700 2013-03-18 Richard Biener <rguenther@suse.de>
26702 PR middle-end/56113
26703 * domwalk.c (bb_postorder): New global static.
26704 (cmp_bb_postorder): New function.
26705 (walk_dominator_tree): Replace scheme imposing an order for
26706 visiting dominator sons by one sorting them at the time they
26707 are pushed on the stack.
26709 2013-03-18 Richard Biener <rguenther@suse.de>
26711 PR tree-optimization/39326
26712 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
26713 (struct mem_ref): Replace mem member with ao_ref typed member.
26714 (MEM_ANALYZABLE): Adjust.
26715 (memref_eq): Likewise.
26716 (mem_ref_alloc): Likewise.
26717 (gather_mem_refs_stmt): Likewise.
26718 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
26719 (execute_sm_if_changed_flag_set): Adjust.
26720 (execute_sm): Likewise.
26721 (ref_always_accessed_p): Likewise.
26722 (refs_independent_p): Likewise.
26723 (can_sm_ref_p): Likewise.
26725 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26728 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
26729 return 1 even for !unsignedp.
26731 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26733 * config/i386/i386.md (isa): Add x64 and nox64.
26734 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
26735 (*pushtf): Enable *roF alternative for x64 isa only.
26736 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
26737 mode attribute of integer alternatives to DImode for TARGET_64BIT.
26738 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
26739 (*movtf_internal): Merge from *movtf_internal_rex64 and
26740 *movtf_internal_sse. Use x64 and nox64 isa attributes.
26741 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
26742 nox64 isa attributes.
26743 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
26744 nox64 isa attributes.
26745 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
26747 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26749 * config/alpha/alpha.c (TARGET_LRA_P): New define.
26751 2013-03-17 Jakub Jelinek <jakub@redhat.com>
26754 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
26755 class names. Remove trailing comma after "ALL_REGS".
26757 2013-03-16 Jan Hubicka <jh@suse.cz>
26759 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
26760 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
26761 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
26762 of cgraph_get_create_node.
26763 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
26765 2013-03-16 Jason Merrill <jason@redhat.com>
26768 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
26769 with DW_AT_default_value.
26771 2013-03-16 Jakub Jelinek <jakub@redhat.com>
26773 * BASE-VER: Set to 4.9.0.
26775 2013-03-14 Andi Kleen <ak@linux.intel.com>
26778 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
26779 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
26780 Document _x* TSX intrinsics.
26782 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
26783 David Holsgrove <david.holsgrove@xilinx.com>
26785 * configure.ac: Add MicroBlaze TLS support detection.
26786 * configure: Regenerate.
26787 * config/microblaze/microblaze-protos.h
26788 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
26789 symbol_mentioned_p, label_mentioned_p): Add prototypes.
26790 * config/microblaze/microblaze.c (microblaze_address_type): Add
26791 ADDRESS_TLS and tls_reloc address types.
26792 (microblaze_address_info): Add tls_reloc.
26793 (TARGET_HAVE_TLS): Define.
26794 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
26795 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
26796 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
26797 load_tls_operand, microblaze_call_tls_get_addr,
26798 microblaze_legitimize_tls_address): New functions.
26799 (microblaze_classify_unspec): Handle UNSPEC_TLS.
26800 (get_base_reg): Use microblaze_tls_symbol_p.
26801 (microblaze_classify_address): Handle TLS.
26802 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
26803 label_mentioned_p and microblaze_tls_referenced_p.
26804 (microblaze_legitimize_address): Handle TLS.
26805 (microblaze_address_insns): Handle ADDRESS_TLS.
26806 (pic_address_needs_scratch): Handle TLS.
26807 (print_operand_address): Handle TLS.
26808 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
26809 (microblaze_expand_move): Handle TLS.
26810 (microblaze_legitimate_constant_p): Check
26811 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
26812 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
26813 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
26814 (PIC_OFFSET_TABLE_REGNUM): Set.
26815 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
26816 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
26817 (addsi3, movsi_internal2, movdf_internal): Update constraints
26818 * config/microblaze/predicates.md (arith_plus_operand): Define
26819 (move_operand): Redefine as move_src_operand,
26820 check microblaze_tls_referenced_p.
26822 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26824 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
26825 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
26827 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26829 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
26832 2013-03-14 Jakub Jelinek <jakub@redhat.com>
26834 PR tree-optimization/53265
26835 * common.opt (Waggressive-loop-optimizations): New option.
26836 * tree-ssa-loop-niter.c: Include tree-pass.h.
26837 (do_warn_aggressive_loop_optimizations): New function.
26838 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
26839 if number_of_latch_executions returned constant.
26840 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
26841 early. If number_of_latch_executions returned constant, set
26842 nb_iterations_upper_bound back to it.
26843 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
26845 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
26846 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
26848 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
26849 (MULTILIB_OSDIRNAMES): Set.
26850 * genmultilib: If defaultosdirname doesn't start with :: , set
26851 defaultosdirname2 instead, clear it and emit two . multilib_raw
26852 entries instead of just one.
26854 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26856 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26857 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26858 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26859 (SUBTARGET_OVERRIDE_OPTIONS): New.
26861 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26864 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26865 (musermode): Convert to Var(TARGET_USERMODE).
26866 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26867 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26868 * config/sh/sh.c (sh_option_override): Use
26869 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26870 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26872 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26874 (udivsi3_i4_single, divsi3_i4_single): Use
26875 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26877 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26879 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26882 2013-03-13 Richard Biener <rguenther@suse.de>
26884 PR tree-optimization/56608
26885 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26886 calls when vectorizing basic-blocks.
26888 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26891 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26894 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26896 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26898 2013-03-11 Jan Hubicka <jh@suse.cz>
26901 * lto-streamer-out.c (output_symbol_p): Skip references from
26902 constructors of external variables.
26904 2013-03-11 Jan Hubicka <jh@suse.cz>
26906 PR middle-end/56571
26907 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26909 * emit-rtl.c (verify_rtx_sharing): Likewise.
26910 (copy_insn_1): Likewise.
26911 * rtl.c (copy_rtx): Likewise.
26913 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26916 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26917 output_operand_lossage message.
26919 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26922 * arm.c (shift_op): Validate RTL pattern on the fly.
26923 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26926 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26929 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26930 new call patterns that clobber %fr12.
26931 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26932 split and postreload patterns.
26933 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26934 registers %fr12 and %fr12R as call used.
26936 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26938 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26939 (canon_address, record_store, replace_read, check_mem_read_rtx,
26940 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26941 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26942 rest_of_handle_dse): Likewise.
26944 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26946 PR middle-end/56524
26947 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26949 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26950 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26951 (save_optabs_if_changed): Replace with...
26952 (init_tree_optimization_optabs): ...this.
26953 * optabs.c (save_optabs_if_changed): Rename to...
26954 (init_tree_optimization_optabs): ...this. Take the optimization node
26955 as argument. Do nothing if the base optabs are already correct.
26956 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26957 to recompute optabs.
26958 * function.h (function): Remove optabs field.
26959 * function.c (invoke_set_current_function_hook): Call
26960 init_tree_optimization_optabs. Use the result to initialize
26963 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26965 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26966 if GTMA_HAS_NO_INSTRUMENTATION.
26967 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26968 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26969 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26970 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26971 GTMA_HAS_NO_INSTRUMENTATION.
26973 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26975 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26978 2013-03-08 Marek Polacek <polacek@redhat.com>
26979 Jakub Jelinek <jakub@redhat.com>
26981 PR tree-optimization/56478
26982 * predict.c (is_comparison_with_loop_invariant_p): Change the
26983 type of loop_step to tree.
26984 (predict_loops): Adjust.
26985 (predict_iv_comparison): Perform the computations on double_ints.
26987 2013-03-08 Richard Biener <rguenther@suse.de>
26989 PR tree-optimization/56570
26990 * tree-cfg.c (verify_expr_location_1): Verify locations for
26992 * tree-sra.c (create_access_replacement): Strip locations
26993 from DECL_DEBUG_EXPRs.
26995 2013-03-08 Richard Biener <rguenther@suse.de>
26997 * tree-inline.c (expand_call_inline): Do not associate
26998 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26999 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
27001 2013-03-08 Richard Biener <rguenther@suse.de>
27003 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
27004 or block changes with -Og. Fix for location / block encoding
27005 changes and PHI arguments with locations.
27007 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
27009 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
27011 (struct output_info): Likewise.
27012 (register_overhead): Remove bad gcc_assert.
27013 (bitmap_find_bit): If there is only a single bitmap element, do not
27014 count a miss as a search.
27015 (print_statistics): Update for counter type changes.
27016 (dump_bitmap_statistics): Likewise. Print headers such that they
27017 are properly lined up with the printed counters.
27019 2013-03-07 Jakub Jelinek <jakub@redhat.com>
27021 PR tree-optimization/56559
27022 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
27023 check that it has only a single use.
27025 2013-03-07 Richard Biener <rguenther@suse.de>
27027 * doc/invoke.texi (fwhole-program): Discourage use in combination
27030 2013-03-06 Jakub Jelinek <jakub@redhat.com>
27032 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
27034 PR tree-optimization/56539
27035 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
27036 instead of GSI_CONTINUE_LINKING as last argument to
27037 force_gimple_operand_gsi. Adjust function comment.
27039 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
27042 PR middle-end/56548
27043 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
27044 promoted mode, convert the result back to the original mode.
27046 2013-03-06 Richard Biener <rguenther@suse.de>
27048 PR middle-end/56294
27049 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
27050 (insert_updated_phi_nodes_compare_uids): New function.
27051 (update_ssa): Sort symbols_to_rename after UID before
27052 traversing it to insert PHI nodes.
27054 2013-03-06 Richard Biener <rguenther@suse.de>
27056 PR middle-end/50494
27057 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
27058 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
27061 2013-02-13 Richard Biener <rguenther@suse.de>
27064 * varasm.c (output_constant_def_1): Get the decl representing
27065 the constant as argument.
27066 (output_constant_def): Wrap output_constant_def_1.
27067 (make_decl_rtl): Use output_constant_def_1 with the decl
27068 representing the constant.
27069 (build_constant_desc): Optionally re-use a decl already
27070 representing the constant.
27071 (tree_output_constant_def): Adjust.
27073 2013-03-06 Joey Ye <joey.ye@arm.com>
27076 * gcc.c (convert_white_space): New function.
27077 (main): Handles white space in function name.
27079 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
27082 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
27083 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
27084 to SH_DIV_CALL_TABLE for TARGET_SH2.
27085 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
27087 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
27088 call-table options.
27090 2013-03-05 Sterling Augustine <saugustine@google.com>
27091 Cary Coutant <ccoutant@google.com>
27094 * dwarf2out.c (resolve_addr): Don't call
27095 remove_loc_list_addr_table_entries a second time for the same
27098 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27101 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
27102 (avoid_complex_debug_insns): New function.
27103 (expand_debug_locations): Call it.
27105 PR rtl-optimization/56484
27106 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
27107 lifetimes of hard registers on small register class machines.
27109 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
27111 * config/microblaze/microblaze-protos.h: Rename
27112 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
27113 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
27115 (microblaze_fast_interrupt_function_p): New function.
27116 (microblaze_is_interrupt_handler): Rename to
27117 microblaze_is_interrupt_variant and add fast_interrupt check.
27118 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
27119 (save_restore_insns): Likewise.
27120 (compute_frame_size): Likewise.
27121 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
27122 (microblaze_globalize_label): Likewise.
27123 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
27124 * config/microblaze/microblaze.md: Use wrapper
27125 microblaze_is_interrupt_variant.
27127 2013-03-05 Kai Tietz <ktietz@redhat.com>
27129 * sdbout.c (sdbout_one_type): Switch to current function's section
27130 supporting cold/hot.
27132 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
27134 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
27137 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27139 PR middle-end/56461
27140 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
27141 if VALGRIND_GET_VBITS is defined, temporarily make object
27142 memory all defined, and restore previous valgrind addressability
27143 and definability afterwards. Free this_object at the end.
27145 PR middle-end/56461
27146 * lra.c (lra): Call lra_clear_live_ranges if live_p,
27147 right before calling lra_create_live_ranges, also call it
27148 when clearing live_p. Only call lra_clear_live_ranges
27149 at the end if live_p.
27151 PR middle-end/56461
27152 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
27154 2013-03-05 Richard Biener <rguenther@suse.de>
27156 PR tree-optimization/56521
27157 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
27160 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
27163 * except.h (remove_unreachable_eh_regions): New prototype.
27164 * except.c (remove_eh_handler_splicer): New function, split out
27165 of remove_eh_handler.
27166 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
27167 warning about running it on many EH regions one at a time.
27168 (remove_unreachable_eh_regions_worker): New function, walk the
27169 EH tree in depth-first order and remove non-marked regions.
27170 (remove_unreachable_eh_regions): New function.
27171 * tree-eh.c (mark_reachable_handlers): New function, split out
27172 from remove_unreachable_handlers.
27173 (remove_unreachable_handlers): Use mark_reachable_handlers and
27174 remove_unreachable_eh_regions.
27175 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
27176 and remove_unreachable_eh_regions.
27178 2013-03-05 Richard Biener <rguenther@suse.de>
27180 PR middle-end/56525
27181 * loop-init.c (fix_loop_structure): Remove loops in two stages,
27182 not freeing them until the end.
27184 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27186 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
27188 2013-03-05 Richard Biener <rguenther@suse.de>
27190 PR tree-optimization/56270
27191 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
27192 of loads after scheduling an SLP instance.
27194 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27196 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
27198 (check_gcc_parallelize): Run guality.exp as a separate job from
27199 vect.exp with unsorted.exp and $(dg_target_exps) separately from
27200 struct-layout-1.exp with stackalign.exp.
27202 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
27204 PR middle-end/56461
27205 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
27206 load_index sbitmap even if some bit in it isn't set.
27208 PR middle-end/56461
27209 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
27210 (discover_iteration_bound_by_body_walk): Change queues to
27211 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
27212 spelling in comment. Call safe_push on queues[bound_index] directly.
27213 Release queues[queue_index] in every iteration unconditionally.
27214 Release bounds vector.
27216 PR middle-end/56461
27217 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
27218 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
27219 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
27222 2013-03-05 Richard Biener <rguenther@suse.de>
27225 * tree-inline.c (remap_blocks_to_null): New function.
27226 (expand_call_inline): When expanding a call stmt without
27227 an associated block inline remap all callee blocks to NULL.
27229 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27231 PR rtl-optimization/56494
27232 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
27233 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
27234 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
27236 PR middle-end/56461
27237 * sel-sched-ir.c (free_sched_pools): Release
27238 succs_info_pool.stack[succs_info_pool.max_top] vectors too
27239 if succs_info_pool.max_top isn't -1.
27242 * opts.c (opts_obstack, opts_concat): Moved to...
27243 * opts-common.c (opts_obstack, opts_concat): ... here.
27245 2013-03-04 Jakub Jelinek <jakub@redhat.com>
27247 PR middle-end/56461
27248 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
27250 2013-03-04 Martin Jambor <mjambor@suse.cz>
27252 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
27253 all appropriate places.
27255 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
27257 PR tree-optimization/56424
27258 * ipa-split.c (split_function): Do not set the RSO flag if result is
27259 not by reference and its type is a register type.
27261 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
27263 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
27264 (microblaze_legitimate_pic_operand): Likewise
27265 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
27266 new function microblaze_legitimate_pic_operand
27267 * config/microblaze/microblaze-protos.h
27268 (microblaze_legitimate_pic_operand): Declare.
27270 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27272 * config/microblaze/predicates.md (call_insn_simple_operand):
27273 New predicate for supported rtx code types.
27274 * config/microblaze/microblaze.md (call_internal1): Use
27275 call_insn_simple_operand predicate.
27277 2013-03-04 Jakub Jelinek <jakub@redhat.com>
27279 PR middle-end/56461
27280 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
27281 partitions.ordered_remove.
27283 PR middle-end/56461
27284 * tree-vect-stmts.c (vectorizable_conversion): Don't call
27285 vec_oprnds0.create (1) for modifier == NONE.
27287 PR middle-end/56461
27288 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
27289 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
27290 vec_oprnds1 right before pushing anything to it for
27293 PR middle-end/56461
27294 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
27295 set nbbs to 0 instead of having separate code path.
27296 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
27297 instead of false as last argument if returning NULL.
27299 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
27301 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
27302 the attribute is now called "target" instead of "option".
27303 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
27304 * doc/tm.texi.in (Target Attributes): Likewise document the correct
27305 attribute/pragma name for TARGET_OPTION_VALID_P and
27306 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
27307 * doc/tm.texi: Regenerated.
27309 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
27311 * config/microblaze/microblaze.c:
27312 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
27313 * config/microblaze/microblaze.h: Add -mxl-reorder to
27315 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
27316 instructions emitted if TARGET_REORDER.
27317 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
27318 or 0 for -m/-mno case, but initialises as 2 to detect default use case
27321 2013-03-01 Xinliang David Li <davidxl@google.com>
27323 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
27326 2013-03-01 Jakub Jelinek <jakub@redhat.com>
27328 PR middle-end/56461
27329 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
27330 vector even when returning true. Fix up function comment formatting.
27332 PR middle-end/56461
27333 * ira-build.c (ira_loop_nodes_count): New variable.
27334 (create_loop_tree_nodes): Initialize it.
27335 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
27337 PR middle-end/56461
27338 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
27339 method on dr_chain and result_chain.
27340 * tree-vect-stmts.c (vectorizable_store): Only call
27341 result_chain.create if j == 0.
27343 PR middle-end/56461
27344 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
27345 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
27346 before overwriting it.
27348 2013-03-01 Tobias Burnus <burnus@net-b.de>
27350 * doc/extended.texi (C Extensions): Change order in @menu
27352 (Other MIPS Built-in Functions): Move last MIPS entry before
27353 "picoChip Built-in Functions".
27354 (SH Built-in Functions): Move after RX Built-in Functions.
27355 * doc/gcc.texi (Introduction): Change order in @menu
27357 * doc/md.texi (Constraints): Ditto.
27358 * gty.texi (Type Information): Ditto.
27359 (User-provided marking routines for template types): Make
27361 * doc/invoke.texi (AArch64 Options): Move before
27362 "Adapteva Epiphany Options".
27364 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
27365 Jakub Jelinek <jakub@redhat.com>
27368 * asan.c (gate_asan): Lookup no_sanitize_address instead of
27369 no_address_safety_analysis attribute.
27370 * doc/extend.texi (no_address_safety_attribute): Rename to
27371 no_sanitize_address attribute, mention no_address_safety_analysis
27372 attribute as deprecated alias.
27374 2013-02-28 Jakub Jelinek <jakub@redhat.com>
27376 PR middle-end/56461
27377 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
27378 type to vec<vec<tree> > *.
27379 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
27380 to be vec<tree> instead of vec<tree> *, set vec_defs
27381 to vNULL and call vec_defs.create (number_of_vects), adjust other
27383 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
27384 vectorizable_condition): Adjust vect_get_slp_defs callers.
27386 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27388 * config/aarch64/aarch64.c
27389 (aarch64_float_const_representable): Remove unused variable.
27391 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27393 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
27395 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27397 * config/aarch64/aarch64-builtins.c
27398 (aarch64_init_simd_builtins): Make static.
27400 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27402 * config/aarch64/aarch64.c
27403 (aarch64_simd_make_constant): Make static.
27405 2013-02-28 Martin Jambor <mjambor@suse.cz>
27407 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
27408 with no initialization to the RHS of debug statements.
27410 2013-02-28 Martin Jambor <mjambor@suse.cz>
27412 PR tree-optimization/56294
27413 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
27415 (get_access_replacement): Do not call create_access_replacement.
27416 Assert a replacement exists.
27417 (get_repl_default_def_ssa_name): Create the replacement declaration
27420 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
27422 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
27423 final_end_function.
27425 2013-02-28 Marek Polacek <polacek@redhat.com>
27427 PR rtl-optimization/56466
27428 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
27429 if we're changing a loop.
27430 (peel_loops_completely): Likewise.
27432 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
27435 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
27437 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27440 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
27441 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
27442 INTX_FTYPE_FX, FX_FTYPE_INTX.
27443 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
27445 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27447 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
27448 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
27449 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
27450 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
27451 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
27452 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
27453 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
27454 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
27455 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
27456 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
27457 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
27458 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
27459 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
27460 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
27461 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
27462 (avrxmega6): Increase max flash segments from 5 to 6.
27463 * config/avr/t-multilib: Regenerate.
27464 * config/avr/avr-tables.opt: Regenerate.
27465 * doc/avr-mmcu.texi: Regenerate.
27467 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27469 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
27470 (avr_device_to_arch): Rename to avr_device_to_ld.
27471 (avr_device_to_as): New prototype.
27472 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
27473 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
27474 * config/avr/driver-avr.c (avr_device_to_as): New.
27475 (avr_device_to_arch): Rename to avr_device_to_ld.
27477 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27479 PR middle-end/56461
27480 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
27481 method on dr_chain and result_chain.
27483 PR middle-end/56461
27484 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
27485 pointer_set_destroy on not_executed_last_iteration.
27487 PR middle-end/56461
27488 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
27490 PR middle-end/56461
27491 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
27492 FOR_EACH_DEFINED_FUNCTION when freeing state.
27494 PR middle-end/56461
27495 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
27497 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
27500 PR middle-end/56461
27501 * ipa-cp.c (decide_whether_version_node): Call vec_free on
27502 known_aggs[i].items and release known_aggs vector.
27504 PR middle-end/56461
27505 * ipa-reference.c (propagate): Free node_info even for alias nodes.
27507 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27509 * config/microblaze/microblaze.c (microblaze_emit_compare):
27510 Use xor for EQ/NE comparisions.
27511 * config/microblaze/microblaze.md (cstoresf4): Add constraints
27512 (cbranchsf4): Adjust operator to comparison_operator.
27514 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27516 PR middle-end/56461
27517 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
27519 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
27520 vec_safe_push, always update *slot.
27521 (redirect_edge_var_map_clear): Use vec_free.
27522 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
27523 (free_var_map_entry): Use vec_free.
27524 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
27525 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
27527 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
27529 PR middle-end/45472
27530 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
27531 when the may_trap_p bit of the exprs being merged differs.
27532 Reorder tests for speculativeness in the logical and operator.
27534 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27536 * incpath.c (add_standard_paths): Use reconcat instead of concat
27537 where appropriate and avoid leaking memory.
27539 * opts.h: Include obstack.h.
27540 (opts_concat): New prototype.
27541 (opts_obstack): New declaration.
27542 * opts.c (opts_concat): New function.
27543 (opts_obstack): New variable.
27544 (init_options_struct): Call gcc_init_obstack on opts_obstack.
27545 (finish_options): Use opts_concat instead of concat
27546 and XOBNEWVEC instead of XNEWVEC.
27547 * opts-common.c (generate_canonical_option, decode_cmdline_option,
27548 generate_option): Likewise.
27549 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
27550 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
27553 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
27554 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
27556 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27558 PR middle-end/56461
27559 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
27561 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27563 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
27564 (arm_block_move_unaligned_straight): Likewise.
27565 (arm_adjust_block_mem): Likewise.
27567 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27570 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
27571 temp, cond and label.
27572 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
27575 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
27576 * config/c6x/c6x.h (dbx_register_map): Update declaration.
27579 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
27580 of prologue/epilogue functions.
27583 * config/tilegx/tilegx.c (tilegx_expand_prologue):
27584 Remove unused variable cfa_offset.
27585 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
27588 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
27589 type promotion to unsigned.
27592 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
27593 for HOST_WIDE_INT of 32 bit / same size as int.
27594 (arm_block_move_unaligned_straight): Likewise.
27595 (arm_adjust_block_mem): Likewise.
27598 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
27601 2013-02-26 Marek Polacek <polacek@redhat.com>
27603 PR tree-optimization/56426
27604 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
27606 2013-02-26 Richard Biener <rguenther@suse.de>
27609 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
27610 unused variable loops.
27612 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27614 PR tree-optimization/56448
27615 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
27616 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
27617 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
27618 later operands of the references, or even first operand for
27619 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
27621 PR tree-optimization/56443
27622 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
27623 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
27624 to type_for_mode langhook.
27626 2013-02-25 Matt Turner <mattst88@gmail.com>
27628 * doc/invoke.texi: Document r4700.
27630 2013-02-25 Richard Biener <rguenther@suse.de>
27632 PR tree-optimization/56175
27633 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
27635 (simplify_bitwise_binary): ... here. Also guard the conversion
27636 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
27638 2013-02-25 Catherine Moore <clm@codesourcery.com>
27641 2013-02-24 Catherine Moore <clm@codesourcery.com>
27642 Maciej W. Rozycki <macro@codesourcery.com>
27643 Tom de Vries <tom@codesourcery.com>
27644 Nathan Sidwell <nathan@codesourcery.com>
27645 Iain Sandoe <iain@codesourcery.com>
27646 Nathan Froyd <froydnj@codesourcery.com>
27647 Chao-ying Fu <fu@mips.com>
27649 * doc/extend.texi: (micromips, nomicromips, nocompression):
27650 Document new function attributes.
27651 * doc/invoke.texi (minterlink-compressed, mmicromips,
27652 m14k, m14ke, m14kec): Document new options.
27653 (minterlink-mips16): Update documentation.
27654 * doc/md.texi (ZC, ZD): Document new constraints.
27655 * configure.ac (gcc_cv_as_micromips): Check if linker
27656 supports the .set micromips directive.
27657 * configure: Regenerate.
27658 * config.in: Regenerate.
27659 * config/mips/mips-tables.opt: Regenerate.
27660 * config/mips/micromips.md: New file.
27661 * constraints.md (ZC, AD): New constraints.
27662 * config/mips/predicates.md (movep_src_register): New predicate.
27663 (movep_src_operand): New predicate.
27664 (non_volatile_mem_operand): New predicate.
27665 * config/mips/mips.md (multimem): New type.
27666 (length): Differentiate between 17-bit and 18-bit branch offsets.
27667 (MOVEP1, MOVEP2): New mode iterator.
27668 (mov_<load>l): Use ZC constraint.
27669 (mov_<load>r): Likewise.
27670 (mov_<store>l): Likewise.
27671 (mov_<store>r): Likewise.
27672 (*branch_equality<mode>_inverted): Add microMIPS support.
27673 (*branch_equality<mode>): Likewise.
27674 (*jump_absolute): Likewise.
27675 (indirect_jump_<mode>): Likewise.
27676 (tablejump_<mode>): Likewise.
27677 (<optab>_internal): Likewise.
27678 (sibcall_internal): Likewise.
27679 (sibcall_value_internal): Likewise.
27680 (prefetch): Use constraint ZD.
27681 * config/mips/mips.opt (minterlink-compressed): New option.
27682 (minterlink-mips16): Now an alias for minterlink-compressed.
27683 (mmicromips): New option.
27684 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27685 (compare_and_swap_12): Likewise.
27686 (sync_add<mode>): Likewise.
27687 (sync_<optab>_12): Likewise.
27688 (sync_old_<optab>_12): Likewise.
27689 (sync_new_<optab>_12): Likewise.
27690 (sync_nand_12): Likewise.
27691 (sync_old_nand_12): Likewise.
27692 (sync_new_nand_12): Likewise.
27693 (sync_sub<mode>): Likewise.
27694 (sync_old_add<mode>): Likewise.
27695 (sync_old_sub<mode>): Likewise.
27696 (sync_new_add<mode>): Likewise.
27697 (sync_new_sub<mode>): Likewise.
27698 (sync_<optab><mode>): Likewise.
27699 (sync_old_<optab><mode>): Likewise.
27700 (sync_new_<optab><mode>): Likewise.
27701 (sync_nand<mode>): Likewise.
27702 (sync_old_nand<mode>): Likewise.
27703 (sync_new_nand<mode>): Likewise.
27704 (sync_lock_test_and_set<mode>): Likewise.
27705 (test_and_set_12): Likewise.
27706 (atomic_compare_and_swap<mode>): Likewise.
27707 (atomic_exchange<mode>_llsc): Likewise.
27708 (atomic_fetch_add<mode>_llsc): Likewise.
27709 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27710 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27711 (umips_save_restore_pattern_p): Likewise.
27712 (umips_load_store_pair_p): Likewise.
27713 (umips_output_load_store_pair): Likewise.
27714 (umips_movep_target_p): Likewise.
27715 (umips_12bit_offset_address_p): Likewise.
27716 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27717 (mips_base_mips16): Rename this...
27718 (mips_base_compression_flags): ...to this. Update all uses.
27719 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27720 (mips_mips16_decl_p): Delete.
27721 (mips_nomips16_decl_p): Delete.
27722 (mips_get_compress_on_flags): New function.
27723 (mips_get_compress_off_flags): New function.
27724 (mips_get_compress_mode): New function.
27725 (mips_get_compress_on_name): New function.
27726 (mips_get_compress_off_name): New function.
27727 (mips_insert_attributes): Support multiple compression types.
27728 (mips_merge_decl_attributes): Likewise.
27729 (umips_12bit_offset_address_p): New function.
27730 (mips_start_function_definition): Emit .set micromips directive.
27731 (mips_call_may_need_jalx_p): New function.
27732 (mips_function_ok_for_sibcall): Add microMIPS support.
27733 (mips_print_operand_punctuation): Support short delay slots and
27735 (umips_swm_mask, umips_swm_encoding): New.
27736 (umips_build_save_restore): New function.
27737 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27738 (was_mips16_p): Remove.
27739 (old_compression_mode): New.
27740 (mips_set_compression_mode): New function.
27741 (mips_set_current_function): Add microMIPS support.
27742 (mips_option_override): Likewise.
27743 (umips_save_restore_pattern_p): New function.
27744 (umips_output_save_restore): New function.
27745 (umips_load_store_pair_p_1): New function.
27746 (umips_load_store_pair_p): New function.
27747 (umips_output_load_store_pair_1): New function.
27748 (umips_output_load_store_pair): New function.
27749 (umips_movep_target_p) New function.
27750 (mips_prepare_pch_save): Add microMIPS support.
27751 * config/mips/mips.h (TARGET_COMPRESSION): New.
27752 (TARGET_CPU_CPP_BUILTINS): Update macro
27753 to use new compression flags and to support microMIPS.
27754 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27755 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27756 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27757 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27758 (ASM_SPEC): Support mmicromips and mno-micromips.
27759 (M16STORE_REG_P): New macro.
27760 (MIPS_CALL): Support TARGET_MICROMIPS.
27761 (MICROMIPS_J): New macro.
27762 (mips_base_mips16): Rename this...
27763 (mips_base_compression_flags): ...to this.
27764 (UMIPS_12BIT_OFFSET_P): New macro.
27765 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27766 (MULTILIB_DIRNAMES): Likewise.
27768 2013-02-25 Tom de Vries <tom@codesourcery.com>
27770 PR rtl-optimization/56131
27771 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
27772 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
27773 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
27775 2013-02-25 Tobias Burnus <burnus@net-b.de>
27777 * doc/invoke.texi (-fsanitize=): Move from optimization
27778 to debugging options.
27780 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27782 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
27784 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27785 Alexander Monakov <amonakov@ispras.ru>
27787 PR middle-end/56077
27788 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
27789 flush pending lists also on non-jumps. Adjust comment.
27791 2013-02-24 Catherine Moore <clm@codesourcery.com>
27792 Maciej W. Rozycki <macro@codesourcery.com>
27793 Tom de Vries <tom@codesourcery.com>
27794 Nathan Sidwell <nathan@codesourcery.com>
27795 Iain Sandoe <iain@codesourcery.com>
27796 Nathan Froyd <froydnj@codesourcery.com>
27797 Chao-ying Fu <fu@mips.com>
27799 * doc/extend.texi: (micromips, nomicromips, nocompression):
27800 Document new function attributes.
27801 * doc/invoke.texi (minterlink-compressed, mmicromips,
27802 m14k, m14ke, m14kec): Document new options.
27803 (minterlink-mips16): Update documentation.
27804 * doc/md.texi (ZC, ZD): Document new constraints.
27805 * configure.ac (gcc_cv_as_micromips): Check if linker
27806 supports the .set micromips directive.
27807 * configure: Regenerate.
27808 * config.in: Regenerate.
27809 * config/mips/mips-tables.opt: Regenerate.
27810 * config/mips/micromips.md: New file.
27811 * constraints.md (ZC, AD): New constraints.
27812 * config/mips/predicates.md (movep_src_register): New predicate.
27813 (movep_src_operand): New predicate.
27814 (non_volatile_mem_operand): New predicate.
27815 * config/mips/mips.md (multimem): New type.
27816 (length): Differentiate between 17-bit and 18-bit branch offsets.
27817 (MOVEP1, MOVEP2): New mode iterator.
27818 (mov_<load>l): Use ZC constraint.
27819 (mov_<load>r): Likewise.
27820 (mov_<store>l): Likewise.
27821 (mov_<store>r): Likewise.
27822 (*branch_equality<mode>_inverted): Add microMIPS support.
27823 (*branch_equality<mode>): Likewise.
27824 (*jump_absolute): Likewise.
27825 (indirect_jump_<mode>): Likewise.
27826 (tablejump_<mode>): Likewise.
27827 (<optab>_internal): Likewise.
27828 (sibcall_internal): Likewise.
27829 (sibcall_value_internal): Likewise.
27830 (prefetch): Use constraint ZD.
27831 * config/mips/mips.opt (minterlink-compressed): New option.
27832 (minterlink-mips16): Now an alias for minterlink-compressed.
27833 (mmicromips): New option.
27834 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27835 (compare_and_swap_12): Likewise.
27836 (sync_add<mode>): Likewise.
27837 (sync_<optab>_12): Likewise.
27838 (sync_old_<optab>_12): Likewise.
27839 (sync_new_<optab>_12): Likewise.
27840 (sync_nand_12): Likewise.
27841 (sync_old_nand_12): Likewise.
27842 (sync_new_nand_12): Likewise.
27843 (sync_sub<mode>): Likewise.
27844 (sync_old_add<mode>): Likewise.
27845 (sync_old_sub<mode>): Likewise.
27846 (sync_new_add<mode>): Likewise.
27847 (sync_new_sub<mode>): Likewise.
27848 (sync_<optab><mode>): Likewise.
27849 (sync_old_<optab><mode>): Likewise.
27850 (sync_new_<optab><mode>): Likewise.
27851 (sync_nand<mode>): Likewise.
27852 (sync_old_nand<mode>): Likewise.
27853 (sync_new_nand<mode>): Likewise.
27854 (sync_lock_test_and_set<mode>): Likewise.
27855 (test_and_set_12): Likewise.
27856 (atomic_compare_and_swap<mode>): Likewise.
27857 (atomic_exchange<mode>_llsc): Likewise.
27858 (atomic_fetch_add<mode>_llsc): Likewise.
27859 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27860 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27861 (umips_save_restore_pattern_p): Likewise.
27862 (umips_load_store_pair_p): Likewise.
27863 (umips_output_load_store_pair): Likewise.
27864 (umips_movep_target_p): Likewise.
27865 (umips_12bit_offset_address_p): Likewise.
27866 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27867 (mips_base_mips16): Rename this...
27868 (mips_base_compression_flags): ...to this. Update all uses.
27869 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27870 (mips_mips16_decl_p): Delete.
27871 (mips_nomips16_decl_p): Delete.
27872 (mips_get_compress_on_flags): New function.
27873 (mips_get_compress_off_flags): New function.
27874 (mips_get_compress_mode): New function.
27875 (mips_get_compress_on_name): New function.
27876 (mips_get_compress_off_name): New function.
27877 (mips_insert_attributes): Support multiple compression types.
27878 (mips_merge_decl_attributes): Likewise.
27879 (umips_12bit_offset_address_p): New function.
27880 (mips_start_function_definition): Emit .set micromips directive.
27881 (mips_call_may_need_jalx_p): New function.
27882 (mips_function_ok_for_sibcall): Add microMIPS support.
27883 (mips_print_operand_punctuation): Support short delay slots and
27885 (umips_swm_mask, umips_swm_encoding): New.
27886 (umips_build_save_restore): New function.
27887 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27888 (was_mips16_p): Remove.
27889 (old_compression_mode): New.
27890 (mips_set_compression_mode): New function.
27891 (mips_set_current_function): Add microMIPS support.
27892 (mips_option_override): Likewise.
27893 (umips_save_restore_pattern_p): New function.
27894 (umips_output_save_restore): New function.
27895 (umips_load_store_pair_p_1): New function.
27896 (umips_load_store_pair_p): New function.
27897 (umips_output_load_store_pair_1): New function.
27898 (umips_output_load_store_pair): New function.
27899 (umips_movep_target_p) New function.
27900 (mips_prepare_pch_save): Add microMIPS support.
27901 * config/mips/mips.h (TARGET_COMPRESSION): New.
27902 (TARGET_CPU_CPP_BUILTINS): Update macro
27903 to use new compression flags and to support microMIPS.
27904 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27905 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27906 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27907 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27908 (ASM_SPEC): Support mmicromips and mno-micromips.
27909 (M16STORE_REG_P): New macro.
27910 (MIPS_CALL): Support TARGET_MICROMIPS.
27911 (MICROMIPS_J): New macro.
27912 (mips_base_mips16): Rename this...
27913 (mips_base_compression_flags): ...to this.
27914 (UMIPS_12BIT_OFFSET_P): New macro.
27915 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27916 (MULTILIB_DIRNAMES): Likewise.
27918 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27921 * target-globals.c (save_target_globals): For init_reg_sets and
27922 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27924 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27926 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27927 * config/aarch64/t-aarch64
27928 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27930 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27932 PR inline-asm/56148
27933 * lra-constraints.c (process_alt_operands): Reload operand
27934 conflicting with earlier clobber only if no more other conflicting
27937 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27940 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27941 if not linking a shared library.
27943 2013-02-22 Seth LaForge <sethml@google.com>
27945 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27947 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27949 * config/arm/arm.md (split for extendsidi): Update condition.
27950 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27951 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27952 (qhs_zextenddi_cstr): Likewise.
27954 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27956 PR middle-end/56420
27957 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27958 avoid signed wrapping.
27959 (expand_mult): Handle properly multiplication by
27960 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27961 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27962 in the compiler if coeff is HOST_WIDE_INT_MIN.
27963 (expand_divmod): Don't make ext_op1 static, change it's type to
27964 uhwi. Avoid undefined behavior in -INTVAL (op1).
27966 PR rtl-optimization/50339
27967 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27969 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27970 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27971 into splitting_ashiftrt field.
27972 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27974 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27977 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27979 PR middle-end/56108
27980 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27981 are sure to go irrevocable.
27983 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27985 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27986 scalars are valid operands.
27988 2013-02-21 Martin Jambor <mjambor@suse.cz>
27990 PR tree-optimization/56310
27991 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27992 only matching indices and non-negative final offsets.
27993 (intersect_aggregates_with_edge): Pass src_idx to
27994 agg_replacements_to_vector. Pass src_idx insstead of index to
27995 intersect_with_agg_replacements.
27997 2013-02-21 Martin Jambor <mjambor@suse.cz>
27999 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
28000 instead of hard-wired defaults.
28002 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
28004 * doc/invoke.texi (MIPS Options): Update documentation of the
28005 floating-point multiply-accumulate instruction restrictions.
28007 2013-02-21 Kostya Serebryany <kcc@google.com>
28009 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
28010 asan_shadow_offset on x86_64 linux.
28012 2013-02-21 Richard Biener <rguenther@suse.de>
28014 PR tree-optimization/56415
28016 2013-02-11 Richard Biener <rguenther@suse.de>
28018 PR tree-optimization/56273
28019 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
28022 2013-02-21 Jakub Jelinek <jakub@redhat.com>
28025 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
28028 PR inline-asm/56405
28029 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
28030 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
28032 2013-02-20 Jan Hubicka <jh@suse.cz>
28034 PR tree-optimization/56265
28035 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
28036 when target is referenced for first time.
28038 2013-02-20 Richard Biener <rguenther@suse.de>
28040 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
28041 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
28042 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
28043 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
28044 not return anything.
28045 (rename_ssa_copies): Do not remove unused locals.
28046 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
28047 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
28048 * passes.c (execute_function_todo): Do not schedule unused locals
28049 removal if cleanup_tree_cfg did something.
28050 * tree-ssa-live.c (remove_unused_locals): Dump statistics
28051 about the number of removed locals.
28053 2013-02-20 Richard Biener <rguenther@suse.de>
28055 PR tree-optimization/56398
28056 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
28058 2013-02-20 Martin Jambor <mjambor@suse.cz>
28060 PR tree-optimization/55334
28061 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
28062 restricted pointers to arrays.
28064 2013-02-20 Richard Biener <rguenther@suse.de>
28065 Jakub Jelinek <jakub@redhat.com>
28067 PR tree-optimization/56396
28068 * tree-ssa-ccp.c (n_const_val): New static variable.
28069 (get_value): Return NULL for SSA names we don't have a lattice
28071 (ccp_initialize): Initialize n_const_val.
28072 * tree-ssa-copy.c (n_copy_of): New static variable.
28073 (init_copy_prop): Initialize n_copy_of.
28074 (get_value): Return NULL_TREE for SSA names we don't have a
28077 2013-02-20 Martin Jambor <mjambor@suse.cz>
28079 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
28081 2013-02-20 Richard Biener <rguenther@suse.de>
28083 * genpreds.c (write_lookup_constraint): Do not compare first
28084 letter of the constraint again.
28086 2013-02-20 Richard Biener <rguenther@suse.de>
28088 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
28090 (get_use_iv_cost): Terminate hashtable walk when coming across
28093 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
28095 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
28096 reassociation for avx2 targets.
28098 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28100 * config/microblaze/microblaze.c: microblaze_has_clz = 0
28101 Add version check for v8.10.a to enable microblaze_has_clz
28102 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
28103 version and TARGET_PATTERN_COMPARE check
28104 * config/microblaze/microblaze.md: New clzsi2 instruction
28106 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28108 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
28109 function before branching.
28111 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
28113 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
28115 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
28117 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
28119 PR middle-end/55889
28120 * sel-sched.c: Include ira.h.
28121 (implicit_clobber_conflict_p): New function.
28122 (moveup_expr): Use it.
28123 * Makefile.in (sel-sched.o): Depend on ira.h.
28125 2013-02-19 Richard Biener <rguenther@suse.de>
28127 PR tree-optimization/56384
28128 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
28129 (vn_hash_type): Split out from ...
28130 (vn_hash_constant_with_type): ... here.
28131 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
28132 (vn_phi_eq): Compare types from vn_phi_s structure.
28133 (vn_phi_lookup): Populate vn_phi_s type.
28134 (vn_phi_insert): Likewise.
28136 2013-02-19 Jakub Jelinek <jakub@redhat.com>
28138 PR tree-optimization/56350
28139 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
28140 if haven't found reduction or nested cycle operand, rather than
28141 asserting we must find it.
28143 PR tree-optimization/56381
28144 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
28147 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
28148 Jakub Jelinek <jakub@redhat.com>
28151 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
28152 (swap_optab_enable): Same.
28153 (init_all_optabs): Use argument instead of global.
28154 * tree.h (struct tree_optimization_option): New field target_optabs.
28155 * expr.h (init_all_optabs): Add argument to prototype.
28156 (TREE_OPTIMIZATION_OPTABS): New.
28157 (save_optabs_if_changed): Protoize.
28158 * optabs.h: Declare this_fn_optabs.
28159 * optabs.c (save_optabs_if_changed): New.
28160 Declare this_fn_optabs.
28161 (init_optabs): Add argument to init_all_optabs() call.
28162 * function.c (invoke_set_current_function_hook): Handle per
28164 * function.h (struct function): New field optabs.
28165 * config/mips/mips.c (mips_set_mips16_mode): Handle when
28166 optimization_current_node has changed.
28167 * target-globals.h (save_target_globals_default_opts): Protoize.
28168 * target-globals.c (save_target_globals_default_opts): New.
28170 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28173 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
28174 registers %fr12 and %fr12R as call used.
28177 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
28178 and HImode, require all displacements to be an integer multiple of
28180 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
28181 only allow QImode and HImode when reload is in progress and strict is
28182 true. Likewise for symbolic addresses. Use base14_operand to check
28183 displacements in REG+BASE addresses.
28185 2013-02-18 Richard Biener <rguenther@suse.de>
28187 PR tree-optimization/56366
28188 * tree-vect-loop.c (get_initial_def_for_induction): Properly
28189 handle sign-conversion of outer-loop initial induction value.
28191 2013-02-18 Richard Biener <rguenther@suse.de>
28193 PR middle-end/56349
28194 * cfghooks.c (merge_blocks): If we merge a latch into another
28195 block adjust references to it.
28196 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
28197 (verify_loop_structure): Verify that a recorded latch is in fact
28200 2013-02-18 Richard Biener <rguenther@suse.de>
28202 PR tree-optimization/56321
28203 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
28204 order SSA name release and virtual operand unlinking.
28206 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28208 * config/microblaze/microblaze.md (save_stack_block): Define.
28209 (restore_stack_block): Likewise.
28211 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28213 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
28214 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
28215 * config/microblaze/microblaze.c (microblaze_option_override):
28216 Bail out early for PIC modes when target does not support PIC.
28218 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28220 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
28221 Replace with a microblaze version.
28222 (microblaze_trampoline_init): Adapt for microblaze.
28223 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
28226 2013-02-16 Jakub Jelinek <jakub@redhat.com>
28227 Dodji Seketeli <dodji@redhat.com>
28230 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
28231 (instrument_mem_region_access): Do not forget to always put
28232 instrumentation of the of 'base' and 'base + len' in a "if (len !=
28233 0) statement, even for cases where either 'base' or 'base + len'
28234 are not instrumented -- because they have been previously
28235 instrumented. Simplify the logic by putting all the statements
28236 instrument 'base + len' inside a sequence, and then insert that
28237 sequence right before the current insertion point. Then, to
28238 instrument 'base + len', just get an iterator on that statement.
28239 And do not forget to update the pointer to iterator the function
28240 received as argument.
28242 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
28244 PR rtl-optimization/56348
28245 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
28247 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
28249 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
28250 (clean_graph_dump_file): Pass base to start_graph_dump.
28252 2013-02-14 Richard Henderson <rth@redhat.com>
28255 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
28257 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
28259 * collect2-aix.h: Define F_LOADONLY.
28261 2013-02-14 Richard Biener <rguenther@suse.de>
28264 * varasm.c (output_constant_def_1): Get the decl representing
28265 the constant as argument.
28266 (output_constant_def): Wrap output_constant_def_1.
28267 (make_decl_rtl): Use output_constant_def_1 with the decl
28268 representing the constant.
28269 (build_constant_desc): Optionally re-use a decl already
28270 representing the constant.
28271 (tree_output_constant_def): Adjust.
28273 2013-02-14 Dodji Seketeli <dodji@redhat.com>
28276 * asan.c (instrument_builtin_call): Really put the length of the
28277 second source argument into src1_len.
28279 2013-02-13 Jakub Jelinek <jakub@redhat.com>
28281 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
28282 argument. If it is false, don't create edge from then_bb to
28284 (insert_if_then_before_iter): Pass true to it.
28285 (build_check_stmt): Pass false to it.
28286 (transform_statements): Flush hash table only on extended basic
28287 block boundaries, rather than at the beginning of every bb.
28288 Don't flush hash table on nonfreeing_call_p calls.
28289 * tree-flow.h (nonfreeing_call_p): New prototype.
28290 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
28292 2013-02-13 David S. Miller <davem@davemloft.net>
28294 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
28296 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
28299 * ira.c (max_regno_before_ira): Move from ...
28301 (fix_reg_equiv_init): Use max_regno_before_ira instead of
28304 2013-02-13 Jakub Jelinek <jakub@redhat.com>
28306 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
28308 2013-02-13 Richard Biener <rguenther@suse.de>
28311 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
28312 globals in MEM_REFs.
28314 2013-02-13 Richard Biener <rguenther@suse.de>
28316 * loop-init.c (loop_optimizer_init): Clear loop state when
28317 re-initializing preserved loops.
28318 * loop-unswitch.c (unswitch_single_loop): Return whether
28319 we unswitched the loop. Do not verify loop state here.
28320 (unswitch_loops): When we unswitched a loop discover new loops.
28322 2013-02-13 Kostya Serebryany <kcc@google.com>
28324 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
28326 * sanitizer.def: Rename __asan_init to __asan_init_v1.
28328 2013-02-12 Dodji Seketeli <dodji@redhat.com>
28330 Avoid instrumenting duplicated memory access in the same basic block
28331 * Makefile.in (asan.o): Add new dependency on hash-table.h
28332 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
28333 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
28334 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
28335 (free_mem_ref_resources, has_mem_ref_been_instrumented)
28336 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
28337 (get_mem_ref_of_assignment): New functions.
28338 (get_mem_refs_of_builtin_call): Extract from
28339 instrument_builtin_call and tweak a little bit to make it fit with
28341 (instrument_builtin_call): Use the new
28342 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
28343 of is_gimple_builtin_call.
28344 (instrument_derefs, instrument_mem_region_access): Insert the
28345 instrumented memory reference into the hash table.
28346 (maybe_instrument_assignment): Renamed instrument_assignment into
28347 this, and change it to advance the iterator when instrumentation
28348 actually happened and return true in that case. This makes it
28349 homogeneous with maybe_instrument_assignment, and thus give a
28350 chance to callers to be more 'regular'.
28351 (transform_statements): Clear the memory reference hash table
28352 whenever we enter a new BB, when we cross a function call, or when
28353 we are done transforming statements. Use
28354 maybe_instrument_assignment instead of instrumentation. No more
28355 need to special case maybe_instrument_assignment and advance the
28356 iterator after calling it; it's now handled just like
28357 maybe_instrument_call. Update comment.
28359 2013-02-13 Richard Biener <rguenther@suse.de>
28361 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
28362 Fix loop discovery code.
28364 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
28366 PR inline-asm/56148
28367 * lra-constraints.c (process_alt_operands): Match early clobber
28368 operand with itself. Check conflicts with earlyclobber only if
28369 the operand is not reloaded. Prefer to reload conflicting operand
28370 if earlyclobber and matching operands are the same.
28372 2013-02-12 Richard Biener <rguenther@suse.de>
28375 * lto-streamer-out.c (write_symbol): Do not output symbols
28376 for hard register variables.
28378 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
28381 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
28382 (umulsidi3_insn, mulsidi3_insn): New insns.
28384 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
28386 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
28387 (struct tune_params): Add vec_costs field.
28388 * config/arm/arm.c (arm_builtin_vectorization_cost)
28389 (arm_add_stmt_cost): New functions.
28390 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
28391 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
28392 (arm_default_vec_cost): New struct of type cpu_vec_costs.
28393 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
28394 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
28395 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
28396 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
28398 2013-02-12 Richard Biener <rguenther@suse.de>
28401 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
28402 decls again if possible.
28404 2013-02-12 Richard Biener <rguenther@suse.de>
28406 PR middle-end/56288
28407 * tree-ssa.c (verify_ssa_name): Fix check, move
28408 SSA_NAME_IN_FREE_LIST check up.
28410 2013-02-12 Jakub Jelinek <jakub@redhat.com>
28411 Steven Bosscher <steven@gcc.gnu.org>
28413 PR rtl-optimization/56151
28414 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
28415 equal to op0 or op1, and last_insn pattern is CODE operation
28416 with MEM dest and one of the operands matches that MEM.
28418 2013-02-11 Sriraman Tallam <tmsriram@google.com>
28420 * doc/extend.texi: Document Function Multiversioning and "default"
28421 parameter string to target attribute.
28422 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
28423 target attribute parameter is "default".
28424 (ix86_compare_version_priority): Remove checks for target attribute.
28425 (ix86_mangle_function_version_assembler_name): Change error to sorry.
28426 Remove check for target attribute equal to NULL. Add assert.
28427 (ix86_generate_version_dispatcher_body): Change error to sorry.
28429 2013-02-11 Iain Sandoe <iain@codesourcery.com>
28430 Jack Howarth <howarth@bromo.med.uc.edu>
28431 Patrick Marlier <patrick.marlier@gmail.com>
28434 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
28435 define ENDFILE_SPEC as TM_DESTRUCTOR.
28436 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
28438 2013-02-11 Alexander Potapenko <glider@google.com>
28439 Jack Howarth <howarth@bromo.med.uc.edu>
28440 Jakub Jelinek <jakub@redhat.com>
28443 * config/darwin.c (cdtor_record): Rename ctor_record.
28444 (sort_cdtor_records): Rename sort_ctor_records.
28445 (finalize_dtors): New routine to sort destructors by
28446 priority before use in assemble_integer.
28447 (machopic_asm_out_destructor): Use finalize_dtors if needed.
28449 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
28451 PR rtl-optimization/56275
28452 * simplify-rtx.c (avoid_constant_pool_reference): Check that
28453 offset is non-negative and less than cmode size before
28454 calling simplify_subreg.
28456 2013-02-11 Richard Biener <rguenther@suse.de>
28458 PR tree-optimization/56264
28459 * cfgloop.h (fix_loop_structure): Adjust prototype.
28460 * loop-init.c (fix_loop_structure): Return the number of
28461 newly discovered loops.
28462 * tree-cfgcleanup.c (repair_loop_structures): When new loops
28463 are discovered, do a full loop-closed SSA rewrite.
28465 2013-02-11 Richard Biener <rguenther@suse.de>
28467 PR tree-optimization/56273
28468 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
28470 (check_array_ref): Fix missing newline in dumps.
28471 (search_for_addr_array): Likewise.
28473 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
28475 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
28477 2013-02-09 Jakub Jelinek <jakub@redhat.com>
28480 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
28482 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
28484 PR rtl-optimization/56246
28485 * lra-constraints.c (simplify_operand_subreg): Try to reuse
28487 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
28488 constraints are satisfied.
28490 2013-02-08 Jeff Law <law@redhat.com>
28493 * emit-rtl.c (reg_is_parm_p): New function.
28494 * regs.h (reg_is_parm_p): New prototype.
28495 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
28496 callee-clobbered registers.
28498 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
28501 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
28502 If there is no implicit builtin declaration, just return NULL.
28504 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28506 * config/i386/sse.md (FMAMODEM): New mode iterator.
28507 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
28508 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
28510 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28512 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
28513 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
28514 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
28516 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28518 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
28519 (microblaze*-*-elf): Likewise.
28520 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
28522 * config/microblaze/microblaze-c.c: Add builtin defines for
28523 _LITTLE_ENDIAN and _BIG_ENDIAN.
28524 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
28525 add to TARGET_DEFAULT flags.
28526 Expand ASM_SPEC and LINK_SPEC.
28527 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
28528 * config/microblaze/microblaze.md: Update extendsidi2 and
28529 movdi_internal instructions to use low-order / high-order reg
28531 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
28532 options and inversemask / mask of LITTLE_ENDIAN.
28533 * config/microblaze/t-microblaze: Expand multilib options to
28534 include mlittle-endian (le) and update exceptions patterns.
28536 2013-02-08 Jakub Jelinek <jakub@redhat.com>
28538 PR rtl-optimization/56195
28539 * lra-constraints.c (get_reload_reg): Don't reuse regs
28540 if they have smaller mode than requested, if they have
28541 wider mode than requested, try to return a SUBREG.
28543 PR tree-optimization/56250
28544 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
28545 if type is unsigned and code isn't MULT_EXPR.
28547 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28549 PR tree-optimization/56064
28550 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
28551 bits according to mode.
28552 * fixed-value.h (fixed_from_double_int)
28553 (const_fixed_from_double_int): Adjust comments.
28555 2013-02-08 Richard Biener <rguenther@suse.de>
28558 * lto-streamer.h (struct data_in): Remove current_file, current_line
28559 and current_col members.
28560 * lto-streamer-out.c (lto_output_location): Stream changed bits
28561 en-block for efficiency.
28562 * lto-streamer-in.c (clear_line_info): Remove.
28563 (lto_input_location): Cache current file, line and column
28564 globally via local statics. Read changed bits en-block.
28565 (input_function): Do not call clear_line_info.
28566 (lto_read_body): Likewise.
28567 (lto_input_toplevel_asms): Likewise.
28569 2013-02-08 Michael Matz <matz@suse.de>
28571 PR tree-optimization/52448
28572 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
28573 (nt_call_phase): New static.
28574 (add_or_mark_expr): Only mark accesses with newer phase than any
28576 (nonfreeing_call_p): New.
28577 (nt_init_block): Update nt_call_phase, mark blocks as visited.
28578 (nt_fini_block): Keep blocks marked as visited.
28579 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
28581 2013-02-08 Richard Biener <rguenther@suse.de>
28583 * ira.c (ira): Free broken dominator information.
28585 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28587 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
28589 2013-02-08 Marek Polacek <polacek@redhat.com>
28591 * cfgloop.c (verify_loop_structure): Add more checking of headers.
28593 2013-02-08 Richard Biener <rguenther@suse.de>
28595 PR middle-end/56181
28596 * cfgloop.h (flow_loops_find): Adjust.
28597 (bb_loop_header_p): Declare.
28598 * cfgloop.c (bb_loop_header_p): New function split out from ...
28599 (flow_loops_find): ... here. Adjust function signature,
28600 support incremental loop structure update.
28601 (verify_loop_structure): Cleanup. Verify a loop is a loop.
28602 * cfgloopmanip.c (fix_loop_structure): Move ...
28603 * loop-init.c (fix_loop_structure): ... here.
28604 (apply_loop_flags): Split out from ...
28605 (loop_optimizer_init): ... here.
28606 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
28607 in incremental mode, only remove dead loops here.
28609 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28612 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
28613 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
28614 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
28615 (*round<mode>3.libgcc): New insns for fixed-modes.
28616 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
28617 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
28618 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
28619 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
28620 implementations. Define to __builtin_avr_absFX,
28621 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
28622 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
28623 __builtin_avr_countlsFX, respectively.
28624 * config/avr/avr-c.c (target.h): Include it.
28625 (enum avr_builtin_id): New enum.
28626 (avr_resolve_overloaded_builtin): New static function.
28627 (avr_register_target_pragmas): Use it to set
28628 targetm.resolve_overloaded_builtin.
28629 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
28630 tree nodes used by DEF_BUILTIN.
28631 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
28632 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
28633 <AVR_BUILTIN_xxBITS>: Same.
28635 2013-02-08 Richard Biener <rguenther@suse.de>
28637 * cfgloop.c (verify_loop_structure): Properly handle
28638 a loop exiting to another loop header.
28639 * ira-int.h (ira_loops): Remove.
28640 * ira.c (ira_loops): Remove.
28641 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
28642 (do_reload): Use loop_optimizer_finalize.
28643 * ira-build.c (create_loop_tree_nodes): Use get_loops and
28644 number_of_loops to access the loop tree.
28645 (more_one_region_p): Likewise.
28646 (finish_loop_tree_nodes): Likewise.
28647 (rebuild_regno_allocno_maps): Likewise.
28648 (mark_loops_for_removal): Likewise.
28649 (mark_all_loops_for_removal): Likewise.
28650 (remove_unnecessary_regions): Likewise.
28651 (ira_build): Likewise.
28652 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
28654 2013-02-08 Richard Biener <rguenther@suse.de>
28656 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
28657 * ipa-pure-const.c (analyze_function): Avoid calling
28658 mark_irreducible_loops twice.
28659 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
28661 2013-02-07 David S. Miller <davem@davemloft.net>
28663 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
28665 * var-tracking.c (vt_add_function_parameter): Test the presence of
28666 HAVE_window_save properly and do not remap argument registers when
28667 we have a leaf function.
28669 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
28672 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
28674 * config/i386/i386.c (ix86_print_operand): Ditto.
28676 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28678 * lra-constraints.c (process_alt_operands): Fix recently added comment.
28680 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28682 PR rtl-optimization/56225
28683 * lra-constraints.c (process_alt_operands): Check that reload hard
28684 reg can hold value for strict_low_part.
28686 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28689 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
28690 dwarf2out_end_function.
28691 (in_first_function_p, maybe_at_text_label_p,
28692 first_loclabel_num_not_at_text_label): New variables.
28693 (dwarf2out_var_location): In the first function find out
28694 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
28695 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
28698 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
28700 PR rtl-optimization/56178
28701 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
28702 SUBREG of a register. Tidy up related block of code.
28703 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
28704 note if the source is a register or a SUBREG of a register.
28706 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28709 * config/rs6000/rs6000.md (ptrm): New mode attr.
28710 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
28711 call_value_indirect_aix<pttrsize>,
28712 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
28715 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28717 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
28718 if -bnortl. Convert to strcmp and strncmp.
28720 2013-02-07 Alan Modra <amodra@gmail.com>
28723 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
28724 addresses won't wrap when offsetting.
28725 (rs6000_secondary_reload): Provide secondary reloads needed for
28726 wrapping LO_SUM addresses.
28728 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
28730 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
28731 MACH, just __MACH__.
28733 2013-02-06 Richard Biener <rguenther@suse.de>
28735 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
28736 instead of calling fix_loop_structure.
28738 2013-02-06 Jakub Jelinek <jakub@redhat.com>
28740 PR middle-end/56217
28741 * omp-low.c (use_pointer_for_field): Return false if
28742 lower_send_shared_vars doesn't generate any copy-out code.
28744 2013-02-06 Tom de Vries <tom@codesourcery.com>
28746 PR rtl-optimization/56131
28747 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
28748 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
28749 of the label is NULL. Add comment.
28751 2013-02-05 Jakub Jelinek <jakub@redhat.com>
28753 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
28756 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
28757 (STATIC_LIBTSAN_LIBS): Likewise.
28758 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
28759 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
28760 is defined, don't add anything else beyond that.
28761 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
28762 (LINK_COMMAND_SPEC): Use them.
28764 PR tree-optimization/56205
28765 * tree-stdarg.c (check_all_va_list_escapes): Return true if
28766 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
28767 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
28769 2013-02-05 Richard Biener <rguenther@suse.de>
28771 PR tree-optimization/53342
28772 PR tree-optimization/53185
28773 * tree-vectorizer.h (vect_check_strided_load): Remove.
28774 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
28775 not disallow peeling for vectorized strided loads.
28776 (vect_check_strided_load): Make static and simplify.
28777 (vect_analyze_data_refs): Adjust.
28778 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
28779 correctly when vectorizing strided loads.
28781 2013-02-05 Richard Biener <rguenther@suse.de>
28783 * doc/install.texi: Refer to ISL, not PPL.
28785 2013-02-05 Jan Hubicka <jh@suse.cz>
28787 PR tree-optimization/55789
28788 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
28790 2013-02-05 Jan Hubicka <jh@suse.cz>
28792 PR tree-optimization/55789
28793 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
28794 the dead call anyway.
28796 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
28799 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
28801 2013-02-04 Alexander Potapenko <glider@google.com>
28802 Jack Howarth <howarth@bromo.med.uc.edu>
28803 Jakub Jelinek <jakub@redhat.com>
28806 * config/darwin.c (sort_ctor_records): Stabilized qsort
28807 on constructor priority by using original position.
28808 (finalize_ctors): New routine to sort constructors by
28809 priority before use in assemble_integer.
28810 (machopic_asm_out_constructor): Use finalize_ctors if needed.
28812 2013-02-04 Jakub Jelinek <jakub@redhat.com>
28815 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
28816 about visibility on artificial decls.
28817 * config/sol2.c (solaris_assemble_visibility): Likewise.
28819 2013-02-04 Kai Tietz <ktietz@redhat.com>
28822 * config/i386/i386.c (function_value_ms_64): Add additional valtype
28823 argument and improve checking of return-argument types for 16-byte
28825 (ix86_function_value_1): Add additional valtype argument on call
28826 of function_value_64.
28827 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
28828 handling infunction_value_64 function.
28830 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
28832 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
28834 2013-02-04 Richard Biener <rguenther@suse.de>
28836 PR tree-optimization/56188
28837 * tree-ssa-structalias.c (label_visit): Consider case with
28838 initially non-empty points-to set.
28839 (perform_var_substitution): Dump node mapping and clean up.
28841 2013-02-04 Richard Guenther <rguenther@suse.de>
28844 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
28845 node prevail as last resort.
28846 (lto_symtab_merge_decls): Remove guard on LTRANS here.
28847 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
28849 2013-02-04 Richard Biener <rguenther@suse.de>
28851 PR tree-optimization/56113
28852 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
28854 (equiv_class_lookup_or_add): ... this.
28855 (label_visit): Adjust and fix error in previous patch.
28856 (perform_var_substitution): Adjust.
28858 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28860 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28861 * config/sh/sh4-300.md: Likewise.
28862 * config/sh/sh4a.md: Likewise.
28863 * config/sh/constraints.md: Likewise.
28864 * config/sh/sh.md: Likewise.
28865 * config/sh/netbsd-elf.h: Likewise.
28866 * config/sh/predicates.md: Likewise.
28867 * config/sh/sh-protos.h: Likewise.
28868 * config/sh/ushmedia.h: Likewise.
28869 * config/sh/linux.h: Likewise.
28870 * config/sh/sh.c: Likewise.
28871 * config/sh/superh.h: Likewise.
28872 * config/sh/elf.h: Likewise.
28873 * config/sh/sh4.md: Likewise.
28874 * config/sh/sh.h: Likewise.
28876 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28878 * config/pa/constraints.md: Adjust unused letters. Change "T"
28879 constraint to match_test floating_point_store_memory_operand().
28880 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28881 (base14_operand): New.
28882 (floating_point_store_memory_operand): New.
28883 (integer_store_memory_operand): Revise to use base14_operand and
28884 reg_plus_base_memory_operand.
28885 (move_dest_operand): Allow symbolic_memory_operands.
28886 (symbolic_memory_operand): Check for LO_SOM.
28887 (symbolic_operand): Change default case to break.
28888 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28889 CONST_DOUBLE values to be reloaded by putting them into memory when
28890 the destination is a floating point register.
28891 (movdf): Remove code to handle CONST_DOUBLE.
28893 (reload_indf_r1): New.
28894 (reload_insf_r1): New.
28895 Consistently use "Q" and "T" constraints with integer and floating
28896 point move instructions, respectively.
28897 (movdi): Remove FAIL.
28898 Change predicate for source operand unamed DImode move from
28899 general_operand to move_src_operand.
28900 (umulsidi3): Change predicate for destination operand to
28902 Likewise for similar unamed patterns.
28903 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28904 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28905 (hppa_legitimize_address): Simplify mask calculation.
28906 (pa_emit_move_sequence): Revised handling of secondary reloads from
28907 REG+D addresses for floating point loads and stores. Directly handle
28908 loading CONST0_RTX (mode) to a floating point register.
28909 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28910 to floating point registers. Don't restrict secondary reloads to
28911 floating point registers to integer modes. Revise some comments and
28913 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28914 (pa_legitimate_address_p): New.
28915 (pa_legitimize_reload_address): New.
28916 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28917 (STRICT_REG_OK_FOR_BASE_P): New.
28918 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28919 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28921 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28922 Andrew Dixie <andrewd@gentrack.com>
28924 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28927 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28929 * expmed.c (extract_bit_field_1): Pass the full width of the
28930 structure to get_best_reg_extraction_insn.
28932 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28935 * configure.ac (use_cxa_atexit): Add AIX.
28936 * configure: Regenerate.
28938 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28940 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28943 * final.c (need_profile_function): New variable.
28944 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28945 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28946 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28947 notes, targetm.asm_out.function_prologue doesn't emit anything,
28948 HAVE_prologue and profiler should be emitted before prologue,
28949 set need_profile_function instead of emitting it.
28950 (final_scan_insn): If need_profile_function, emit
28951 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28952 NOTE_INSN_FUNCTION_BEG note.
28954 2013-02-01 Richard Henderson <rth@redhat.com>
28956 * config/rs6000/rs6000.md (smulditi3): New.
28959 * config/alpha/alpha.md (umulditi3): New.
28961 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28963 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28964 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28966 2013-02-01 Richard Biener <rguenther@suse.de>
28968 PR tree-optimization/56113
28969 * tree-ssa-structalias.c (label_visit): Reduce work for
28970 single-predecessor nodes.
28972 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28974 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28975 range isn't testing for zero.
28977 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28979 PR middle-end/56113
28980 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28982 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28983 Nick Clifton <nickc@redhat.com>
28985 * config/v850/constraints.md (Q): Define as a memory constraint.
28986 * config/v850/predicates.md (label_ref_operand): New predicate.
28987 (e3v5_shift_operand): New predicate.
28988 (ior_operator): New predicate.
28989 * config/v850/t-v850: Add e3v5 multilib.
28990 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28991 (v850_gen_movdi): Prototype.
28992 * config/v850/v850.c: Add support for e3v5 architecture.
28993 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28995 (construct_save_jarl): Add e3v5 long JARL support.
28996 (v850_adjust_insn_length): New function. Adjust length of call
28997 insns when using e3v5 instructions.
28998 (v850_gen_movdi): New function: Generate instructions to move a
29000 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
29001 (CPP_SPEC): Define __v850e3v5__ as appropriate.
29002 (TARGET_USE_FPU): Enable for e3v5.
29003 (CONST_OK_FOR_W): New macro.
29004 (ADJUST_INSN_LENGTH): Define.
29005 * config/v850/v850.md (UNSPEC_LOOP): Define.
29006 (attr cpu): Add v850e3v5.
29007 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
29008 (movdi): New pattern.
29009 (movdi_internal): New pattern.
29010 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
29011 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
29012 (cstoresf4): Likewise.
29013 (cstoredf4): Likewise.
29014 (insv): New pattern.
29015 (rotlso3_a): New pattern.
29016 (rotlsi3_b): New pattern
29017 (rotlsi3_v850e3v5): New pattern.
29018 (doloop_begin): New pattern.
29019 (fix_loop_counter): New pattern.
29020 (doloop_end): New pattern.
29021 (branch_normal): Add e3v5 long branch support.
29022 (branch_invert): Likewise.
29023 (branch_z_normal): Likewise.
29024 (branch_z_invert): Likewise.
29025 (branch_nz_normal): Likewise.
29026 (branch_nz_invert): Likewise.
29027 (call_internal_short): Add e3v5 register-indirect JARL support.
29028 (call_internal_long): Likewise.
29029 (call_value_internal_short): Likewise.
29030 (call_value_internal_long): Likewise.
29031 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
29032 (mloop): New option.
29033 * config.gcc: Add support for configuring v840e3v5 target.
29034 * doc/invoke.texi: Document new v850 specific command line options.
29036 2013-01-31 Paul Koning <ni1d@arrl.net>
29040 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
29041 children if parent is a class.
29042 (prune_unused_types_prune): Don't add DW_AT_declaration.
29044 2013-01-31 Richard Biener <rguenther@suse.de>
29046 PR tree-optimization/56157
29047 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
29048 match up operand with SLP child.
29050 2013-01-31 Jason Merrill <jason@redhat.com>
29053 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
29054 parameters the first time.
29055 (gen_scheduled_generic_parms_dies): Check completeness here.
29057 2013-01-31 Richard Biener <rguenther@suse.de>
29059 PR middle-end/53073
29060 * common.opt (faggressive-loop-optimizations): New flag,
29061 enabled by default.
29062 * doc/invoke.texi (faggressive-loop-optimizations): Document.
29063 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
29064 infer_loop_bounds_from_undefined by it.
29066 2013-01-31 Richard Biener <rguenther@suse.de>
29068 PR tree-optimization/56150
29069 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
29070 visit virtual operands.
29071 (find_uses_to_rename_bb): Likewise.
29073 2013-01-31 Richard Biener <rguenther@suse.de>
29075 PR tree-optimization/56150
29076 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
29077 mixed store non-store stmts.
29079 2013-01-30 Jakub Jelinek <jakub@redhat.com>
29082 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
29083 LIBASAN_EARLY_SPEC is defined.
29084 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
29085 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
29087 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
29090 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
29091 invalid args instead of ICEing on it.
29092 (ix86_valid_target_attribute_tree): Return error_mark_node if
29093 ix86_valid_target_attribute_inner_p failed.
29094 (ix86_valid_target_attribute_p): Return false only if
29095 ix86_valid_target_attribute_tree returned error_mark_node. Allow
29096 target("default") attribute.
29097 (sorted_attr_string): Change argument from const char * to tree,
29098 merge in all target attribute arguments rather than just one.
29099 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
29100 instead of free. Avoid using strcat.
29101 (ix86_mangle_function_version_assembler_name): Mangle
29102 target("default") as if no target attribute is present. Adjust
29103 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
29104 instead of xmalloc and XDELETEVEC instead of free.
29105 (ix86_function_versions): Don't return true if one of the decls
29106 doesn't have target attribute. If they don't and one of the decls
29107 is DECL_FUNCTION_VERSIONED, report an error. Adjust
29108 sorted_attr_string caller. Use XDELETEVEC instead of free.
29109 (ix86_supports_function_versions): Remove.
29110 (make_name): Fix up formatting.
29111 (make_dispatcher_decl): Remove resolver_name and its initialization.
29112 Avoid leaking memory.
29113 (is_function_default_version): Return true if there is
29114 target("default") attribute rather than no target attribute at all.
29115 (make_resolver_func): Avoid leaking memory.
29116 (ix86_generate_version_dispatcher_body): Likewise.
29117 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
29118 * target.def (supports_function_versions): Remove.
29119 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
29120 * doc/tm.texi: Regenerated.
29122 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
29124 PR rtl-optimization/56144
29125 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
29126 for values with side effects.
29128 2013-01-30 Richard Biener <rguenther@suse.de>
29130 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
29131 (sparseset_pop): Likewise.
29132 * cfganal.c (compute_idf): Likewise. Increase work-stack size
29133 to be able to use quick_push in the worker loop.
29135 2013-01-30 Marek Polacek <polacek@redhat.com>
29137 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
29139 2013-01-30 Richard Biener <rguenther@suse.de>
29142 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
29144 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
29146 PR tree-optimization/56064
29147 * fixed-value.c (fixed_from_double_int): New function.
29148 * fixed-value.h (fixed_from_double_int): New prototype.
29149 (const_fixed_from_double_int): New static inline function.
29150 * fold-const.c (native_interpret_fixed): New static function.
29151 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
29152 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
29153 (native_encode_fixed): New static function.
29154 (native_encode_expr) <FIXED_CST>: Use it.
29155 (native_interpret_int): Move double_int worker code to...
29156 * double-int.c (double_int::from_buffer): ...this new static method.
29157 * double-int.h (double_int::from_buffer): Prototype it.
29159 2013-01-30 Richard Biener <rguenther@suse.de>
29161 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
29162 New pointer-map and obstack.
29163 (init_alias_vars): Allocate pointer-map and obstack.
29164 (delete_points_to_sets): Free them.
29165 (find_what_var_points_to): Cache result.
29166 (find_what_p_points_to): Adjust for changed interface of
29167 find_what_var_points_to.
29168 (compute_points_to_sets): Likewise.
29169 (ipa_pta_execute): Likewise.
29171 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
29173 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
29174 * configure: Regenerate.
29175 * config.in: Regenerate.
29176 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
29177 #nobits/#progbits if supported.
29179 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
29182 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
29183 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
29184 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
29186 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29188 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
29189 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
29191 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29193 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
29194 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
29196 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29198 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
29200 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
29201 * config/arm/cortex-a7.md: New bypasses using
29202 arm_mac_accumulator_is_result.
29204 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29206 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
29207 (cortex_a7_neon_mla): Likewise.
29208 (cortex_a7_fpfmad): New reservation.
29209 (cortex_a7_fpmacs): Use ffmas and update required units.
29210 (cortex_a7_fpmuld): Update required units and latency.
29211 (cortex_a7_fpmacd): Likewise.
29212 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
29213 (cortex_a7_neon). Likewise.
29214 (bypass) Update participating units.
29216 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29218 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
29219 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
29221 * config/arm/vfp11.md (vfp_farith): Use ffmas.
29222 (vfp_fmul): Use ffmad.
29223 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
29224 (cortex_r4_fmacd): Use ffmad.
29225 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
29226 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
29227 (cortex_a9_fmacd): Use ffmad.
29228 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
29229 (cortex_a8_vfp_macd): Use ffmad.
29230 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
29231 (cortex_a5_fpmacd): Use ffmad.
29232 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
29233 (cortex_a15_vfp_macd): Use ffmad.
29234 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29236 2013-01-29 Jason Merrill <jason@redhat.com>
29239 * varasm.c (default_assemble_visibility): Don't warn about
29240 visibility on artificial decls.
29242 2013-01-29 Richard Biener <rguenther@suse.de>
29244 PR tree-optimization/56113
29245 * tree-ssa-structalias.c (equiv_class_lookup): Also return
29247 (label_visit): Free duplicate bitmaps and record the leader instead.
29248 (perform_var_substitution): Adjust.
29250 2013-01-29 Richard Biener <rguenther@suse.de>
29252 PR tree-optimization/55270
29253 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
29254 the CFG, schedule loops for fixup.
29256 2013-01-29 Nick Clifton <nickc@redhat.com>
29258 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
29261 2013-01-28 Leif Ekblad <leif@rdos.net>
29263 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
29264 * config/i386/i386.h (TARGET_RDOS): New macro.
29265 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
29266 * config/i386/i386.c (ix86_option_override_internal): For 64bit
29267 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
29268 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
29269 DEFAULT_LARGE_SECTION_THRESHOLD.
29270 * config/i386/i386.md (R14_REG, R15_REG): New constants.
29271 * config/i386/rdos.h: New file.
29272 * config/i386/rdos64.h: New file.
29274 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
29277 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
29280 2013-01-28 Jakub Jelinek <jakub@redhat.com>
29282 PR rtl-optimization/56117
29283 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
29284 call cselib_lookup_from_insn on the MEM before calling
29285 add_insn_mem_dependence.
29287 2013-01-28 Richard Biener <rguenther@suse.de>
29289 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
29290 to a stmt that didn't have one.
29291 (copy_phis_for_bb): Likewise for PHI arguments.
29292 (copy_debug_stmt): Likewise for debug stmts.
29294 2013-01-28 Richard Biener <rguenther@suse.de>
29296 PR tree-optimization/56034
29297 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
29298 (partition_builtin_p): Adjust.
29299 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
29300 it is the last partition.
29301 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
29302 up the vertex for the definition.
29303 (classify_partition): Classify whether a partition is a
29304 PKIND_REDUCTION, thus has uses outside of the loop.
29305 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
29306 Merge all PKIND_REDUCTION partitions into the last partition.
29307 (tree_loop_distribution): Seed partitions from reductions as well.
29309 2013-01-28 Jakub Jelinek <jakub@redhat.com>
29311 PR tree-optimization/56125
29312 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
29313 pow(x,c) into sqrt(x) * powi(x, n/2) or
29314 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
29315 optimizing for size.
29316 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
29317 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
29320 PR tree-optimization/56094
29321 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
29322 to UNKNOWN_LOCATION while gimplifying expr.
29324 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
29327 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
29328 operand 0 in movabs insn template for -masm=intel asm alternative.
29329 (*movabs<mode>_2): Ditto for operand 1.
29331 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
29334 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
29337 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29339 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
29342 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29344 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
29345 Undef to avoid warning.
29347 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
29349 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
29350 * configure: Regenerate.
29352 2013-01-25 Jakub Jelinek <jakub@redhat.com>
29354 PR tree-optimization/56098
29355 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
29356 for stmts with volatile ops.
29357 (cond_store_replacement): Don't optimize if assign has volatile ops.
29358 (cond_if_else_store_replacement_1): Don't optimize if either
29359 then_assign or else_assign have volatile ops.
29360 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
29363 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29365 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
29367 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29369 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
29370 missing ':' in asm example.
29372 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
29374 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
29375 entries into lane and laneq entries.
29376 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
29377 Remove AdvSIMD scalar modes.
29378 (aarch64_sq<r>dmulh_laneq<mode>): New.
29379 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
29381 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
29382 builtin implementations to relfect changes in RTL in aarch64-simd.md.
29383 * config/aarch64/iterators.md (VCOND): New.
29386 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29389 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
29390 Add NULL LIBNAME argument to existing definitions.
29391 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
29392 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
29393 * config/avr/avr.c (DEF_BUILTIN): Same.
29394 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
29395 (avr_expand_builtin): Expand to a vanilla call if a libgcc
29396 implementation is available (DECL_ASSEMBLER_NAME is set).
29397 (avr_fold_absfx): New static function.
29398 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
29399 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
29400 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
29401 AVR_BUILTIN_ABSLLK.
29402 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
29403 (abshk, absk, abslk, absllk): Provide as static inline functions.
29405 2013-01-25 Marek Polacek <polacek@redhat.com>
29407 PR tree-optimization/56035
29408 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
29410 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
29412 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
29413 (*movtf_internal_rex64): Add (!o,C) alternative
29414 (*movxf_internal_rex64): Ditto.
29415 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
29417 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
29419 * doc/invoke.texi: fix typo.
29420 * doc/objc.texi: fix typo.
29422 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
29424 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
29425 for the first two alternatives.
29427 2013-01-24 Diego Novillo <dnovillo@google.com>
29429 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
29430 (ggc-zone.o): Remove.
29431 * configure.ac: Remove option --with-gc.
29432 * configure: Re-generate.
29433 * doc/install.texi: Remove documentation for --with-gc.
29434 * gengtype.c (write_enum_defn): Remove. Update all users.
29435 (write_Types_process_field): Remove generation of gt_e_* argument.
29436 (output_type_enum): Remove. Update all users.
29437 (write_enum_defn): Remove. Update all users.
29438 (enum alloc_zone): Remove. Update all users.
29439 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
29440 * ggc-common.c (ggc_splay_alloc): Remove first argument.
29441 Update all callers.
29442 (struct ptr_data): Remove field TYPE. Update all users.
29443 (gt_pch_note_object): Remove argument TYPE. Update all users.
29444 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
29446 * ggc-none.c (ggc_alloc_typed_stat): Remove.
29447 (struct alloc_zone): Remove.
29448 (ggc_internal_alloc_zone_stat): Remove.
29449 (ggc_internal_cleared_alloc_zone_stat): Remove.
29450 * ggc-page.c (ggc_alloc_typed_stat): Remove.
29451 (ggc_pch_count_object): Remove last argument. Update all users.
29452 (ggc_pch_alloc_object): Remove last argument. Update all users.
29453 (struct alloc_zone): Remove.
29454 * ggc-zone.c: Remove.
29455 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
29456 (struct alloc_zone): Remove.
29457 (ggc_alloc_typed_stat): Remove.
29458 (ggc_alloc_typed): Remove.
29459 (ggc_splay_alloc): Remove first argument.
29460 (rtl_zone): Remove. Update all users.
29461 (tree_zone): Remove. Update all users.
29462 (tree_id_zone): Remove. Update all users.
29463 (ggc_internal_zone_alloc_stat): Remove. Update all users.
29464 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
29465 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
29466 * tree-ssanames.c: Remove references to zone allocator in comments.
29468 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
29470 * config/avr/avr.c (avr_out_fract): Make register numbers that
29471 might be outside of source operand signed.
29473 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
29475 * config/i386/constraints.md (Yf): New constraint.
29476 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
29477 of f constraint to conditionaly disable x87 register preferences.
29478 (*movdf_internal): Ditto.
29479 (*movsf_internal): Ditto.
29481 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
29483 PR inline-asm/55934
29484 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
29485 that have operands with impossible constraints.
29486 Add a FIXME for a speed-up opportunity.
29487 * lra-constraints.c (process_alt_operands): Verify that a class
29488 selected from constraints on asms is valid for the operand mode.
29489 (curr_insn_transform): Remove incorrect comment.
29491 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
29493 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
29494 TOC operand is a valid symbol ref in the constant pool.
29496 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29498 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
29500 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
29503 * config/avr/stdfix.h: New file.
29504 * t-avr (stdfix-gcc.h): New rule to build it.
29505 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
29507 2013-01-23 Kostya Serebryany <kcc@google.com>
29509 * config/darwin.h: remove dependency on
29510 CoreFoundation (asan on Mac OS).
29512 2013-01-23 Jakub Jelinek <jakub@redhat.com>
29515 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
29516 instead of cmpdi_operand for first comparison operand.
29517 Don't assert that comparison operands aren't both constants.
29519 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29521 * doc/install.texi (Downloading the Source): Update references to
29522 downloading separate components.
29524 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29526 * doc/extend.texi (__int128): Improve grammar.
29528 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
29531 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
29532 alternative to (o,r).
29533 (*movdi_internal_rex64): Remove (!o,n) alternative.
29534 (DImode immediate->memory splitter): Remove.
29535 (DImode immediate->memory peephole2): Remove.
29536 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
29537 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
29538 alternative to (!o,*r).
29539 (*movtf_internal_sse): New pattern.
29540 (*movxf_internal_rex64): New pattern.
29541 (*movxf_internal): Disable for TARGET_64BIT.
29542 (*movdf_internal_rex64): Remove (!o,F) alternative.
29544 2013-01-22 Jakub Jelinek <jakub@redhat.com>
29546 PR middle-end/56074
29547 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
29548 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
29549 * tree-vect-loop-manip.c (find_loop_location): Also ignore
29550 stmt locations where LOCATION_LOCUS of the stmt location is
29551 UNKNOWN_LOCATION or BUILTINS_LOCATION.
29554 * config/i386/i386.md (UNSPEC_STOS): New.
29555 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
29556 *strsetqi_1): Add UNSPEC_STOS.
29558 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
29561 * doc/invoke.texi: Remove left over -Wsynth example.
29563 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29565 PR tree-optimization/56051
29566 * fold-const.c (fold_binary_loc): Don't fold
29567 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
29568 a narrowing conversion, or widening conversion from signed
29571 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29573 PR rtl-optimization/56023
29574 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
29575 dependent on debug instruction.
29577 2013-01-21 Martin Jambor <mjambor@suse.cz>
29579 PR middle-end/56022
29580 * function.c (allocate_struct_function): Call
29581 invoke_set_current_function_hook earlier.
29583 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29585 * reload1.c (init_reload): Only initialize reload_obstack
29586 during the first call.
29588 2013-01-21 Marek Polacek <polacek@redhat.com>
29590 * cfgloop.c (verify_loop_structure): Fix up grammar.
29592 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
29594 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
29595 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
29597 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29600 * config/arm/marvell-pj4.md: Update copyright year.
29601 Fix up use of alu to alu_reg and simple_alu_imm.
29603 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29605 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
29607 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
29610 * lra-constraints.c (curr_insn_transform): Don't reuse original
29611 insn for secondary memory move when memory mode should be different.
29613 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29615 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
29616 atomic_storedi_1): New patterns.
29618 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29620 btver2 pipeline descriptions.
29621 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
29623 * config/i386/i386.md (btver2_decode): New type attributes.
29624 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
29626 * config/i386/btver2.md: New file describing btver2 pipelines.
29628 2013-01-19 Andrew Pinski <apinski@cavium.com>
29630 PR tree-optimization/52631
29631 * tree-ssa-sccvn (visit_use): Before looking up the original
29632 statement, try looking up the simplified expression.
29634 2013-01-19 Anthony Green <green@moxielogic.com>
29636 * config/moxie/moxie.c (moxie_expand_prologue): Set
29637 current_function_static_stack_size.
29639 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29641 PR tree-optimization/56029
29642 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
29643 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
29645 2013-01-18 Sharad Singhai <singhai@google.com>
29647 PR tree-optimization/55995
29648 * dumpfile.c (dump_loc): Print location only if available.
29649 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
29651 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
29654 * lra-constraints.c (curr_insn_transform): Reuse original insn for
29655 secondary memory move.
29656 (inherit_reload_reg): Use rclass instead of cl for
29657 check_secondary_memory_needed_p.
29659 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29661 PR middle-end/56015
29662 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
29663 the case where writing real complex part of target modifies op1.
29665 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
29667 * config/aarch64/aarch64-simd.md
29668 (aarch64_vcond_internal<mode>): Handle unordered cases.
29669 * config/aarch64/iterators.md (v_cmp_result): New.
29671 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
29672 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29674 * config/arm/marvell-pj4.md: New file.
29675 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
29676 * config/arm/arm.md (generic_sched): Add marvell_pj4.
29677 (generic_vfp): Likewise.
29678 * config/arm/arm-cores.def: Add marvell-pj4.
29679 * config/arm/arm-tune.md: Regenerate.
29680 * config/arm/arm-tables.opt: Regenerate.
29681 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
29682 * doc/invoke.texi: Document marvell-pj4.
29684 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
29686 * config/aarch64/arm_neon.h: Map scalar types to standard types.
29688 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29693 * var-tracking.c (negative_power_of_two_p): New.
29694 (global_get_addr_cache, local_get_addr_cache): New.
29695 (get_addr_from_global_cache, get_addr_from_local_cache): New.
29696 (vt_canonicalize_addr): Rewrite using the above. Adjust the
29698 (vt_stack_offset_p): Remove.
29699 (vt_canon_true_dep): Always canonicalize loc's address.
29700 (clobber_overlapping_mems): Make sure we have a MEM.
29701 (local_get_addr_clear_given_value): New.
29702 (val_reset): Clear local cached entries.
29703 (compute_bb_dataflow): Create and release the local cache.
29704 Disable duplicate MEMs clobbering.
29705 (emit_notes_in_bb): Clobber MEMs likewise.
29706 (vt_emit_notes): Create and release the local cache.
29707 (vt_initialize, vt_finalize): Create and release the global
29708 cache, respectively.
29709 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
29711 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29713 PR libmudflap/53359
29714 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
29715 not found in the symtab.
29717 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29720 PR rtl-optimization/55547
29721 PR rtl-optimization/53827
29724 * alias.c (offset_overlap_p): New, factored out of...
29725 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
29726 the conservative special case for symbolic constants. Don't
29727 adjust zero sizes on alignment.
29729 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
29731 PR rtl-optimization/52573
29732 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
29733 REG_UNUSED for the same register.
29735 2013-01-17 Richard Biener <rguenther@suse.de>
29736 Marek Polacek <polacek@redhat.com>
29738 PR rtl-optimization/55833
29739 * loop-unswitch.c (unswitch_loops): Move loop verification...
29740 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
29741 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
29742 Set it to true when we're removing a loop from hierarchy tree in
29743 an irreducible region.
29744 (fix_bb_placements): Adjust caller.
29745 (fix_loop_placements): Likewise.
29747 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
29749 * config/avr/builtins.def (DEF_BUILTIN): Factor out
29750 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
29751 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
29752 Remove ID. Adjust comments.
29753 * config/avr/avr-c.c (avr_builtin_name): Remove.
29754 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
29755 * config/avr/avr.c (avr_tolower): New static function.
29756 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
29757 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
29758 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
29761 2013-01-17 Jan Hubicka <jh@suse.cz>
29763 PR tree-optimization/55273
29764 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
29766 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
29769 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
29770 store through atomic_store<mode>_1.
29771 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
29773 2013-01-17 Martin Jambor <mjambor@suse.cz>
29775 PR tree-optimizations/55264
29776 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
29777 for virtual methods.
29778 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
29779 virtual methods before inlining is over.
29780 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
29782 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
29785 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
29787 PR rtl-optimization/56005
29788 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
29789 pending reads for prefetch.
29791 2013-01-16 Ian Bolton <ian.bolton@arm.com>
29793 * config/aarch64/aarch64.md
29794 (*cstoresi_neg_uxtw): New pattern.
29795 (*cmovsi_insn_uxtw): New pattern.
29796 (*<optab>si3_uxtw): New pattern.
29797 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
29798 (*<optab>si3_insn_uxtw): New pattern.
29799 (*bswapsi2_uxtw): New pattern.
29801 2013-01-16 Richard Biener <rguenther@suse.de>
29803 * tree-inline.c (tree_function_versioning): Remove set but
29804 never used variable.
29806 2013-01-16 Richard Biener <rguenther@suse.de>
29808 PR tree-optimization/55964
29809 * tree-flow.h (rename_variables_in_loop): Remove.
29810 (rename_variables_in_bb): Likewise.
29811 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
29812 (copy_loop_before): Adjust and delete update-ssa status.
29813 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
29814 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
29815 (rename_variables_in_loop): Remove.
29816 (slpeel_update_phis_for_duplicate_loop): Likewise.
29817 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
29818 use available cfg machinery instead of duplicating it.
29819 Update PHI nodes and perform poor-mans SSA update here.
29820 (slpeel_tree_peel_loop_to_edge): Adjust.
29822 2013-01-16 Richard Biener <rguenther@suse.de>
29824 PR tree-optimization/54767
29825 PR tree-optimization/53465
29826 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
29827 (vrp_visit_phi_node): For PHI arguments coming via backedges
29828 drop all symbolical range information.
29829 (execute_vrp): Compute backedges.
29831 2013-01-16 Richard Biener <rguenther@suse.de>
29833 * doc/install.texi: Update CLooG and ISL requirements to
29836 2013-01-16 Christian Bruel <christian.bruel@st.com>
29839 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
29840 (broken_move): Handle UNSPECV_SP_SWITCH_B.
29841 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
29843 2013-01-16 DJ Delorie <dj@redhat.com>
29845 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
29846 (UNSPECV_SP_SWITCH_E): New.
29847 (sp_switch_1): Change to an unspec.
29848 (sp_switch_2): Change to an unspec. Don't use post-inc when we
29851 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
29853 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29854 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29855 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29856 (expand_mem_thread_fence): Ditto.
29857 (expand_mem_signal_fence): Ditto.
29858 (expand_atomic_load): Ditto.
29859 (expand_atomic_store): Ditto.
29861 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29863 PR rtl-optimization/55547
29864 PR rtl-optimization/53827
29867 * alias.c (memrefs_conflict_p): Set sizes to negative after
29870 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29873 * function.c (thread_prologue_and_epilogue_insns): Always
29874 add crtl->drap_reg to set_up_by_prologue.set, even if
29875 stack_realign_drap is false.
29877 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29879 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29880 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29881 *call): Fix indention.
29883 2013-01-15 Tom de Vries <tom@codesourcery.com>
29886 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29889 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29891 PR rtl-optimization/55153
29892 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29894 2013-01-15 Martin Jambor <mjambor@suse.cz>
29896 PR tree-optimization/55920
29897 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29898 accesses as grp_to_be_debug_replaced.
29900 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29902 PR tree-optimization/55920
29903 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29904 there is non-useless type conversion needed from debug rhs to lhs,
29905 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29907 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29908 Mikael Pettersson <mikpe@it.uu.se>
29911 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29913 (ASM_OUTPUT_CASE_LABEL): Remove.
29914 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29915 * final.c (shorten_branches): Update alignment of labels before
29916 jump tables if CASE_VECTOR_SHORTEN_MODE.
29918 2013-01-15 Richard Biener <rguenther@suse.de>
29921 * system.h: Do not include gmp.h for building host tools.
29923 2013-01-15 Richard Biener <rguenther@suse.de>
29925 PR middle-end/55882
29926 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29927 account for bitpos when computing alignment.
29929 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29931 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29932 (ix86_target_macros_internal): Likewise.
29934 * config/i386/i386.c (m_CORE2I7): Removed.
29935 (m_CORE_HASWELL): New macro.
29936 (m_CORE_ALL): Likewise.
29937 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29938 (initial_ix86_arch_features): Likewise.
29939 (processor_target_table): Initializations for Core avx2.
29940 (cpu_names): New names "core-avx2".
29941 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29942 PROCESSOR_CORE_HASWELL.
29943 (ix86_issue_rate): New case.
29944 (ia32_multipass_dfa_lookahead): Likewise.
29945 (ix86_sched_init_global): Likewise.
29947 * config/i386/i386.h (TARGET_HASWELL): New macro.
29948 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29949 (processor_type): New PROCESSOR_HASWELL.
29951 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29953 PR tree-optimization/55955
29954 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29955 *SHIFT_EXPR and *ROTATE_EXPR codes.
29957 PR tree-optimization/48766
29958 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29959 -ftrapv disable -fwrapv.
29961 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29964 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29965 etc. to 1 and not to __flash.
29966 Use LL suffix for __INT24_MAX__ with -mint8.
29967 Use ULL suffix for __UINT24_MAX__ with -mint8.
29969 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29971 * config/avr/avr-arch.h
29972 (struct base_arch_s): Use typedef avr_arch_t instead.
29973 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29974 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29975 * config/avr/avr.c: Same.
29976 * config/avr/avr-devices.c: Same.
29977 * config/avr/driver-avr.c: Same.
29978 * config/avr/gen-avr-mmcu-texi.c: Same.
29979 * config/avr/avr-mcus.def: Adjust comment.
29981 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29983 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29984 * config/aarch64/iterators.md (VALLDI): New.
29986 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29987 Andi Kleen <ak@linux.intel.com>
29990 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29991 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29994 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29996 * config/avr/avr-stdint.h: Remove trailing blanks.
29997 * config/avr/avr-log.h: Same.
29998 * config/avr/avr-arch.h: Same.
29999 * config/avr/avr-devices.c: Same.
30000 * config/avr/avr-dimode.md: Same.
30001 * config/avr/predicates.md: Same.
30002 * config/avr/avr-c.c: Same. And fix typo.
30004 * config/avr/avr-protos.h: Same. And:
30005 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
30006 (init_cumulative_args): Rename to avr_init_cumulative_args.
30007 (expand_prologue): Rename to avr_expand_prologue.
30008 (expand_epilogue): Rename to avr_expand_epilogue.
30009 (adjust_insn_length): Rename to avr_adjust_insn_length.
30010 (notice_update_cc): Rename to avr_notice_update_cc.
30011 (final_prescan_insn): Rename to avr_final_prescan_insn.
30012 * config/avr/avr.c: Same.
30013 * config/avr/avr.h: Same.
30014 * config/avr/avr.md: Remove trailing blanks.
30015 (prologue): Use avr_expand_prologue.
30016 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
30018 2013-01-14 Richard Biener <rguenther@suse.de>
30020 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
30021 verify_location, collect_subblocks): New functions.
30022 (verify_gimple_in_cfg): Verify that locations only reference
30023 BLOCKs in the functions BLOCK tree.
30025 2013-01-14 Richard Biener <rguenther@suse.de>
30027 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
30029 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
30031 (insert_out_of_ssa_copy_on_edge): Likewise.
30032 (rewrite_close_phi_out_of_ssa): Likewise.
30033 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
30035 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
30036 propagated constants.
30037 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
30040 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
30042 * config/avr/avr-modes.def: Add GPL copyright notice.
30044 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
30046 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
30047 MEMMODEL_MASK to determine memory model.
30048 (atomic_store<mode>): Ditto from operands[2].
30049 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
30051 2013-01-13 Jakub Jelinek <jakub@redhat.com>
30054 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
30055 (fold_gimple_assign): Don't call unshare_expr here.
30056 (fold_ctor_reference): Call unshare_expr.
30058 2013-01-13 Terry Guo <terry.guo@arm.com>
30060 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
30061 * doc/fragments.texi: Document MULTILIB_REUSE.
30062 * gcc.c (multilib_reuse): New internal spec.
30063 (set_multilib_dir): Also search multilib from multilib_reuse.
30064 * genmultilib (tmpmultilib3): Refactor code.
30065 (tmpmultilib4): Ditto.
30066 (multilib_reuse): New multilib argument.
30068 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
30070 * Makefile.in: Update copyright.
30072 2013-01-12 Tom de Vries <tom@codesourcery.com>
30074 PR middle-end/55890
30075 * calls.c (expand_call): Check if arg_nr is valid.
30077 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
30079 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
30080 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
30081 documentation. Add missing '__' in front of
30082 __builtin_ia32_packssdw256.
30084 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
30087 * config/s390/s390.c (s390_preferred_reload_class): Do not return
30088 NO_REGS for larl operands.
30089 (s390_reload_larl_operand): Use s390_load_address instead of
30092 2013-01-11 Richard Biener <rguenther@suse.de>
30094 * tree-cfg.c (verify_node_sharing_1): Split out from ...
30095 (verify_node_sharing): ... here.
30096 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
30098 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
30100 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
30101 Substitute TREECHECKING.
30102 * configure: Regenerate.
30103 * Makefile.in (TREECHECKING): New.
30105 2013-01-11 Richard Guenther <rguenther@suse.de>
30107 PR tree-optimization/44061
30108 * tree-vrp.c (extract_range_basic): Compute zero as
30109 value-range for __builtin_constant_p of function parameters.
30111 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
30113 Update copyright years.
30115 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
30117 PR rtl-optimization/55672
30118 * lra-eliminations.c (mark_not_eliminable): Permit addition with
30119 const to be eliminable.
30121 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
30123 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
30124 * configure: Regenerate.
30126 2013-01-10 Richard Biener <rguenther@suse.de>
30128 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
30130 2013-01-10 Richard Biener <rguenther@suse.de>
30133 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
30134 locations for virtual PHI arguments.
30135 (rewrite_update_phi_arguments): Likewise.
30137 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
30139 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
30142 2013-01-10 Jakub Jelinek <jakub@redhat.com>
30144 PR tree-optimization/55921
30145 * tree-complex.c (expand_complex_asm): New function.
30146 (expand_complex_operations_1): Call it for GIMPLE_ASM.
30148 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
30151 * config/s390/s390.c (s390_symref_operand_p)
30152 (s390_loadrelative_operand_p): Merge the two functions.
30153 (s390_check_qrst_address, print_operand_address): Add parameters
30154 to s390_loadrelative_operand_p invokation.
30155 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
30156 (s390_reload_larl_operand, s390_secondary_reload): Use
30157 s390_loadrelative_operand_p instead of s390_symref_operand_p.
30158 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
30160 2013-01-09 Mike Stump <mikestump@comcast.net>
30162 * dse.c (record_store): Remove unnecessary assert.
30164 2013-01-09 Jan Hubicka <jh@suse.cz>
30166 PR tree-optimization/55569
30167 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
30168 * cfgloop.h (scale_loop_profile): Likewise.
30170 2013-01-09 Jan Hubicka <jh@suse.cz>
30173 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
30175 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
30177 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
30179 PR middle-end/55114
30180 * expr.h (maybe_emit_group_store): Declare.
30181 * expr.c (maybe_emit_group_store): New function.
30182 * builtins.c (expand_builtin_int_roundingfn): Call it.
30183 (expand_builtin_int_roundingfn_2): Likewise.
30185 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
30187 PR rtl-optimization/55829
30188 * lra-constraints.c (match_reload): Add code for absent output.
30189 (curr_insn_transform): Add code for reloads of matched inputs
30192 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
30194 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
30195 attribute of movddup insn to DF.
30196 (*vec_interleave_lowv2df): Ditto.
30197 (vec_dupv2df): Ditto.
30199 2013-01-09 Jan Hubicka <jh@suse.cz>
30201 PR tree-optimiation/55875
30202 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
30203 EVERY_ITERATION parameter.
30204 (number_of_iterations_exit): Check if exit is executed every iteration.
30205 (idx_infer_loop_bounds): Similarly here.
30206 (n_of_executions_at_most): Simplify
30207 to only test for cases where statement is dominated by the
30208 particular bound; handle correctly the "postdominance" test.
30209 (scev_probably_wraps_p): Use max loop iterations info
30210 as a global bound first.
30212 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
30213 Nick Clifton <nickc@redhat.com>
30215 * config/v850/v850.md (cbranchsf4): New pattern.
30216 (cstoresf4): New pattern.
30217 (cbranchdf4): New pattern.
30218 (cstoredf4): New pattern.
30219 (movsicc): Disallow floating point comparisons.
30220 (cmpsf_le_insn): Fix order of operators.
30221 (cmpsf_lt_insn): Likewise.
30222 (cmpsf_eq_insn): Likewise.
30223 (cmpdf_le_insn): Likewise.
30224 (cmpdf_lt_insn): Likewise.
30225 (cmpdf_eq_insn): Likewise.
30226 (cmpsf_ge_insn): Use LE comparison.
30227 (cmpdf_ge_insn): Likewise.
30228 (cmpsf_gt_insn): Use LT comparison.
30229 (cmpdf_gt_insn): Likewise.
30230 (cmpsf_ne_insn): Delete pattern.
30231 (cmpdf_ne_insn): Delete pattern.
30232 * config/v850/v850.c (v850_gen_float_compare): Use
30233 gen_cmpdf_eq_insn for NE comparison.
30234 (v850_float_z_comparison_operator)
30235 (v850_float_nz_comparison_operator): Move from here ...
30236 * config/v850/predicates.md: ... to here. Move GT and GE
30237 comparisons into v850_float_z_comparison_operator.
30238 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
30240 (v850_float_nz_comparison_operator): Likewise.
30242 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30244 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
30245 with calls to gen_insvsi/gen_insvdi.
30247 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
30249 * config/i386/i386.c (initial_ix86_tune_features): Set up
30250 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
30252 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
30253 Jakub Jelinek <jakub@redhat.com>
30255 PR tree-optimization/48189
30256 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
30257 If nitercst is 0, don't predict the exit edge.
30259 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
30261 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
30262 in asm_fprintf with reg_names.
30263 (aarch64_print_operand_address): Likewise.
30264 (aarch64_return_addr): Likewise.
30265 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
30267 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30269 * config/pa/pa.h (VAL_U6_BITS_P): Define.
30270 (INT_U6_BITS): Likewise.
30271 * config/pa/predicates.md (uint6_operand): New predicate.
30272 (shift5_operand, shift6_operand): Likewise.
30273 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
30275 (lshrdi3): Use shift6_operand.
30276 (shrpsi4, shrpdi4): New insn patterns.
30277 (extzv): Delete expander.
30278 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
30279 predicates in unamed zero extract patterns. Tighten common constraint.
30280 (extv): Delete expander.
30281 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
30282 predicates in unamed sign extract patterns. Tighten common constraint.
30283 (insv): Delete expander.
30284 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
30285 predicates in unamed insert patterns. Tighten common constraint.
30286 Change uint32_operand predicate to uint6_operand predicate in unamed
30287 DImode pattern to insert constant values of type 1...1xxxx.
30289 2013-01-04 Jan Hubicka <jh@suse.cz>
30291 PR tree-optimization/55823
30292 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
30295 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30296 Uros Bizjak <ubizjak@gmail.com>
30298 PR rtl-optimization/55845
30299 * df-problems.c (can_move_insns_across): Stop scanning at
30300 volatile_insn_p source instruction or give up if
30301 across_from .. across_to range contains any volatile_insn_p
30304 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
30306 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
30307 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
30309 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
30310 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
30312 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30315 * asan.c (asan_clear_shadow): New function.
30316 (asan_emit_stack_protection): Use it.
30318 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
30320 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
30321 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
30322 with tab instead of space.
30324 2013-01-08 Nick Clifton <nickc@redhat.com>
30326 * config/rl78/rl78.c (rl78_expand_prologue): Always select
30327 register bank 0 at the start of an interrupt handler.
30328 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
30331 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
30333 * config/aarch64/aarch64-simd.md
30334 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
30335 (aarch64_simd_bsl): Likewise.
30336 (aarch64_vcond_internal<mode>): Likewise.
30337 (vcond<mode><mode>): Likewise.
30338 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
30339 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
30341 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
30343 * config/aarch64/aarch64-builtins.c
30344 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
30346 2013-01-08 Martin Jambor <mjambor@suse.cz>
30349 * tree-sra.c (analyze_access_subtree): Return true also after
30350 potentially creating a debug-only replacement.
30352 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30354 PR middle-end/55890
30355 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
30357 PR tree-optimization/54120
30358 * tree-vrp.c (range_fits_type_p): Don't allow
30359 src_precision < precision from signed vr to unsigned_p
30360 if vr->min or vr->max is negative.
30361 (simplify_float_conversion_using_ranges): Test can_float_p
30362 against CODE_FOR_nothing.
30364 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30365 Richard Biener <rguenther@suse.de>
30367 PR middle-end/55851
30368 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
30369 types instead of just INTEGER_TYPE types.
30371 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
30373 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
30376 2013-01-07 Steve Ellcey <sellcey@mips.com>
30379 * config/mips/mips.opt: Change mad to mmad to match documentation.
30381 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30384 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
30385 .progmemx.data now.
30387 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30390 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
30391 (avr_addrspace_t): Add .section_name field.
30392 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
30394 (avr_addrspace): Same. Initialize .section_name. Remove last
30395 NULL entry. Put __memx into .progmemx.data.
30396 (progmem_section_prefix): Remove.
30397 (avr_asm_init_sections): No need to initialize progmem_section.
30398 (avr_asm_named_section): Use avr_addrspace[].section_name to get
30399 section name prefix.
30400 (avr_asm_select_section): Ditto. And use get_unnamed_section to
30401 retrieve the progmem section.
30402 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
30403 boundary to run over avr_addrspace[].
30404 (avr_register_target_pragmas): Ditto.
30406 2013-01-06 Jakub Jelinek <jakub@redhat.com>
30408 * varasm.c (output_constant_def_contents): For asan_protect_global
30409 protected strings, adjust DECL_ALIGN if needed, before testing for
30411 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
30412 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
30414 (output_object_block): For asan protected decls, emit asan padding
30415 after their contents.
30416 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
30417 (asan_finish_file): Test it here instead.
30419 2013-01-07 Nick Clifton <nickc@redhat.com>
30420 Matthias Klose <doko@debian.org>
30421 Doug Kwan <dougkwan@google.com>
30422 H.J. Lu <hongjiu.lu@intel.com>
30425 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
30427 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
30429 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
30431 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
30433 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
30435 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30438 * doc/install.texi (Cross-Compiler-Specific Options): Document
30441 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
30443 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
30444 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
30445 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
30446 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
30447 vqmovun_high_s64): Fix source operand number and update copyright.
30449 2013-01-07 Richard Biener <rguenther@suse.de>
30451 PR middle-end/55890
30452 * gimple.h (gimple_call_builtin_p): New overload.
30453 * gimple.c (validate_call): New function.
30454 (gimple_call_builtin_p): Likewise.
30455 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
30456 Use gimple_call_builtin_p.
30457 (find_func_clobbers): Likewise.
30458 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
30459 (strlen_optimize_stmt): Likewise.
30461 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30463 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
30464 (vld1q_dup_*): Likewise.
30465 (vld1_*): Likewise.
30466 (vld1q_*): Likewise.
30467 (vld1_lane_*): Likewise.
30468 (vld1q_lane_*): Likewise.
30470 2013-01-07 Richard Biener <rguenther@suse.de>
30472 * lto-streamer.h (LTO_minor_version): Bump to 2.
30474 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30476 * config/aarch64/aarch64-protos.h
30477 (aarch64_const_double_zero_rtx_p): Rename to...
30478 (aarch64_float_const_zero_rtx_p): ...this.
30479 (aarch64_float_const_representable_p): New.
30480 (aarch64_output_simd_mov_immediate): Likewise.
30481 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
30482 move immediate case.
30483 * config/aarch64/aarch64.c
30484 (aarch64_const_double_zero_rtx_p): Rename to...
30485 (aarch64_float_const_zero_rtx_p): ...this.
30486 (aarch64_print_operand): Allow printing of new constants.
30487 (aarch64_valid_floating_const): New.
30488 (aarch64_legitimate_constant_p): Check for valid floating-point
30490 (aarch64_simd_valid_immediate): Likewise.
30491 (aarch64_vect_float_const_representable_p): New.
30492 (aarch64_float_const_representable_p): Likewise.
30493 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
30494 (aarch64_output_simd_mov_immediate): New.
30495 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
30496 (*movdf_aarch64): Likewise.
30497 * config/aarch64/constraints.md (Ufc): New.
30498 (Y): call aarch64_float_const_zero_rtx.
30499 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
30501 2013-01-07 Richard Biener <rguenther@suse.de>
30503 PR tree-optimization/55888
30504 PR tree-optimization/55862
30505 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
30506 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
30507 not if it is contained therein.
30509 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30511 * config/avr/t-avr: Typo.
30513 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30516 * config/avr/t-avr: Don't automatically rebuild
30517 $(srcdir)/config/avr/t-multilib
30518 $(srcdir)/config/avr/avr-tables.opt
30519 $(srcdir)/doc/avr-mmcu.texi
30520 (avr-mcus): New phony target to build them on request.
30521 (s-avr-mlib, s-avr-mmcu-texi): Remove.
30522 * avr/avr-mcus.def: Adjust comments.
30524 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
30526 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
30528 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30530 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
30532 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30534 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
30536 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
30538 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
30539 to generate profiling.
30540 * config/rs6000/aix64.h (LIB_SPEC): Same.
30542 2013-01-04 Andrew Pinski <apinski@cavium.com>
30544 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
30546 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
30548 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
30550 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
30552 (ix86_expand_move): Ditto.
30553 (ix86_zero_extend_to_Pmode): Ditto.
30554 (ix86_expand_call): Ditto.
30555 (ix86_expand_special_args_builtin): Ditto.
30556 (ix86_expand_builtin): Ditto.
30558 2013-01-04 Richard Biener <rguenther@suse.de>
30560 PR tree-optimization/55862
30561 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
30562 translating them through PHI nodes.
30564 2013-01-04 Martin Jambor <mjambor@suse.cz>
30566 PR tree-optimization/55755
30567 * tree-sra.c (sra_modify_assign): Do not check that an access has no
30568 children when trying to avoid producing a VIEW_CONVERT_EXPR.
30570 2013-01-04 Marek Polacek <polacek@redhat.com>
30572 PR middle-end/55859
30573 * opts.c (default_options_optimization): Clarify error message.
30575 2013-01-04 Richard Biener <rguenther@suse.de>
30577 PR middle-end/55863
30578 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
30581 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30584 * config/pa/pa.md (movsi): Revert previous change.
30585 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
30588 2013-01-03 Richard Henderson <rth@redhat.com>
30590 * config/i386/i386.c (ix86_expand_move): Always assign to op1
30591 after eliminating TLS symbols.
30593 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
30596 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
30597 * graphite-poly.c (debug_gmp_value): Likewise.
30599 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
30602 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
30603 selected code model, define __code_mode_small__, __code_model_medium__,
30604 __code_model_large__, __code_model_32__ or __code_model_kernel__.
30605 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
30606 xchg temporary register with %k. Declare temporary register as
30608 [__x86_64__]: For medium and large code models, preserve %rbx register.
30610 2013-01-03 Richard Biener <rguenther@suse.de>
30612 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
30613 (dump_subscript): Adjust.
30614 (finalize_ddr_dependent): Do not dump redundant info.
30615 (analyze_siv_subscript): Adjust.
30616 (subscript_dependence_tester): Likewise.
30617 (compute_affine_dependence): Likewise.
30619 2013-01-03 Richard Biener <rguenther@suse.de>
30622 2013-01-03 Richard Biener <rguenther@suse.de>
30624 PR tree-optimization/55857
30625 * tree-vect-stmts.c (vectorizable_load): Do not setup
30626 re-alignment for invariant loads.
30628 2013-01-02 Richard Biener <rguenther@suse.de>
30630 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30631 invariant load do not generate a vector load from the scalar location.
30633 2013-01-03 Richard Biener <rguenther@suse.de>
30635 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
30636 for not vectorizing.
30637 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
30638 not build INDIRECT_REFs, call get_name once only.
30639 (vect_create_data_ref_ptr): Likewise. Dump base object kind
30640 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
30642 2013-01-03 Richard Biener <rguenther@suse.de>
30644 PR tree-optimization/55857
30645 * tree-vect-stmts.c (vectorizable_load): Do not setup
30646 re-alignment for invariant loads.
30648 2013-01-03 Richard Biener <rguenther@suse.de>
30651 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
30652 prefer a built-in decl.
30654 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30656 * gcc.c (process_command): Update copyright notice dates.
30657 * gcov.c (print_version): Likewise.
30658 * gcov-dump.c (print_version): Likewise.
30660 PR rtl-optimization/55838
30661 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
30662 iv0.step, iv1.step and step.
30664 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30665 Marc Glisse <marc.glisse@inria.fr>
30667 PR tree-optimization/55832
30668 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
30669 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
30670 integer_{one,zero}_node.
30672 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30675 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
30676 * var-tracking.c (reverse_op): Don't add reverse ops to
30677 VALUEs that have already
30678 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
30680 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
30682 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
30684 2013-01-02 Teresa Johnson <tejohnson@google.com>
30686 * dumpfile.c (dump_loc): Print filename with location.
30687 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
30688 new location_t parameter to emit complete unroll message with
30689 new dump framework.
30690 (canonicalize_loop_induction_variables): Compute loops location
30691 and pass to try_unroll_loop_completely.
30692 * loop-unroll.c (report_unroll_peel): New function.
30693 (peel_loops_completely): Use new dump format with location
30694 for main dumpfile message, and invoke report_unroll_peel on success.
30695 (decide_unrolling_and_peeling): Ditto.
30696 (decide_peel_once_rolling): Remove old dumpfile message subsumed
30697 by report_unroll_peel.
30698 (decide_peel_completely): Ditto.
30699 (decide_unroll_constant_iterations): Ditto.
30700 (decide_unroll_runtime_iterations): Ditto.
30701 (decide_peel_simple): Ditto.
30702 (decide_unroll_stupid): Ditto.
30703 * cfgloop.c (get_loop_location): New function.
30704 * cfgloop.h (get_loop_location): Declare.
30706 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30708 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
30711 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30713 PR middle-end/55198
30714 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
30715 BLKmode objects when EXPAND_MEMORY is specified.
30717 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30719 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
30721 (fold_builtin_cpu): Do not share cpu model decls across statements.
30723 2013-01-02 Jason Merrill <jason@redhat.com>
30726 * tree.c (build_array_type_1): Revert earlier change.
30728 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
30730 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
30732 * config/aarch64/aarch64-tune.md: Re-generate.
30734 2013-01-02 Richard Biener <rguenther@suse.de>
30736 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30737 invariant load do not generate a vector load from the scalar location.
30739 2013-01-02 Richard Biener <rguenther@suse.de>
30742 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
30743 * configure: Regenerate.
30745 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30747 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
30748 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
30749 (expand_builtin_int_roundingfn_2): Keep the original target around
30750 for the fallback case.
30752 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30754 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
30755 to be clear for sign changes.
30757 2013-01-01 Jan Hubicka <jh@suse.cz>
30759 * ipa-inline-analysis.c: Fix formatting.
30761 2013-01-01 Jakub Jelinek <jakub@redhat.com>
30763 PR tree-optimization/55831
30764 * tree-vect-loop.c (get_initial_def_for_induction): Use
30765 gsi_after_labels instead of gsi_start_bb.
30767 Copyright (C) 2013 Free Software Foundation, Inc.
30769 Copying and distribution of this file, with or without modification,
30770 are permitted in any medium without royalty provided the copyright
30771 notice and this notice are preserved.