1 2013-11-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4 * config/rs6000/rs6000.md (movdi_internal32): Eliminate
5 constraints that would allow DImode into the traditional Altivec
6 registers, but cause undesirable code generation when loading 0 as
8 (movdi_internal64): Likewise.
9 (cmp<mode>_fpr): Do not use %x for CR register output.
10 (extendsfdf2_fpr): Fix constraints when -mallow-upper-df and
11 -mallow-upper-sf debug switches are used.
13 2013-11-12 Andrew MacLeod <amacleod@redhat.com>
15 * gimple-expr.h (create_tmp_var_name, create_tmp_var_raw,
16 create_tmp_var, create_tmp_reg, mark_addressable, is_gimple_reg_rhs):
17 Relocate prototypes from gimple.h.
18 * gimplify.h: New File. Relocate some prototypes from gimple.h here.
19 (gimple_predicate, enum fallback, enum gimplify_status): Relocate
21 * gimple.h: Move some prototypes to gimplify.h.
22 (gimple_predicate, enum fallback, enum gimplify_status): Move to
24 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location):
25 Relocate from gimpify.c.
26 * gimple-expr.c (remove_suffix, tmp_var_id_num, create_tmp_var_name,
27 create_tmp_var_raw, create_tmp_var, create_tmp_reg, mark_addressable,
28 is_gimple_reg_rhs) Relocate from gimplify.c.
29 * gimplify.c (mark_addressable): Move to gimple-expr.c.
30 (gimple_seq_add_stmt_without_update): Move to gimple.c.
31 (remove_suffix, tmp_var_id_num, create_tmp_var_name, create_tmp_var_raw,
32 create_tmp_var, create_tmp_reg, is_gimple_reg_rhs): Move to
34 (should_carry_location_p): Move to gimple.c.
35 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location): Move
37 (annotate_one_with_location, annotate_all_with_location_after,
38 annotate_all_with_location): Move to gimple.c.
39 (compare_case_labels, sort_case_labels,
40 preprocess_case_label_vec_for_gimple): Move to gimple.c.
41 (rhs_predicate_for): Make static.
42 (gimplify_assign): Relocate from gimple.c.
43 * gimple.c (gimplify_assign): Move to gimplify.c.
44 (gimple_seq_add_stmt_without_update, should_carry_location_p,
45 annotate_one_with_location, annotate_all_with_location_after,
46 annotate_all_with_location, compare_case_labels, sort_case_labels,
47 preprocess_case_label_vec_for_gimple): Relocate from gimplify.c.
48 * tree.h (unshare_expr, unshare_expr_without_location,
49 mark_addressable): Move prototypes to gimplify.h.
50 * Makefile.in (GTFILES): gimple-expr.c now has the GTY tag for
52 * asan.c: Include gimplify.h rather than gimple.h.
53 * cfgloopmanip.c: Likewise.
54 * cgraphunit.c: Likewise.
55 * cilk-common.c: Likewise.
56 * dwarf2out.c: Dont include gimple.h.
57 * fold-const.c: Include gimplify.h rather than gimple.h.
58 * function.c: Likewise.
59 * gimple-fold.c: Likewise.
60 * gimple-ssa-strength-reduction.c: Likewise.
61 * graphite-clast-to-gimple.c: Likewise.
62 * graphite-sese-to-poly.c: Likewise.
63 * ipa-prop.c: Likewise.
64 * ipa-split.c: Likewise.
66 * langhooks.c: Dont include gimple.h.
67 * loop-init.c: Include gimplify.h rather than gimple.h.
68 * omp-low.c: Likewise.
70 * stor-layout.c: Likewise.
71 * targhooks.c: Likewise.
72 * trans-mem.c: Likewise.
73 * tree-affine.c: Likewise.
74 * tree-cfg.c: Likewise.
75 * tree-cfgcleanup.c: Likewise.
76 * tree-complex.c: Likewise.
77 * tree-if-conv.c: Likewise.
78 * tree-inline.c: Likewise.
79 * tree-iterator.c: Likewise.
80 * tree-loop-distribution.c: Likewise.
81 * tree-nested.c: Likewise.
82 * tree-parloops.c: Likewise.
83 * tree-predcom.c: Likewise.
84 * tree-profile.c: Likewise.
85 * tree-scalar-evolution.c: Likewise.
86 * tree-sra.c: Likewise.
87 * tree-ssa-address.c: Likewise.
88 * tree-ssa-ccp.c: Likewise.
89 * tree-ssa-dce.c: Likewise.
90 * tree-ssa-forwprop.c: Likewise.
91 * tree-ssa-ifcombine.c: Likewise.
92 * tree-ssa-loop-im.c: Likewise.
93 * tree-ssa-loop-ivopts.c: Likewise.
94 * tree-ssa-loop-manip.c: Likewise.
95 * tree-ssa-loop-niter.c: Likewise.
96 * tree-ssa-loop-prefetch.c: Likewise.
97 * tree-ssa-loop-unswitch.c: Likewise.
98 * tree-ssa-math-opts.c: Likewise.
99 * tree-ssa-phiopt.c: Likewise.
100 * tree-ssa-phiprop.c: Likewise.
101 * tree-ssa-pre.c: Likewise.
102 * tree-ssa-propagate.c: Likewise.
103 * tree-ssa-reassoc.c: Likewise.
104 * tree-ssa-sccvn.c: Likewise.
105 * tree-ssa-strlen.c: Likewise.
106 * tree-ssa.c: Likewise.
107 * tree-switch-conversio: Likewise.n.c
108 * tree-tailcall.c: Likewise.
109 * tree-vect-data-refs.c: Likewise.
110 * tree-vect-generic.c: Likewise.
111 * tree-vect-loop-manip.c: Likewise.
112 * tree-vect-loop.c: Likewise.
113 * tree-vect-patterns.c: Likewise.
114 * tree-vect-stmts.c: Likewise.
116 * value-prof.c: Likewise.
117 * config/aarch64/aarch64.c: Include gimplify.h instead of gimple.h.
118 * config/alpha/alpha.c: Likewise.
119 * config/darwin.c: Likewise.
120 * config/i386/i386.c: Likewise.
121 * config/ia64/ia64.c: Likewise.
122 * config/mep/mep.c: Likewise.
123 * config/mips/mips.c: Likewise.
124 * config/rs6000/rs6000.c: Likewise.
125 * config/s390/s390.c: Likewise.
126 * config/sh/sh.c: Likewise.
127 * config/sparc/sparc.c: Likewise.
128 * config/spu/spu.c: Likewise.
129 * config/stormy16/stormy16.c: Likewise.
130 * config/tilegx/tilegx.c: Likewise.
131 * config/tilepro/tilepro.c: Likewise.
132 * config/xtensa/xtensa.c: Likewise.
134 2013-11-12 Adam Butcher <adam@jessamine.co.uk>
136 * tree.c (grow_tree_vec_stat): New function ...
137 * tree.h (grow_tree_vec_stat) (grow_tree_vec): ... and its declaration
140 2013-11-12 Marek Polacek <polacek@redhat.com>
142 * final.c (update_alignments): Initialize label to NULL_RTX.
144 2013-11-12 Jeff Law <law@redhat.com>
146 * gimple-ssa-isolate-paths.c (check_loadstore): New function.
147 (insert_trap_and_remove_trailing_statements): New argument OP which
148 is the NULL pointer. Emit the trap after the load/store through
149 the NULL pointer. Simplify the RHS of a store through a NULL pointer
150 when trivial to do so.
151 (isolate_path): Corresponding changes.
152 (gimple_ssa_isolate_erroneous_path): Likewise.
154 2013-11-12 Teresa Johnson <tejohnson@google.com>
155 Jan Hubicka <jh@suse.cz>
157 * predict.c (drop_profile): New function.
158 (handle_missing_profiles): Ditto.
159 (counts_to_freqs): Don't overwrite estimated frequencies
160 when function has no profile counts.
161 * predict.h (handle_missing_profiles): Declare.
162 * tree-inline.c (freqs_to_counts): New function.
163 (copy_cfg_body): Invoke freqs_to_counts as needed.
164 * tree-profile.c (tree_profiling): Invoke handle_missing_profiles.
166 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
169 * config/i386/x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL):
171 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Set for m_HASWELL.
173 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
176 * config/i386/i386.c (ix86_option_override_internal): Check
177 X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL and
178 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL for
179 MASK_AVX256_SPLIT_UNALIGNED_LOAD and
180 MASK_AVX256_SPLIT_UNALIGNED_STORE.
182 * config/i386/x86-tune.def (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL):
183 Clear m_COREI7_AVX and update comments.
184 (X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): Likewise.
186 2013-11-12 Martin Jambor <mjambor@suse.cz>
188 PR rtl-optimization/10474
189 * ira.c (interesting_dest_for_shprep): New function.
190 (split_live_ranges_for_shrink_wrap): Likewise.
191 (find_moveable_pseudos): Move calculation of dominance info,
192 df_analysios and the final anlyses to...
193 (ira): ...here, call split_live_ranges_for_shrink_wrap.
195 2013-11-12 Bin Cheng <bin.cheng@arm.com>
197 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Refactor the code.
198 Handle type conversion.
200 2013-11-11 Martin Liska <marxin.liska@gmail.com>
201 Jan Hubicka <jh@suse.cz>
203 * cgraph.c (dump_cgraph_node): Profile dump added.
204 * cgraph.h (struct cgraph_node): New time profile variable added.
205 * cgraphclones.c (cgraph_clone_node): Time profile is cloned.
206 * gcov-io.h (gcov_type): New profiler type introduced.
207 * ipa-profile.c (lto_output_node): Streaming for time profile added.
208 (input_node): Time profiler is read from LTO stream.
209 * predict.c (maybe_hot_count_p): Hot prediction changed.
210 * profile.c (instrument_values): New case for time profiler added.
211 (compute_value_histograms): Read of time profile.
212 * tree-pretty-print.c (dump_function_header): Time profiler is dumped.
213 * tree-profile.c (init_ic_make_global_vars): Time profiler function added.
214 (gimple_init_edge_profiler): TP function instrumentation.
215 (gimple_gen_time_profiler): New.
216 * value-prof.c (gimple_add_histogram_value): Support for time profiler
218 (dump_histogram_value): TP type added to dumps.
219 (visit_hist): More sensitive check that takes TP into account.
220 (gimple_find_values_to_profile): TP instrumentation.
221 * value-prof.h (hist_type): New histogram type added.
222 (struct histogram_value_t): Pointer to struct function added.
223 * libgcc/Makefile.in: New GCOV merge function for TP added.
224 * libgcov.c: function_counter variable introduced.
225 (_gcov_merge_time_profile): New.
226 (_gcov_time_profiler): New.
228 2013-11-11 Marc Glisse <marc.glisse@inria.fr>
229 Jeff Law <law@redhat.com>
231 * tree-ssa-alias.c (stmt_kills_ref_p_1): Use
232 ao_ref_init_from_ptr_and_size for builtins.
234 2013-11-11 Uros Bizjak <ubizjak@gmail.com>
235 H.J. Lu <hongjiu.lu@intel.com>
238 * config/i386/x86-tune.def
239 (X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
240 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES.
242 (TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
243 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES. Update for renamed
244 X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES.
245 * config/i386/i386.c (ix86_expand_set_or_movmem): Use
246 TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES to calculate
247 misaligned_prologue_used. Check that
248 desired_aling <= epilogue_size_needed.
250 2013-11-11 Cong Hou <congh@google.com>
252 PR tree-optimization/59050
253 * tree-vect-data-refs.c (comp_dr_addr_with_seg_len_pair): Bug fix.
255 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
258 * expmed.c (emit_store_flag): Fail for const-const comparison.
260 2013-11-11 Tristan Gingold <gingold@adacore.com>
261 Eric Botcazou <ebotcazou@adacore.com>
263 * tree.h (CONSTRUCTOR_NO_CLEARING): Define.
264 * tree-core.h (CONSTRUCTOR_NO_CLEARING): Document it.
265 * tree.def (CONSTRUCTOR): Likewise.
266 * doc/generic.texi (CONSTRUCTOR): Likewise. Update description.
267 * gimplify.c (gimplify_init_constructor): Do not clear the object when
268 the constructor is incomplete and CONSTRUCTOR_NO_CLEARING is set.
270 2013-11-11 Basile Starynkevitch <basile@starynkevitch.net>
272 * toplev.c (toplev_main): Move PLUGIN_FINISH invocation before
275 2013-11-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
277 * config/arm/arm.c (arm_new_rtx_costs): Return after handling
280 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
282 * config/arc/arc.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
284 2013-11-08 Jeff Law <law@redhat.com>
286 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
287 threading paths first, then perform PHI node checks if applicable.
289 2013-11-10 Karlson2k <k2k@narod.ru>
290 Kai Tietz <ktietz@redhat.com>
293 * configure.ac: Adding for exported symbols check
294 and for rdynamic-check executable-extension.
295 * configure: Regenerated.
297 2013-11-10 Uros Bizjak <ubizjak@gmail.com>
299 * mode-switching.c (optimize_mode_switching): Mark block as
300 nontransparent, if last_mode at block exit is different from no_mode.
302 2013-11-09 Jan-Benedict Glaw <jbglaw@lug-owl.de>
304 * function.c (NAME__MAIN): Move to...
305 * cfgexpand.c (NAME__MAIN): ...here.
307 2013-11-09 Richard Sandiford <rdsandiford@googlemail.com>
309 * target.def (can_use_doloop_p): New hook.
310 * doc/tm.texi.in (TARGET_CAN_USE_DOLOOP_P): Add.
311 * doc/tm.texi: Regenerate.
312 * doc/md.texi (doloop_begin, doloop_end): Update documentation.
313 * hooks.h (hook_bool_dint_dint_uint_true): Declare.
314 * hooks.c (hook_bool_dint_dint_uint_true): New function.
315 * targhooks.h (can_use_doloop_if_innermost): Declare.
316 * targhooks.c (can_use_doloop_if_innermost): New function.
317 * target.h: Include double-int.h.
318 * loop-doloop.c (doloop_optimize): Call targetm.can_use_doloop_p.
319 Remove iteration count, maximum iteration count, loop depth and
320 enter-at-top inputs from doloop_begin and doloop_end.
321 * config/arc/arc.md (doloop_begin, doloop_end): Update for new
323 * config/arc/arc.c (arc_can_use_doloop_p): New function.
324 (TARGET_CAN_USE_DOLOOP_P): Define.
325 * config/arm/thumb2.md (doloop_end): Update for new interface.
326 * config/arm/arm.c (TARGET_CAN_USE_DOLOOP_P): Define.
327 * config/bfin/bfin.md (doloop_end): Update for new interface.
328 * config/bfin/bfin.c (bfin_can_use_doloop_p): New function.
329 (TARGET_CAN_USE_DOLOOP_P): Define.
330 * config/c6x/c6x.md (doloop_end): Update for new interface.
331 * config/ia64/ia64.md (doloop_end): Update for new interface.
332 * config/ia64/ia64.c (TARGET_CAN_USE_DOLOOP_P): Define.
333 * config/mep/mep.md (doloop_begin, doloop_end): Update for new
335 * config/mep/mep.c (mep_emit_doloop): Likewise.
336 (TARGET_CAN_USE_DOLOOP_P): Define.
337 * config/rs6000/rs6000.md (doloop_end): Update for new interface.
338 * config/rs6000/rs6000.c (TARGET_CAN_USE_DOLOOP_P): Define.
339 * config/s390/s390.md (doloop_end): Update for new interface.
340 * config/sh/sh.md (doloop_end): Likewise.
341 * config/spu/spu.md (doloop_end): Likewise.
342 * config/spu/spu.c (TARGET_CAN_USE_DOLOOP_P): Define.
343 * config/tilegx/tilegx.md (doloop_end): Update for new interface.
344 * config/tilegx/tilegx.c (TARGET_CAN_USE_DOLOOP_P): Define.
345 * config/tilepro/tilepro.md (doloop_end): Update for new interface.
346 * config/tilepro/tilepro.c (TARGET_CAN_USE_DOLOOP_P): Define.
347 * config/v850/v850.md (doloop_begin, doloop_end): Update for new
349 * config/v850/v850.c (TARGET_CAN_USE_DOLOOP_P): Define.
351 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
354 * doc/extend.texi: Move Cilk Plus Builtins node before Other
357 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
358 Joseph Myers <joseph@codesourcery.com>
360 * ginclude/stdatomic.h: New file.
361 * Makefile.in (USER_H): Add stdatomic.h.
363 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
365 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
368 2013-11-08 Jeff Law <law@redhat.com>
370 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
371 * tree-ssa-threadupdate.c (delete_thread_path): New function.
372 (ssa_redirect_edges, thread_block_1): Use it.
373 (thread_through_loop_header, mark_threaded_blocks): Likewise.
374 (thread_through_all_blocks, register_jump_thread): Likewise.
375 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
377 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
379 * config/arm/aarch-common.c
380 (search_term): New typedef.
381 (shift_rtx_costs): New array.
382 (arm_rtx_shift_left_p): New.
383 (arm_find_sub_rtx_with_search_term): Likewise.
384 (arm_find_sub_rtx_with_code): Likewise.
385 (arm_early_load_addr_dep): Add sanity checking.
386 (arm_no_early_alu_shift_dep): Likewise.
387 (arm_no_early_alu_shift_value_dep): Likewise.
388 (arm_no_early_mul_dep): Likewise.
389 (arm_no_early_store_addr_dep): Likewise.
391 2013-11-08 Richard Biener <rguenther@suse.de>
393 PR tree-optimization/59047
394 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses properly.
396 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
398 * common.opt (fcheck-pointer-bounds): Move to ...
399 * c-family/c.opt: ... here.
400 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
401 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
402 * langhooks.h (lang_hooks): Remove chkp_supported field.
403 * toplev.c (process_options): Remove chkp_supported check.
405 2013-11-08 Richard Biener <rguenther@suse.de>
407 PR tree-optimization/59038
408 PR tree-optimization/58955
409 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
410 previous change. Handle known dependences correctly.
412 2013-11-08 Tom de Vries <tom@codesourcery.com>
414 * config/rs6000/t-xilinx: Remove duplicate contents.
416 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
417 Joseph Myers <joseph@codesourcery.com>
419 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
420 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
421 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
422 (struct tree_base): Add atomic_flag field.
423 * tree.h (TYPE_ATOMIC): New accessor macro.
424 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
425 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
426 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
427 (atomicDI_type_node, atomicTI_type_node): New macros for type nodes.
428 * tree.c (set_type_quals): Set TYPE_ATOMIC.
429 (find_atomic_core_type): New function.
430 (build_qualified_type): Adjust alignment for qualified types.
431 (build_atomic_base): New function
432 (build_common_tree_nodes): Build atomicQI_type_node,
433 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
435 * print-tree.c (print_node): Print atomic qualifier.
436 * tree-pretty-print.c (dump_generic_node): Print atomic type attribute.
437 * target.def (atomic_assign_expand_fenv): New hook.
438 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
439 * doc/tm.texi: Regenerate.
440 * targhooks.c (default_atomic_assign_expand_fenv): New function.
441 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
442 * sync-builtins.def (__atomic_feraiseexcept): New built-in function.
443 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
445 * config/i386/i386.c (enum ix86_builtins): Add
446 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
448 (bdesc_special_args): Add __builtin_ia32_fnstenv,
449 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and __builtin_ia32_fnclex.
450 (ix86_expand_builtin): Handle the new built-in functions.
451 (ix86_atomic_assign_expand_fenv): New function.
452 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
453 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
454 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
455 (fnstenv, fldenv, fnstsw, fnclex): New insns.
457 2013-11-07 Steve Ellcey <sellcey@mips.com>
459 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
460 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
461 (MULTILIB_DIRNAMES): Add fp64 directory.
462 (MULTILIB_EXCEPTIONS): Add new exclusions.
464 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
466 * gimplify.c (gimple_regimplify_operands): Do not set
468 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
469 (rewrite_close_phi_out_of_ssa): Same.
470 (rewrite_phi_out_of_ssa): Same.
471 (rewrite_degenerate_phi): Same.
472 (handle_scalar_deps_crossing_scop_limits): Same.
473 * tree-if-conv.c (predicate_scalar_phi): Same.
474 * tree-parloops.c (create_loads_for_reductions): Same.
475 (create_final_loads_for_reduction): Same.
476 (create_loads_and_stores_for_name): Same.
477 (transform_to_exit_first_loop): Same.
478 (create_parallel_loop): Same.
480 (move_computations_dom_walker::before_dom_children): Same.
481 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
482 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Same.
483 * tree-ssa-propagate.c (substitute_and_fold): Same.
484 * tree-vect-loop.c (vect_finalize_reduction): Same.
485 * tree-vect-stmts.c (vectorizable_call): Same.
487 2013-11-07 Mike Stump <mikestump@comcast.net>
489 * config/pdp11/pdp11.c: Include dbxout.h.
490 * config/picochip/picochip.c: Likewise.
492 2013-11-07 Cong Hou <congh@google.com>
494 PR tree-optimization/56764
495 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
496 Combine alias checks if it is possible to amortize the runtime
497 overhead. Return the number of alias checks after merging.
498 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
499 Use the function vect_create_cond_for_alias_checks () to check
500 the number of alias checks.
502 2013-11-07 Jeff Law <law@redhat.com>
504 * varpool.c (ctor_for_folding): Fix typo in comment.
506 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
508 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
509 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
511 2013-11-07 Jeff Law <law@redhat.com>
513 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
515 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
516 No longer check if we have __builtin_trap, assume it's available.
518 2013-11-07 Diego Novillo <dnovillo@google.com>
520 * attribs.c (lookup_scoped_attribute_spec): Make static.
521 (get_attribute_namespace): Likewise.
522 * builtins.c (more_const_call_expr_args_p): Move from tree.h.
523 (validate_arglist): Move earlier in the file. Make static.
524 (expand_stack_restore): Move from stmt.c
525 (expand_stack_save): Move from stmt.c
526 (rewrite_call_expr_array): Move earlier in the file.
527 (rewrite_call_expr_valist): Likewise.
528 * cfgexpand.c: Include hard-reg-set.h before tree.h
531 (expand_asm_loc): Move from stmt.c.
532 (n_occurrences): Move from stmt.c.
533 (check_operand_nalternatives): Move from stmt.c.
534 (tree_conflicts_with_clobbers_p): Move from stmt.c.
535 (expand_asm_operands): Move from stmt.c
536 (expand_asm_stmt): Move from stmt.c
537 (expand_computed_goto): Move from stmt.c
538 (expand_goto): Move from stmt.c
539 (expand_null_return_1): Move from stmt.c
540 (expand_null_return): Move from stmt.c
541 (expand_value_return): Move from stmt.c
542 (expand_return): Move from stmt.c
543 (expand_main_function): Move from function.c
544 (stack_protect_prologue): Move from function.c
545 * cgraphclones.c (build_function_type_skip_args): Move from tree.c.
546 (build_function_decl_skip_args): Move from tree.c.
547 * explow.c (tree_expr_size): Move from tree.c.
548 * expr.c (addr_expr_of_non_mem_decl_p): Remove.
549 (fields_length): Move from tree.c.
550 * fold-const.c (size_low_cst): Move from tree.c.
551 (tree_expr_nonzero_warnv_p): Make static. Move earlier in the file.
552 (tree_expr_nonzero_p): Make static. Move earlier in the file.
553 (fold_build3_initializer_loc): Remove.
554 (tree_invalid_nonnegative_warnv_p): Make static.
555 * function.c (expand_main_function): Move to cfgexpand.c.
556 (stack_protect_prologue): Move to cfgexpand.c.
557 (set_insn_locations): Move earlier in the file.
558 * gimple-fold.c: Include langhooks.h.
559 (truth_type_for): Move from tree.c.
560 * print-tree.c (print_vec_tree): Remove.
561 * stmt.c (expand_computed_goto): Move to cfgexpand.c.
562 (expand_goto): Move to cfgexpand.c.
563 (n_occurrences): Move to cfgexpand.c.
564 (expand_asm_loc): Move to cfgexpand.c
565 (tree_conflicts_with_clobbers_p): Move to cfgexpand.c.
566 (expand_asm_operands): Move to cfgexpand.c.
567 (expand_asm_stmt): Move to cfgexpand.c.
568 (check_operand_nalternatives): Move to cfgexpand.c
569 (expand_null_return): Move to cfgexpand.c.
570 (expand_value_return): Move to cfgexpand.c.
571 (expand_null_return_1): Move to cfgexpand.c.
572 (expand_return): Move to cfgexpand.c.
573 (expand_stack_save): Move to builtins.c.
574 (expand_stack_restore): Move to builtins.c
575 * symtab.c: Include output.h.
576 (decl_assembler_name_hash): Move from tree.c.
577 (decl_assembler_name_equal): Move from tree.c.
578 * trans-mem.c (is_tm_safe_or_pure): Move from tree.h.
579 * tree-eh.c (in_array_bounds_p): Move from tree.c.
580 (range_in_array_bounds_p): Move from tree.c.
581 * tree-object-size.c (fini_object_sizes): Make static.
582 * tree-ssa-dom.c (iterative_hash_exprs_commutative): Move from tree.h.
583 * tree-vrp.c (ssa_name_nonnegative_p): Remove.
584 * tree.c (decl_assembler_name_equal): Move to symtab.c.
585 (tree_expr_size): Move to explow.c.
586 (decl_assembler_name_hash): Move to symtab.c.
588 (tree_expr_size): Move to explow.c.
589 (stabilize_reference_1): Move earlier in the file. Make static.
590 (omp_remove_redundant_declare_simd_attrs): Remove.
591 (simple_cst_list_equal): Move earlier in the file. Make static.
592 (size_low_cst): Move to fold-const.c.
593 (build_type_no_quals): Remove.
594 (build_function_type_skip_args): Move to cgraphclones.c.
595 (build_function_decl_skip_args): Move to cgraphclones.c.
596 (in_array_bounds_p): Move to tree-eh.c.
597 (range_in_array_bounds_p): Move to tree-eh.c.
598 (truth_type_for): Move to gimple-fold.c.
599 (list_equal_p): Remove.
600 * tree.h (decl_assembler_name_equal): Remove.
601 (decl_assembler_name_hash): Remove.
602 (truth_type_for): Remove.
603 (build_type_no_quals): Remove.
604 (build_function_decl_skip_args): Remove.
605 (in_array_bounds_p): Remove.
606 (range_in_array_bounds_p): Remove.
607 (size_low_cst): Remove.
608 (omp_remove_redundant_declare_simd_attrs): Remove.
609 (tree_expr_size): Remove.
610 (fields_length): Remove.
611 (stabilize_reference_1): Remove.
612 (expand_goto): Remove.
613 (expand_stack_save): Remove.
614 (expand_stack_restore): Remove.
615 (expand_return): Remove.
616 (fold_build3_initializer_loc): Remove.
617 (tree_expr_nonzero_p): Remove.
618 (tree_invalid_nonnegative_warnv_p): Remove.
619 (tree_expr_nonzero_warnv_p): Remove.
620 (fold_builtin_snprintf_chk): Remove.
621 (validate_arglist): Remove.
622 (iterative_hash_exprs_commutative): Move to tree-ssa-dom.c.
623 (simple_cst_list_equal): Remove.
625 (expand_main_function): Remove.
626 (stack_protect_prologue): Remove.
627 (print_vec_tree): Remove.
628 (lookup_scoped_attribute_spec): Remove.
629 (get_attribute_namespace): Remove.
630 (expand_computed_goto): Remove.
631 (expand_asm_stmt): Remove.
632 (list_equal_p): Remove.
633 (ssa_name_nonnegative_p): Remove.
634 (fini_object_sizes): Remove.
635 (addr_expr_of_non_mem_decl_p): Remove.
636 (is_tm_safe_or_pure): Move to trans-mem.c.
637 (more_const_call_expr_args_p): Remove.
638 (save_vtable_map_decl): Remove.
640 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
642 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
645 * config/arc/arc.h (LINK_COMMAND_SPEC): For -ftree-parallelize-loops=*,
646 link to libgomp and its dependencies.
647 * config/ia64/hpux.h (LIB_SPEC): Likewise.
648 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
649 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
650 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's dependencies.
652 2013-11-07 Jakub Jelinek <jakub@redhat.com>
654 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
655 (determine_value_range): Add loop argument. Use get_range_info to
657 (bound_difference): Adjust caller.
659 2013-11-07 Richard Biener <rguenther@suse.de>
660 Jakub Jelinek <jakub@redhat.com>
662 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
663 latches from header PHI arguments from the latch edge.
665 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
668 * varasm.c (output_constant): Handle NULLPTR_TYPE.
670 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
672 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
673 misaligned_prologue_used when it has been set.
675 2013-11-07 Yury Gribov <y.gribov@samsung.com>
676 Jakub Jelinek <jakub@redhat.com>
679 * asan.c (get_mem_refs_of_builtin_call): Allow
680 integer literals as addresses in instrumented builtins.
682 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
684 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
685 Explain why plus_constant is not used.
687 2013-11-07 Richard Biener <rguenther@suse.de>
689 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
690 (canonicalize_value): ... this. Also handle stripping of
692 (get_value, set_lattice_value, get_value_for_expr): Adjust.
693 * gimple-fold.c (canonicalize_constructor_val): Strip TREE_OVERFLOW.
694 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
696 2013-11-07 Richard Biener <rguenther@suse.de>
698 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
700 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
703 * config/i386/i386.md (push peepholer/splitter): Use Pmode
704 with stack_pointer_rtx.
706 2013-11-07 Bin Cheng <bin.cheng@arm.com>
708 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
709 using operand_equal_p.
711 2013-11-07 Bin Cheng <bin.cheng@arm.com>
713 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
714 * tree-affine.c (get_inner_reference_aff): Return base.
715 * tree-affine.h (get_inner_reference_aff): Change prototype.
717 2013-11-06 Tobias Burnus <burnus@net-b.de>
719 * doc/invoke.texi (Wdate-time): Fix typo.
721 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
723 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
724 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
725 fp_arith_reg_operand instead of arith_reg_operand.
727 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
729 * config/sh/sh.md (adddi3): Remove empty constraints.
730 Remove can_create_pseudo_p and arith_reg_operand check.
731 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
734 2013-11-06 Jeff Law <law@redhat.com>
735 Tom Tromey <tromey@redhat.com>
737 * gdbinit.in: Disable strict type checking.
739 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
741 * tree-pass.h (make_pass_live_range_shrinkage): New external.
742 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
743 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
744 (rest_of_handle_live_range_shrinkage): Ditto
745 (class pass_live_range_shrinkage): Ditto.
746 (pass_data_live_range_shrinkage): Ditto.
747 (make_pass_live_range_shrinkage): Ditto.
748 * sched-int.h (initialize_live_range_shrinkage): New prototype.
749 (finish_live_range_shrinkage): Ditto.
750 * sched-deps.c (create_insn_reg_set): Make void return value.
751 * passes.def: Add pass_live_range_shrinkage.
752 * ira.c (update_equiv_regs): Don't move if flag_live_range_shrinkage.
753 * haifa-sched.c (live_range_shrinkage_p): New.
754 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
756 (rank_for_schedule): Add code for pressure relief through live
758 (schedule_insn): Print more debug info.
759 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
760 through live range shrinkage.
761 * doc/invoke.texi (-flive-range-shrinkage): New.
762 * common.opt (flive-range-shrinkage): New.
764 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
767 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
768 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
769 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
770 RTXes that return in AVX256 register.
772 2013-11-06 Richard Biener <rguenther@suse.de>
774 PR tree-optimization/58653
775 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
776 (prepare_initializers_chain): Adjust.
778 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
780 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
781 (gimple_in_transaction): Use bb_in_transaction.
782 * basic-block.h (bb_in_transaction): Relocate here and rename.
783 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
785 2013-11-06 Richard Biener <rguenther@suse.de>
787 * tree.c (drop_tree_overflow): New function.
788 * tree.h (drop_tree_overflow): Declare.
789 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
790 * tree-vrp.c (range_int_cst_singleton_p): Use
791 is_overflow_infinity instead of testing TREE_OVERFLOW.
792 (extract_range_from_assert): Likewise.
793 (zero_nonzero_bits_from_vr): Likewise.
794 (extract_range_basic): Likewise.
795 (register_new_assert_for): Use drop_tree_overflow.
796 (vrp_visit_phi_node): Likewise.
798 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
800 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
801 checking for leaf functions without dynamic stack allocation.
802 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
803 (ia64_expand_prologue): Likewise.
804 * config/mips/mips.c (mips_expand_prologue): Likewise.
805 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
806 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
807 (sparc_flat_expand_prologue): Likewise.
809 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
811 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
812 (__ST3_LANE_FUNC): Likewise.
813 (__ST4_LANE_FUNC): Likewise.
815 2013-11-06 Nick Clifton <nickc@redhat.com>
817 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
818 name returned by msp430_mcu_name.
819 (LIB_SPEC): If a -T option has not been specified then set a
820 default, mcu-specific, linker script.
821 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
822 * config/msp430/msp430.c (msp430x_names): Likewise.
823 Alpha sort the names for ease of comparison.
824 (msp430_mcu_name): New function: Returns a string suitable for
825 use as a C preprocessor symbol based upon the name of the MCU
827 (msp430_option_override): Accept msp430x and msp430xv2 as generic
829 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
831 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
832 unless it is the last part of a spec substring.
833 * doc/invoke.texi (Spec Files): Document space insertion
836 2013-11-06 Christian Bruel <christian.bruel@st.com>
838 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
839 Factorize probabilities, Use adjust_address instead of
840 adjust_automodify_address when possible. Enable for optimize.
841 (sh_expand_strlen): New function.
842 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
843 * config/sh/sh.md (strlensi): New pattern.
844 (UNSPEC_BUILTIN_STRLEN): Define.
846 2013-11-06 Jakub Jelinek <jakub@redhat.com>
849 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
850 (expand_assignment): If *bitpos is negative, set *offset
851 and adjust *bitpos, so that it is not negative.
853 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
855 * config/i386/bdver3.md : Added two additional decoder units
856 to support issue rate of 4 and remodeled vector unit.
857 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
858 architectures is set to 4.
859 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
860 lookahead is set to 4 for BD architectures.
862 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
864 * config/rs6000/rs6000.c (rs6000_option_override_internal):
865 Remove restriction against use of VSX instructions when generating
866 code for little endian mode.
868 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
870 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
871 for both big and little endian.
872 (mulv8hi3): Swap input operands for merge high and merge low
873 instructions for little endian.
875 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
877 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
878 define_insn to define_expand that uses even patterns for big
879 endian and odd patterns for little endian.
880 (vec_widen_smult_even_v16qi): Likewise.
881 (vec_widen_umult_even_v8hi): Likewise.
882 (vec_widen_smult_even_v8hi): Likewise.
883 (vec_widen_umult_odd_v16qi): Likewise.
884 (vec_widen_smult_odd_v16qi): Likewise.
885 (vec_widen_umult_odd_v8hi): Likewise.
886 (vec_widen_smult_odd_v8hi): Likewise.
887 (altivec_vmuleub): New define_insn.
888 (altivec_vmuloub): Likewise.
889 (altivec_vmulesb): Likewise.
890 (altivec_vmulosb): Likewise.
891 (altivec_vmuleuh): Likewise.
892 (altivec_vmulouh): Likewise.
893 (altivec_vmulesh): Likewise.
894 (altivec_vmulosh): Likewise.
896 2013-11-05 Mike Stump <mikestump@comcast.net>
898 * Makefile.in (mostlyclean): Remove c-family objects.
900 2013-11-05 Ian Lance Taylor <iant@google.com>
902 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
903 If possible, add .cfi directives to record change to bx.
904 * config/i386/i386.c (ix86_emit_cfi): New function.
905 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
907 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
909 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
910 immediately after a label for a tablejump pattern.
912 * config/arm/arm.c (is_jump_table): Remove.
913 (create_fix_barrier): Use tablejump_p instead.
914 (arm_reorg): Likewise.
915 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
916 (thumb2_output_casesi): Likewise.
917 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
918 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
919 casesi_shift_media, casesi_load_media): Likewise.
920 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
921 * config/microblaze/microblaze.md: Likewise.
923 2013-11-05 Tobias Burnus <burnus@net-b.de>
925 * doc/invoke.texi (-Wdate-time): Document.
927 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
929 * double-int.c (lshift_double, rshift_double): Remove
930 SHIFT_COUNT_TRUNCATED handling.
932 2013-11-05 Jeff Law <law@redhat.com>
934 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
935 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
936 * gimple-ssa-isolate-paths.c: New file.
937 * gimple.c (check_loadstore): New function.
938 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
939 Verify OP is in the argument list and the argument corresponding
940 to OP is a pointer type. Use operand_equal_p rather than
941 pointer equality when testing if OP is on the nonnull list.
942 Use check_loadstore rather than count_ptr_derefs. Handle
943 GIMPLE_RETURN statements.
944 * tree-vrp.c (infer_nonnull_range): Remove.
945 * gimple.h (infer_nonnull_range): Declare.
946 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
947 * passes.def: Add pass_isolate_erroneous_paths.
948 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
949 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
950 * tree-ssa.c (struct count_ptr_d): Remove.
951 (count_ptr_derefs, count_uses_and_derefs): Remove.
952 * tree-ssa.h (count_uses_and_derefs): Remove.
954 2013-11-05 Jakub Jelinek <jakub@redhat.com>
956 PR rtl-optimization/58997
957 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
958 get_iv_value to be in iv->mode rather than iv->extend_mode.
959 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
960 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
961 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
964 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
966 * gimple.h: Move some prototypes to gimple-expr.h and add to include
968 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
969 Move to gimple-expr.h.
970 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
971 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
972 is_gimple_id, virtual_operand_p, is_gimple_addressable,
973 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
975 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
976 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
977 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
978 is_gimple_address, is_gimple_invariant_address,
979 is_gimple_ip_invariant_address, is_gimple_min_invariant,
980 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
981 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
982 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
983 gimple_decl_printable_name, useless_type_conversion_p,
984 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
986 * gimple-expr.c: New File.
987 (useless_type_conversion_p, gimple_set_body, gimple_body,
988 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
989 gimple_can_coalesce_p, extract_ops_from_tree_1,
990 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
991 is_gimple_address, is_gimple_invariant_address,
992 is_gimple_ip_invariant_address, is_gimple_min_invariant,
993 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
994 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
995 is_gimple_mem_ref_addr): Relocate here.
996 * Makefile.in (OBJS): Add gimple-expr.o.
998 2013-11-05 David Malcolm <dmalcolm@redhat.com>
1000 * gengtype-parse.c (struct_field_seq): Support empty structs.
1002 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1004 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
1006 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1008 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
1010 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
1012 (LONG_DOUBLE_TYPE_SIZE): New define.
1013 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
1015 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
1018 * doc/extend.texi [visibility ("visibility_type")]: Add example
1019 about visibility attribute on namespace declaration.
1021 2013-11-05 Richard Biener <rguenther@suse.de>
1024 * passes.def (all_passes): Start with pass_fixup_cfg again.
1026 2013-11-05 Richard Biener <rguenther@suse.de>
1028 PR tree-optimization/58955
1029 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
1032 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1034 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
1036 (vec_pack_ufix_trunc_v2df): Likewise.
1038 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
1041 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
1042 pattern, instead of word_mode.
1044 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
1046 (set_storage_via_setmem): Likewise.
1048 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
1050 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
1051 gimple.h and the rest of the condition in eliminate_build.
1052 (eliminate_build): Call new routine.
1053 * gimple.h (phi_ssa_name_p): Delete.
1055 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
1057 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
1058 case of no prefix and reserving zero slots, because when that's the
1059 case we'll never get here.
1060 * vec.h (va_heap::reserve): Don't try and handle
1061 vec_prefix::calculate_allocation returning zero because that should
1064 2013-11-05 Richard Biener <rguenther@suse.de>
1067 * tree-dfa.c (get_ref_base_and_extent): Merge common code
1068 in MEM_REF and TARGET_MEM_REF handling. Make sure to
1069 process trailing array detection before diving into the
1070 view-converted object (and possibly apply some extra offset).
1072 2013-11-05 Joseph Myers <joseph@codesourcery.com>
1074 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
1076 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
1078 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
1080 PR tree-optimization/58958
1081 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
1082 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
1084 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
1086 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
1087 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
1089 2013-11-05 Jakub Jelinek <jakub@redhat.com>
1091 PR tree-optimization/58984
1092 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
1093 set *SIZE_P if non-NULL on success.
1094 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
1096 (ipcp_transform_function): Likewise. Punt if size of access
1097 is different from TYPE_SIZE on v->value's type.
1099 2013-11-05 Tobias Burnus <burnus@net-b.de>
1101 * doc/invoke.texi (-fopenmp-simd): Document new option.
1102 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
1103 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
1104 * tree.c (attribute_value_equal): Ditto.
1106 2013-11-04 Wei Mi <wmi@google.com>
1108 * sched-rgn.c (add_branch_dependences): Keep insns in
1109 a SCHED_GROUP at the end of BB to remain their location.
1111 2013-11-04 Wei Mi <wmi@google.com>
1113 * config/i386/i386.c (memory_address_length): Extract a part
1114 of code to rip_relative_addr_p.
1115 (rip_relative_addr_p): New Function.
1116 (ix86_macro_fusion_p): Ditto.
1117 (ix86_macro_fusion_pair_p): Ditto.
1118 * config/i386/i386.h: Add new tune features about macro-fusion.
1119 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
1120 * doc/tm.texi: Generated.
1121 * doc/tm.texi.in: Ditto.
1122 * haifa-sched.c (try_group_insn): New Function.
1123 (group_insns_for_macro_fusion): Ditto.
1124 (sched_init): Call group_insns_for_macro_fusion.
1125 * target.def: Add two hooks: macro_fusion_p and
1126 macro_fusion_pair_p.
1128 2013-11-04 Kostya Serebryany <kcc@google.com>
1130 Update to match the changed asan API.
1131 * asan.c (asan_function_start): New function.
1132 (asan_emit_stack_protection): Update the string stored in the
1133 stack red zone to match new API. Store the PC of the current
1134 function in the red zone.
1135 (asan_global_struct): Update the __asan_global definition to match
1137 (asan_add_global): Ditto.
1138 * asan.h (asan_function_start): New prototype.
1139 * final.c (final_start_function): Call asan_function_start.
1140 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
1143 2013-11-04 Wei Mi <wmi@google.com>
1145 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
1146 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
1147 * config/i386/i386.c (ix86_option_override_internal): Ditto.
1148 (ix86_issue_rate): Ditto.
1149 (ix86_adjust_cost): Ditto.
1150 (ia32_multipass_dfa_lookahead): Ditto.
1151 (ix86_sched_init_global): Ditto.
1152 (get_builtin_code_for_version): Ditto.
1153 * config/i386/i386.h (enum target_cpu_default): Ditto.
1154 (enum processor_type): Ditto.
1155 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
1157 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
1159 PR rtl-optimization/58967
1160 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
1161 !lra_in_progress for mode sizes bigger word.
1163 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1165 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
1166 arguments to merge instruction for little endian.
1167 (vec_widen_umult_lo_v16qi): Likewise.
1168 (vec_widen_smult_hi_v16qi): Likewise.
1169 (vec_widen_smult_lo_v16qi): Likewise.
1170 (vec_widen_umult_hi_v8hi): Likewise.
1171 (vec_widen_umult_lo_v8hi): Likewise.
1172 (vec_widen_smult_hi_v8hi): Likewise.
1173 (vec_widen_smult_lo_v8hi): Likewise.
1175 2013-11-04 Ian Lance Taylor <iant@google.com>
1177 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
1178 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
1179 builtins that take pointers.
1180 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
1181 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
1182 (run_gcc): Pass -fnon-call-exceptions.
1184 2013-11-04 Jakub Jelinek <jakub@redhat.com>
1186 * optabs.c (expand_vec_perm): Revert one incorrect line from
1189 PR tree-optimization/58978
1190 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
1191 use_stmt by single_imm_use directly. Only call single_imm_use
1194 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
1196 PR rtl-optimization/58968
1197 * lra-spills.c (return_regno_p): New function.
1198 (lra_final_code_change): Use it.
1200 2013-11-04 Joseph Myers <joseph@codesourcery.com>
1202 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
1203 * target.def (float_exceptions_rounding_supported_p): New hook.
1204 * targhooks.c (default_float_exceptions_rounding_supported_p): New
1206 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
1207 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
1209 * doc/tm.texi: Regenerate.
1210 * config.gcc (powerpc*-*-linux*): Set extra_objs.
1211 * config/rs6000/rs6000-linux.c: New file.
1212 * config/rs6000/rs6000-protos.h
1213 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
1214 * config/rs6000/linux.h
1215 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
1216 * config/rs6000/linux64.h
1217 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
1218 * config/rs6000/t-linux (rs6000-linux.o): New rule.
1219 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
1221 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1223 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
1224 Replace the define_insn_and_split with a define_insn and two
1225 define_splits, with the split after reload re-permuting the source
1226 register to its original value.
1227 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
1228 (*vsx_le_perm_store_v8hi): Likewise.
1229 (*vsx_le_perm_store_v16qi): Likewise.
1231 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1233 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
1236 2013-11-04 Jakub Jelinek <jakub@redhat.com>
1238 PR tree-optimization/58946
1239 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
1240 bbs with bbinfo[idx].op != NULL before all blocks with
1241 bbinfo[idx].op == NULL.
1243 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
1245 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
1246 (avr_dump_double_int_hex): Likewise.
1247 (avr_log_vadump): Remove %D and %X handling.
1248 * config/avr/avr.c (avr_double_int_push_digit): Delete.
1249 (avr_map_op_t): Change map from double_int to unsigned int.
1250 (avr_map_op): Update accordingly.
1251 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
1252 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
1253 unsigned ints rather than double_ints.
1255 2013-11-03 Marek Polacek <polacek@redhat.com>
1257 Implement -fsanitize=vla-bound.
1258 * opts.c (common_handle_option): Handle vla-bound.
1259 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
1260 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
1261 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
1263 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1265 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
1268 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
1270 * config/i386/constraints.md (Ts, Tv): New address constrains.
1271 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
1272 constraint for address_no_seg_operand.
1273 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
1274 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
1275 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
1276 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
1277 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
1278 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
1279 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
1280 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
1281 constraint for vsib_address_operand.
1283 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
1285 * gcse.c (pre_delete): Remove references to regmove from comments.
1286 * recog.c: (validate_replace_rtx_1): Likewise.
1287 * config/rl78/rl78.c: Likewise.
1288 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
1289 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
1290 * common/config/mmix/mmix-common.c: Likewise.
1292 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
1294 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
1295 * gimplify.c (gimplify_compound_lval): Likewise.
1296 * graphite-clast-to-gimple.c (gloog): Likewise.
1297 * graphite-dependences.c (loop_is_parallel_p): Likewise.
1298 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
1299 (limit_scop); Likewise.
1300 (build_scops): Likewise.
1301 (dot_scop): Likewise.
1302 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
1303 (build_scop_drs): Likewise.
1304 (insert_stmts): Likewise.
1305 (insert_out_of_ssa_copy): Likewise.
1306 (remove_phi): Likewise.
1307 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
1308 * hw-doloop.c (discover_loop): Likewise.
1309 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
1310 * tree-dfa.c (dump_enumerated_decls): Likewise.
1311 * tree-if-conv.c (if_convertable_loop_p): Likewise.
1312 * tree-inline.c (tree_function_versioning): Likewise.
1313 * tree-loop-distribution.c (build_rdg): Likewise.
1314 (rdg_flag_vertex_and_dependent): Likewise.
1315 (distribute_loop): Likewise.
1316 * tree-parloops.c (loop_parallel_p): Likewise.
1317 (eliminate_local_variables): Likewise.
1318 (separate_decls_in_region): Likewise.
1319 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
1320 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
1321 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
1322 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
1323 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
1324 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
1325 (vectorizable_condition): Likewise.
1327 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
1329 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
1330 * configure: Regenerate.
1331 * config/i386/i386.md (*movdi_internal): Change
1332 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
1333 (*movdf_internal): Ditto.
1334 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
1335 * config/i386/sse.md (vec_concatv2di): Output interunit movq
1336 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
1338 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
1340 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
1343 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1345 Automated part of renaming of symtab_node_base to symtab_node.
1347 Patch autogenerated by rename_symtab.py from
1348 https://github.com/davidmalcolm/gcc-refactoring-scripts
1349 revision 58bb219cc090b2f4516a9297d868c245495ee622
1350 with ChangeLog entry fixed up by hand.
1352 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
1354 (cgraph_node_for_asm): Likewise.
1355 * cgraph.h (symtab_node_base): Likewise.
1356 (cgraph_node): Likewise.
1357 (varpool_node): Likewise.
1358 (is_a_helper <cgraph_node>::test): Likewise.
1359 (is_a_helper <varpool_node>::test): Likewise.
1360 (symtab_nodes): Likewise.
1361 (symtab_register_node): Likewise.
1362 (symtab_unregister_node): Likewise.
1363 (symtab_remove_node): Likewise.
1364 (symtab_get_node): Likewise.
1365 (symtab_node_for_asm): Likewise.
1366 (symtab_node_asm_name): Likewise.
1367 (symtab_node_name): Likewise.
1368 (symtab_insert_node_to_hashtable): Likewise.
1369 (symtab_add_to_same_comdat_group): Likewise.
1370 (symtab_dissolve_same_comdat_group_list): Likewise.
1371 (dump_symtab_node): Likewise.
1372 (debug_symtab_node): Likewise.
1373 (dump_symtab_base): Likewise.
1374 (verify_symtab_node): Likewise.
1375 (verify_symtab_base): Likewise.
1376 (symtab_used_from_object_file_p): Likewise.
1377 (symtab_alias_ultimate_target): Likewise.
1378 (symtab_resolve_alias): Likewise.
1379 (fixup_same_cpp_alias_visibility): Likewise.
1380 (symtab_for_node_and_aliases): Likewise.
1381 (symtab_nonoverwritable_alias): Likewise.
1382 (availability symtab_node_availability): Likewise.
1383 (symtab_semantically_equivalent_p): Likewise.
1384 (fixup_same_cpp_alias_visibility): Likewise.
1385 (symtab_prevail_in_asm_name_hash): Likewise.
1387 (varpool): Likewise.
1388 (varpool_first_variable): Likewise.
1389 (varpool_next_variable): Likewise.
1390 (varpool_first_static_initializer): Likewise.
1391 (varpool_next_static_initializer): Likewise.
1392 (varpool_first_defined_variable): Likewise.
1393 (varpool_next_defined_variable): Likewise.
1394 (cgraph_first_defined_function): Likewise.
1395 (cgraph_next_defined_function): Likewise.
1396 (cgraph_first_function): Likewise.
1397 (cgraph_next_function): Likewise.
1398 (cgraph_first_function_with_gimple_body): Likewise.
1399 (cgraph_next_function_with_gimple_body): Likewise.
1400 (symtab_alias_target): Likewise.
1401 (symtab_real_symbol_p): Likewise.
1402 (symtab_can_be_discarded): Likewise.
1403 * cgraphbuild.c (mark_address): Likewise.
1404 (mark_load): Likewise.
1405 (mark_store): Likewise.
1406 * cgraphunit.c (decide_is_symbol_needed): Likewise.
1408 (enqueue_node): Likewise.
1409 (referred_to_p): Likewise.
1410 (cgraph_process_same_body_aliases): Likewise.
1411 (analyze_functions): Likewise.
1412 (handle_alias_pairs): Likewise.
1413 (output_weakrefs): Likewise.
1414 (compile): Likewise.
1415 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1416 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1417 * ipa-prop.c (remove_described_reference): Likewise.
1418 (try_decrement_rdesc_refcount): Likewise.
1419 (ipa_edge_duplication_hook): Likewise.
1420 * ipa-ref.c (ipa_record_reference): Likewise.
1421 (ipa_maybe_record_reference): Likewise.
1422 (ipa_clone_ref): Likewise.
1423 (ipa_clone_references): Likewise.
1424 (ipa_clone_referring): Likewise.
1425 (ipa_find_reference): Likewise.
1426 (ipa_remove_stmt_references): Likewise.
1427 (ipa_clear_stmts_in_references): Likewise.
1428 * ipa-ref.h (symtab_node_base): Likewise.
1429 (ipa_ref): Likewise.
1430 (ipa_record_reference): Likewise.
1431 (ipa_maybe_record_reference): Likewise.
1432 (ipa_clone_references): Likewise.
1433 (ipa_clone_referring): Likewise.
1434 (ipa_clone_ref): Likewise.
1435 (ipa_find_reference): Likewise.
1436 (ipa_remove_stmt_references): Likewise.
1437 (ipa_clear_stmts_in_references): Likewise.
1438 * ipa-reference.c (ipa_reference_write_optimization_summary):
1440 * ipa.c (enqueue_node): Likewise.
1441 (process_references): Likewise.
1442 (walk_polymorphic_call_targets): Likewise.
1443 (symtab_remove_unreachable_nodes): Likewise.
1444 (address_taken_from_non_vtable_p): Likewise.
1445 (comdat_can_be_unshared_p_1): Likewise.
1446 (comdat_can_be_unshared_p): Likewise.
1447 (can_replace_by_local_alias): Likewise.
1448 (function_and_variable_visibility): Likewise.
1449 * is-a.h: Likewise (within example in comment).
1450 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
1451 (lto_symtab_encoder_encode): Likewise.
1452 (lto_symtab_encoder_delete_node): Likewise.
1453 (lto_symtab_encoder_in_partition_p): Likewise.
1454 (lto_set_symtab_encoder_in_partition): Likewise.
1455 (output_refs): Likewise.
1456 (compute_ltrans_boundary): Likewise.
1457 (output_symtab): Likewise.
1458 (input_node): Likewise.
1459 (input_ref): Likewise.
1460 (input_edge): Likewise.
1461 (input_cgraph_1): Likewise.
1462 (input_refs): Likewise.
1463 (output_cgraph_opt_summary): Likewise.
1464 (input_node_opt_summary): Likewise.
1465 (input_cgraph_opt_section): Likewise.
1466 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1468 * lto-streamer-out.c (lto_output): Likewise.
1469 (output_symbol_p): Likewise.
1470 (produce_symtab): Likewise.
1471 * lto-streamer.h (lto_encoder_entry): Likewise.
1472 (lto_free_function_in_decl_state_for_node): Likewise.
1473 (lto_symtab_encoder_encode): Likewise.
1474 (lto_symtab_encoder_delete_node): Likewise.
1475 (lto_symtab_encoder_in_partition_p): Likewise.
1476 (lto_set_symtab_encoder_in_partition): Likewise.
1477 (lto_symtab_encoder_lookup): Likewise.
1478 (lsei_node): Likewise.
1479 (lto_symtab_encoder_deref): Likewise.
1480 * symtab.c (symtab_hash): Likewise.
1481 (assembler_name_hash): Likewise.
1482 (symtab_nodes): Likewise.
1483 (hash_node): Likewise.
1484 (eq_node): Likewise.
1485 (hash_node_by_assembler_name): Likewise.
1486 (eq_assembler_name): Likewise.
1487 (insert_to_assembler_name_hash): Likewise.
1488 (unlink_from_assembler_name_hash): Likewise.
1489 (symtab_prevail_in_asm_name_hash): Likewise.
1490 (symtab_register_node): Likewise.
1491 (symtab_insert_node_to_hashtable): Likewise.
1492 (symtab_unregister_node): Likewise.
1493 (symtab_get_node): Likewise.
1494 (symtab_remove_node): Likewise.
1495 (symtab_initialize_asm_name_hash): Likewise.
1496 (symtab_node_for_asm): Likewise.
1497 (symtab_add_to_same_comdat_group): Likewise.
1498 (symtab_dissolve_same_comdat_group_list): Likewise.
1499 (symtab_node_asm_name): Likewise.
1500 (symtab_node_name): Likewise.
1501 (dump_symtab_base): Likewise.
1502 (dump_symtab_node): Likewise.
1503 (dump_symtab): Likewise.
1504 (debug_symtab_node): Likewise.
1505 (verify_symtab_base): Likewise.
1506 (verify_symtab_node): Likewise.
1507 (verify_symtab): Likewise.
1508 (symtab_used_from_object_file_p): Likewise.
1509 (symtab_node_availability): Likewise.
1510 (symtab_alias_ultimate_target): Likewise.
1511 (fixup_same_cpp_alias_visibility): Likewise.
1512 (symtab_resolve_alias): Likewise.
1513 (symtab_for_node_and_aliases): Likewise.
1514 (symtab_for_node_and_aliases): Likewise.
1515 (symtab_nonoverwritable_alias_1): Likewise.
1516 (symtab_nonoverwritable_alias): Likewise.
1517 (symtab_semantically_equivalent_p): Likewise.
1518 * value-prof.c (init_node_map): Likewise.
1519 * varasm.c (find_decl): Likewise.
1520 * varpool.c (varpool_node_for_asm): Likewise.
1521 (varpool_remove_unreferenced_decls): Likewise.
1523 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1525 Manual part of renaming of symtab_node_base to symtab_node.
1527 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
1528 clashes with the preferred name for the base class.
1529 (const_symtab_node): Remove redundant typedef.
1531 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1533 * optabs.c (expand_vec_perm): Avoid vector mode punning
1534 SUBREGs in SET_DEST.
1535 * expmed.c (store_bit_field_1): Likewise.
1536 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
1537 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
1538 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
1539 * config/i386/i386.c (ix86_expand_vector_move_misalign,
1540 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
1541 ix86_expand_sse_unpack, ix86_expand_args_builtin,
1542 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
1543 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
1544 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
1545 expand_vec_perm_vpshufb2_vpermq,
1546 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
1547 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
1548 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
1549 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
1552 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
1554 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
1555 Use gcc_unreachable for unhandled modes. Do not check results of
1556 expand_simple_binop. If not expanded to target, move the result.
1558 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
1559 Shiva Chen <shiva0217@gmail.com>
1561 * config.gcc (nds32*-*-*): Add nds32 target.
1562 * config/nds32/nds32.c: New file.
1563 * config/nds32/nds32.h: New file.
1564 * config/nds32/nds32.md: New file.
1565 * config/nds32/constants.md: New file.
1566 * config/nds32/constraints.md: New file.
1567 * config/nds32/iterators.md: New file.
1568 * config/nds32/nds32-doubleword.md: New file.
1569 * config/nds32/nds32-intrinsic.md: New file.
1570 * config/nds32/nds32_intrinsic.h: New file.
1571 * config/nds32/nds32-modes.def: New file.
1572 * config/nds32/nds32-multiple.md: New file.
1573 * config/nds32/nds32.opt: New file.
1574 * config/nds32/nds32-opts.h: New file.
1575 * config/nds32/nds32-protos.h: New file.
1576 * config/nds32/nds32-peephole2.md: New file.
1577 * config/nds32/pipelines.md: New file.
1578 * config/nds32/predicates.md: New file.
1579 * config/nds32/t-mlibs: New file.
1580 * common/config/nds32: New directory and files.
1582 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
1583 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
1584 * doc/install.texi (Cross-Compiler-Specific Options): Document
1585 --with-nds32-lib for nds32 target.
1586 * doc/extend.texi (Function Attributes, Target Builtins): Document
1587 nds32 specific attributes.
1589 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
1591 * lra-constraints (process_alt_operands): Use the result
1592 elimination register for operand when matching constraints.
1594 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1596 * tree-vrp.c (maybe_set_nonzero_bits): New function.
1597 (remove_range_assertions): Call it.
1599 * tree.c (tree_ctz): New function.
1600 * tree.h (tree_ctz): New prototype.
1601 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
1602 first argument from tree to const_tree.
1603 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
1604 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
1605 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
1607 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
1608 * tree-vect-loop.c (vect_analyze_loop_operations,
1609 vect_transform_loop): Don't force scalar loop for bound just because
1610 number of iterations is unknown, only do it if it is not known to be
1611 a multiple of vectorization_factor.
1612 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
1614 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
1615 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
1616 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
1617 (set_nonzero_bits, get_nonzero_bits): New prototypes.
1618 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
1619 a default def isn't partially constant.
1620 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
1621 is known to be partially zero.
1622 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
1623 to see if a default def isn't partially constant.
1624 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
1625 creation of a range, if VR_RANGE, try to improve nonzero_bits from
1627 (set_nonzero_bits, get_nonzero_bits): New functions.
1629 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
1630 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
1631 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
1632 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
1633 uses but in the condition and ASSERT_EXPR and the other successor of
1634 the predecessor bb is __builtin_unreachable (), set_range_info of the
1635 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
1637 2013-10-31 Martin Jambor <mjambor@suse.cz>
1639 PR rtl-optimization/58934
1641 2013-10-30 Martin Jambor <mjambor@suse.cz>
1642 PR rtl-optimization/10474
1643 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1645 (interesting_dest_for_shprep): New function.
1646 (split_live_ranges_for_shrink_wrap): Likewise.
1647 (ira): Calculate dominance info and df analysis. Call
1648 split_live_ranges_for_shrink_wrap.
1650 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1651 Yury Gribov <y.gribov@samsung.com>
1654 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
1655 shadow pointer to avoid clobbering the main one.
1657 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1659 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
1661 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1664 * ira-color.c (update_costs_from_copies): Add new parameter. Use
1665 it for calling update_costs_from_allocno.
1666 (assign_hard_reg): Call restore_costs_from_copies only for
1667 !retry_p. Pass new argument to update_costs_from_copies.
1668 (color_pass): Pass new argument to update_costs_from_copies.
1669 (ira_mark_allocation_change): Ditto.
1671 2013-10-30 Sharad Singhai <singhai@google.com>
1674 * opts.c (common_handle_option): Remove deprecated option
1675 -ftree-vectorizer-verbose.
1676 * doc/invoke.texi (Debugging Options): Ditto.
1677 * opts-global.c (handle_common_deferred_options): Ditto.
1678 (dump_remap_tree_vectorizer_verbose): Delete.
1679 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
1681 2013-10-30 DJ Delorie <dj@redhat.com>
1683 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
1684 that take no arguments.
1686 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
1689 * reload1.c (update_eliminables_and_spill): New function, broken
1691 (reload): Use it. Check for frame size change after frame size
1692 alignment, and call update_eliminables_and_spill first if continue-ing.
1694 2013-10-30 Cong Hou <congh@google.com>
1697 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
1698 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
1699 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
1701 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
1703 PR rtl-optimization/58369
1704 * reload1.c (compute_reload_subreg_offset): New function.
1705 (choose_reload_regs): Use it to pass endian-correct
1706 offset to subreg_regno_offset.
1708 2013-10-30 Tobias Burnus <burnus@net-b.de>
1711 * tree-cfg.c (replace_loop_annotate): Replace warning by
1714 2013-10-30 Jason Merrill <jason@redhat.com>
1716 * configure.ac (loose_warn): Add -Wno-format if
1717 --disable-build-format-warnings.
1719 2013-10-30 David Malcolm <dmalcolm@redhat.com>
1721 * cgraphunit.c (analyze_functions): Split symtab_node declarations
1722 onto multiple lines to make things easier for rename_symtab.py.
1724 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
1725 (symtab_semantically_equivalent_p): Likewise.
1727 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1730 * lra.c (check_rtl): Remove address check before LRA work.
1732 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
1734 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
1735 POINTER_PLUS_EXPR in the defining statement.
1737 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1739 * regmove.c: Remove.
1740 * tree-pass.h (make_pass_regmove): Remove.
1741 * timevar.def (TV_REGMOVE): Remove.
1742 * passes.def (pass_regmove): Remove.
1743 * opts.c (default_options_table): Remove entry for regmove.
1744 * doc/passes.texi: Remove regmove pass description.
1745 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
1747 (-fdump-rtl-regmove): Ditto.
1748 * common.opt (foptimize-register-move, fregmove): Ignore.
1749 * Makefile.in (OBJS): Remove regmove.o.
1750 * regmove.c: Remove.
1751 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
1753 (struct ira_allocno) New member allocno_prefs.
1754 (ALLOCNO_PREFS): New macro.
1755 (ira_prefs, ira_prefs_num): New external vars.
1756 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
1757 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
1758 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
1760 (ira_add_allocno_copy_to_list): Remove prototype.
1761 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
1762 (ira_pref_iterator): New type.
1763 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
1764 (FOR_EACH_PREF): New macro.
1765 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
1766 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
1768 (ira_setup_alts): New function.
1769 (decrease_live_ranges_number): New function.
1770 (ira): Call the above function.
1771 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
1772 (ira_create_allocno): Initialize allocno prefs.
1773 (pref_pool, pref_vec): New static vars.
1774 (initiate_prefs, find_allocno_pref, ira_create_pref): New
1776 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
1777 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
1778 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
1779 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
1780 (ira_add_allocno_copy_to_list): Make static. Rename to
1781 add_allocno_copy_to_list.
1782 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
1783 swap_allocno_copy_ends_if_necessary.
1784 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
1785 ira_remove_allocno_prefs.
1786 (ira_flattening): Ditto.
1787 (ira_build): Call initiate_prefs, print_prefs.
1788 (ira_destroy): Call finish_prefs.
1789 * ira-color.c (struct update_cost_record): New.
1790 (struct allocno_color_data): Add new member update_cost_records.
1791 (update_cost_record_pool): New static var.
1792 (init_update_cost_records, get_update_cost_record): New functions.
1793 (free_update_cost_record_list, finish_update_cost_records): Ditto.
1794 (struct update_cost_queue_elem): Add member from.
1795 (initiate_cost_update): Call init_update_cost_records.
1796 (finish_cost_update): Call finish_update_cost_records.
1797 (queue_update_cost, get_next_update_cost): Add new param from.
1798 (Update_allocno_cost, update_costs_from_allocno): New functions.
1799 (update_costs_from_prefs): Ditto.
1800 (update_copy_costs): Rename to update_costs_from_copies.
1801 (restore_costs_from_copies): New function.
1802 (update_conflict_hard_regno_costs): Don't go back.
1803 (assign_hard_reg): Call restore_costs_from_copies. Add printing
1805 (pop_allocnos): Add priniting more debug info.
1806 (color_allocnos): Remove prefs for conflicting hard regs.
1807 Call update_costs_from_prefs.
1808 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
1809 (get_dup_num): Rename, modify, and move to ira.c
1810 (process_regs_for_copy): Add prefs.
1811 (add_insn_allocno_copies): Put src as first arg of
1812 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
1813 * ira-costs.c (record_reg_classes): Modify and move code into
1814 record_operands_costs.
1815 (find_costs_and_classes): Create prefs for the hard reg of small
1817 (process_bb_node_for_hard_reg_moves): Add prefs.
1819 2013-10-30 Richard Biener <rguenther@suse.de>
1822 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
1823 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
1825 (pre_and_rev_post_order_compute): ... which now wraps it.
1826 * graph.c (draw_cfg_nodes_no_loops): Use
1827 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
1830 2013-10-30 Christian Bruel <christian.bruel@st.com>
1832 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
1833 (sh_expand_cmpstr): Handle known align and schedule improvements.
1834 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
1835 * config/sh/sh.md (cmpstrnsi): New pattern.
1837 2013-10-30 Martin Jambor <mjambor@suse.cz>
1839 PR rtl-optimization/10474
1840 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1842 (interesting_dest_for_shprep): New function.
1843 (split_live_ranges_for_shrink_wrap): Likewise.
1844 (ira): Calculate dominance info and df analysis. Call
1845 split_live_ranges_for_shrink_wrap.
1847 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1850 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1852 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1854 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1855 * tree.h (POINTER_BOUNDS_P): New.
1856 (BOUNDED_TYPE_P): New.
1858 (pointer_bounds_type_node): New.
1859 * tree.c (build_common_tree_nodes): Initialize
1860 pointer_bounds_type_node.
1861 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1862 (gimple_call_num_nobnd_args): New.
1863 (gimple_call_nobnd_arg): New.
1864 (gimple_return_retbnd): New.
1865 (gimple_return_set_retbnd): New
1866 * gimple.c (gimple_build_return): Increase number of ops
1867 for return statement.
1868 (gimple_call_get_nobnd_arg_index): New.
1869 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1871 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1873 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1874 with "chkp ctor" and "bnd_legacy" attributes.
1875 * gimplify.c (gimplify_init_constructor): Avoid infinite
1876 loop during gimplification of bounds initializer.
1878 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1880 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1881 (handle_bnd_legacy): New.
1882 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1883 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1886 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1888 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1889 (BT_FN_PTR_CONST_PTR): New.
1890 (BT_FN_CONST_PTR_CONST_PTR): New.
1891 (BT_FN_PTR_CONST_PTR_SIZE): New.
1892 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1893 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1894 (BT_FN_VOID_CONST_PTR_SIZE): New.
1895 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1896 * chkp-builtins.def: New.
1897 * builtins.def: include chkp-builtins.def.
1898 (DEF_CHKP_BUILTIN): New.
1899 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1900 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1901 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1902 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1903 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1904 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1905 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1906 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1907 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1908 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1909 * common.opt (fcheck-pointer-bounds): New.
1910 * toplev.c (process_options): Check Pointer Bounds Checker is
1912 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1914 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1916 * target.def (builtin_chkp_function): New.
1917 (chkp_bound_type): New.
1918 (chkp_bound_mode): New.
1919 (fn_abi_va_list_bounds_size): New.
1920 (load_bounds_for_arg): New.
1921 (store_bounds_for_arg): New.
1922 * targhooks.h (default_load_bounds_for_arg): New.
1923 (default_store_bounds_for_arg): New.
1924 (default_fn_abi_va_list_bounds_size): New.
1925 (default_chkp_bound_type): New.
1926 (default_chkp_bound_mode): New.
1927 (default_builtin_chkp_function): New.
1928 * targhooks.c (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 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1935 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1936 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1937 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1938 (TARGET_CHKP_BOUND_TYPE): New.
1939 (TARGET_CHKP_BOUND_MODE): New.
1940 * doc/tm.texi: Regenerated.
1941 * langhooks.h (lang_hooks): Add chkp_supported field.
1942 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1943 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1945 2013-10-29 Andrew Pinski <apinski@cavium.com>
1947 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1948 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1949 fails, combining the branches anyways.
1950 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1951 increases the number of combinings.
1952 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1954 2013-10-29 Mike Stump <mikestump@comcast.net>
1956 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1957 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1958 (make_vector_mode): Increase namebuf to 16.
1959 (emit_insn_modes_h): When processing BImode, don't
1960 also match partial int modes.
1961 (emit_class_narrowest_mode): Likewise.
1963 * config/bfin/bfin-modes.def: Add precision to PDI.
1964 * config/m32c/m32c-modes.def: Add precision to PSI.
1965 * config/msp430/msp430-modes.def: Add precision to PSI.
1966 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1967 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1969 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1972 * config/sh/sh.md (*addc): Rename existing variations to ...
1973 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1974 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1975 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1976 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1978 2013-10-29 Teresa Johnson <tejohnson@google.com>
1981 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1982 insanities when updating probabilities.
1984 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1986 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1987 prettyprinter for cgraph_node to reflect the conversion of the
1988 symtable types to a C++ class hierarchy: it now *is* a
1989 symtab_node_base, rather than having one (named "symbol").
1991 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1993 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
1994 __cilkrts_pop_frame. If matched, then return true for built-in
1996 (expand_builtin): Added BUILT_IN_CILK_DETACH and
1997 BUILT_IN_CILK_POP_FRAME case.
1998 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
1999 (lhs_cilk_detect_spawn): Likewise.
2000 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
2001 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
2002 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
2003 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
2004 (LANG_HOOKS_CILKPLUS): Likewise.
2005 * tree.h (CILK_SPAWN_FN): Likewise.
2006 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
2007 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
2008 (OBJS): Added cilk-common.o.
2009 (BUILTINS_DEF): Added cilk-builtins.def.
2010 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
2011 (lhd_cilk_detect_spawn): Likewise.
2012 * langhooks.h (lang_hooks_for_cilkplus): New struct.
2013 (struct lang_hooks): Added new field called "cilkplus."
2014 * cilk-common.c: New file.
2016 * cilk-builtins.def: Likewise.
2017 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
2018 "__cilk" macro and set it to 200.
2019 * function.h (struct function::cilk_frame_decl): New field.
2020 (struct function::is_cilk_function): Likewise.
2021 (struct function::calls_cilk_spawn): Likewise.
2022 * gimplify.c (gimplify_call_expr): Added a check if the function call
2023 being gimplified is a spawn detach point. If so, then add pop_frame
2024 and detach function calls.
2025 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
2026 for gimplifying _Cilk_spawn and _Cilk_sync statements.
2027 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
2028 function. If so, added a _Cilk_sync and gimplified it.
2029 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
2030 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
2031 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
2033 (can_inline_edge_p): Prevent inling of spawnee function.
2034 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
2035 for functions that use Cilk keywords.
2036 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
2037 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
2038 CILK_SYNC_STMT cases.
2039 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
2041 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
2042 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
2043 * passes.texi (Cilk Keywords): New section that describes the compiler
2044 code changes for handling Cilk Keywords.
2046 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2048 Patch autogenerated by refactor_symtab.py from
2049 https://github.com/davidmalcolm/gcc-refactoring-scripts
2050 revision 58bb219cc090b2f4516a9297d868c245495ee622
2052 * asan.c (asan_finish_file): Update for conversion of symtab types to
2053 a true class hierarchy.
2054 * cfgexpand.c (estimated_stack_frame_size): Likewise.
2055 * cgraph.c (cgraph_get_body): Likewise.
2056 (cgraph_get_create_real_symbol_node): Likewise.
2057 (verify_cgraph_node): Likewise.
2058 (verify_edge_corresponds_to_fndecl): Likewise.
2059 (verify_edge_count_and_frequency): Likewise.
2060 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
2061 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
2062 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
2063 (cgraph_node_cannot_return): Likewise.
2064 (cgraph_set_pure_flag_1): Likewise.
2065 (cgraph_set_const_flag_1): Likewise.
2066 (cgraph_set_nothrow_flag_1): Likewise.
2067 (cgraph_make_node_local_1): Likewise.
2068 (cgraph_for_node_and_aliases): Likewise.
2069 (cgraph_for_node_thunks_and_aliases): Likewise.
2070 (cgraph_node_can_be_local_p): Likewise.
2071 (cgraph_node_cannot_be_local_p_1): Likewise.
2072 (cgraph_function_body_availability): Likewise.
2073 (dump_cgraph_node): Likewise.
2074 (cgraph_rtl_info): Likewise.
2075 (cgraph_mark_address_taken_node): Likewise.
2076 (cgraph_remove_node): Likewise.
2077 (cgraph_release_function_body): Likewise.
2078 (cgraph_update_edges_for_call_stmt_node): Likewise.
2079 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
2080 (cgraph_make_edge_direct): Likewise.
2081 (cgraph_resolve_speculation): Likewise.
2082 (cgraph_speculative_call_info): Likewise.
2083 (cgraph_turn_edge_to_speculative): Likewise.
2084 (cgraph_create_edge_1): Likewise.
2085 (cgraph_set_call_stmt): Likewise.
2086 (cgraph_node_for_asm): Likewise.
2087 (cgraph_add_thunk): Likewise.
2088 (cgraph_same_body_alias): Likewise.
2089 (cgraph_create_function_alias): Likewise.
2090 (cgraph_create_node): Likewise.
2091 (cgraph_create_empty_node): Likewise.
2092 (record_function_versions): Likewise.
2093 (used_from_object_file_p): Likewise.
2094 * cgraph.h (symtab_can_be_discarded): Likewise.
2095 (symtab_real_symbol_p): Likewise.
2096 (cgraph_mark_force_output_node): Likewise.
2097 (cgraph_edge_recursive_p): Likewise.
2098 (symtab_alias_target): Likewise.
2099 (varpool_all_refs_explicit_p): Likewise.
2100 (varpool_can_remove_if_no_refs): Likewise.
2101 (cgraph_only_called_directly_or_aliased_p): Likewise.
2102 (cgraph_next_function_with_gimple_body): Likewise.
2103 (cgraph_first_function_with_gimple_body): Likewise.
2104 (cgraph_function_with_gimple_body_p): Likewise.
2105 (cgraph_next_function): Likewise.
2106 (cgraph_first_function): Likewise.
2107 (cgraph_next_defined_function): Likewise.
2108 (cgraph_first_defined_function): Likewise.
2109 (varpool_next_defined_variable): Likewise.
2110 (varpool_first_defined_variable): Likewise.
2111 (varpool_next_static_initializer): Likewise.
2112 (varpool_first_static_initializer): Likewise.
2113 (varpool_next_variable): Likewise.
2114 (varpool_first_variable): Likewise.
2115 (varpool_node_name): Likewise.
2116 (varpool): Likewise.
2118 (is_a_helper <varpool_node>::test): Likewise.
2119 (is_a_helper <cgraph_node>::test): Likewise.
2120 (varpool_variable_node): Likewise.
2121 (cgraph_function_or_thunk_node): Likewise.
2122 (varpool_alias_target): Likewise.
2123 (cgraph_alias_target): Likewise.
2124 (cgraph_node_name): Likewise.
2125 (varpool_node_asm_name): Likewise.
2126 (cgraph_node_asm_name): Likewise.
2127 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
2128 (cgraph_rebuild_references): Likewise.
2129 (rebuild_cgraph_edges): Likewise.
2130 (record_eh_tables): Likewise.
2131 (build_cgraph_edges): Likewise.
2132 (mark_store): Likewise.
2133 (mark_load): Likewise.
2134 (mark_address): Likewise.
2135 (record_type_list): Likewise.
2136 (record_reference): Likewise.
2137 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
2138 (cgraph_materialize_clone): Likewise.
2139 (cgraph_function_versioning): Likewise.
2140 (cgraph_copy_node_for_versioning): Likewise.
2141 (update_call_expr): Likewise.
2142 (cgraph_find_replacement_node): Likewise.
2143 (cgraph_create_virtual_clone): Likewise.
2144 (cgraph_clone_node): Likewise.
2145 * cgraphunit.c (compile): Likewise.
2146 (output_weakrefs): Likewise.
2147 (output_in_order): Likewise.
2148 (expand_function): Likewise.
2149 (assemble_thunks_and_aliases): Likewise.
2150 (expand_thunk): Likewise.
2151 (mark_functions_to_output): Likewise.
2152 (handle_alias_pairs): Likewise.
2153 (analyze_functions): Likewise.
2154 (walk_polymorphic_call_targets): Likewise.
2155 (varpool_finalize_decl): Likewise.
2156 (process_function_and_variable_attributes): Likewise.
2157 (cgraph_process_same_body_aliases): Likewise.
2158 (analyze_function): Likewise.
2159 (cgraph_add_new_function): Likewise.
2160 (cgraph_finalize_function): Likewise.
2161 (referred_to_p): Likewise.
2162 (cgraph_reset_node): Likewise.
2163 (cgraph_process_new_functions): Likewise.
2164 (enqueue_node): Likewise.
2165 (decide_is_symbol_needed): Likewise.
2166 * coverage.c (coverage_compute_profile_id): Likewise.
2167 * dbxout.c (dbxout_expand_expr): Likewise.
2168 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
2169 (reference_to_unused): Likewise.
2170 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2171 * gimplify.c (unvisit_body): Likewise.
2172 (unshare_body): Likewise.
2173 * ipa-cp.c (ipcp_generate_summary): Likewise.
2174 (ipcp_decision_stage): Likewise.
2175 (identify_dead_nodes): Likewise.
2176 (decide_whether_version_node): Likewise.
2177 (decide_about_value): Likewise.
2178 (perhaps_add_new_callers): Likewise.
2179 (create_specialized_node): Likewise.
2180 (update_profiling_info): Likewise.
2181 (ipcp_propagate_stage): Likewise.
2182 (estimate_local_effects): Likewise.
2183 (good_cloning_opportunity_p): Likewise.
2184 (devirtualization_time_bonus): Likewise.
2185 (propagate_constants_accross_call): Likewise.
2186 (initialize_node_lattices): Likewise.
2187 (ipcp_cloning_candidate_p): Likewise.
2188 (determine_versionability): Likewise.
2189 (print_all_lattices): Likewise.
2190 (print_lattice): Likewise.
2191 (ipcp_discover_new_direct_edges): Likewise.
2192 * ipa-devirt.c (ipa_devirt): Likewise.
2193 (likely_target_p): Likewise.
2194 (update_type_inheritance_graph): Likewise.
2195 (possible_polymorphic_call_target_p): Likewise.
2196 (dump_possible_polymorphic_call_targets): Likewise.
2197 (devirt_variable_node_removal_hook): Likewise.
2198 (record_binfo): Likewise.
2199 (maybe_record_node): Likewise.
2200 (build_type_inheritance_graph): Likewise.
2201 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2202 (inline_generate_summary): Likewise.
2203 (inline_analyze_function): Likewise.
2204 (do_estimate_growth): Likewise.
2205 (simple_edge_hints): Likewise.
2206 (estimate_node_size_and_time): Likewise.
2207 (estimate_edge_devirt_benefit): Likewise.
2208 (compute_inline_parameters): Likewise.
2209 (estimate_function_body_sizes): Likewise.
2210 (compute_bb_predicates): Likewise.
2211 (initialize_inline_failed): Likewise.
2212 (dump_inline_summary): Likewise.
2213 (dump_inline_edge_summary): Likewise.
2214 * ipa-inline-transform.c (inline_transform): Likewise.
2215 (preserve_function_body_p): Likewise.
2216 (save_inline_function_body): Likewise.
2217 (inline_call): Likewise.
2218 (clone_inlined_nodes): Likewise.
2219 (can_remove_node_now_p): Likewise.
2220 (can_remove_node_now_p_1): Likewise.
2221 * ipa-inline.c (early_inliner): Likewise.
2222 (early_inline_small_functions): Likewise.
2223 (inline_always_inline_functions): Likewise.
2224 (ipa_inline): Likewise.
2225 (flatten_function): Likewise.
2226 (inline_small_functions): Likewise.
2227 (speculation_useful_p): Likewise.
2228 (recursive_inlining): Likewise.
2229 (update_caller_keys): Likewise.
2230 (reset_edge_caches): Likewise.
2231 (update_edge_key): Likewise.
2232 (edge_badness): Likewise.
2233 (relative_time_benefit): Likewise.
2234 (want_inline_self_recursive_call_p): Likewise.
2235 (want_inline_small_function_p): Likewise.
2236 (want_early_inline_function_p): Likewise.
2237 (num_calls): Likewise.
2238 (can_early_inline_edge_p): Likewise.
2239 (can_inline_edge_p): Likewise.
2240 (report_inline_failed_reason): Likewise.
2241 * ipa-profile.c (ipa_profile): Likewise.
2242 (ipa_propagate_frequency): Likewise.
2243 (ipa_propagate_frequency_1): Likewise.
2244 (ipa_profile_generate_summary): Likewise.
2245 * ipa-prop.c (ipcp_transform_function): Likewise.
2246 (read_replacements_section): Likewise.
2247 (ipa_prop_read_section): Likewise.
2248 (ipa_modify_call_arguments): Likewise.
2249 (ipa_print_node_params): Likewise.
2250 (propagate_controlled_uses): Likewise.
2251 (update_indirect_edges_after_inlining): Likewise.
2252 (remove_described_reference): Likewise.
2253 (ipa_make_edge_direct_to_target): Likewise.
2254 (ipa_analyze_node): Likewise.
2255 (ipa_analyze_params_uses): Likewise.
2256 (ipa_compute_jump_functions): Likewise.
2257 (ipa_get_callee_param_type): Likewise.
2258 (ipa_print_node_jump_functions): Likewise.
2259 (ipa_initialize_node_params): Likewise.
2260 (ipa_populate_param_decls): Likewise.
2261 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
2262 (write_agg_replacement_chain): Likewise.
2263 (ipa_write_node_info): Likewise.
2264 (ipa_edge_duplication_hook): Likewise.
2265 (try_decrement_rdesc_refcount): Likewise.
2266 * ipa-pure-const.c (propagate_nothrow): Likewise.
2267 (propagate_pure_const): Likewise.
2268 (pure_const_read_summary): Likewise.
2269 (pure_const_write_summary): Likewise.
2270 (analyze_function): Likewise.
2271 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
2272 (ipa_ref_referring_ref_list): Likewise.
2273 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
2274 (ipa_remove_stmt_references): Likewise.
2275 (ipa_find_reference): Likewise.
2276 (ipa_dump_referring): Likewise.
2277 (ipa_dump_references): Likewise.
2278 (ipa_record_reference): Likewise.
2279 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
2280 (ipa_reference_write_optimization_summary): Likewise.
2281 (write_node_summary_p): Likewise.
2282 (propagate): Likewise.
2283 (read_write_all_from_decl): Likewise.
2284 (generate_summary): Likewise.
2285 (analyze_function): Likewise.
2286 (propagate_bits): Likewise.
2287 (ipa_reference_get_not_written_global): Likewise.
2288 (ipa_reference_get_not_read_global): Likewise.
2289 * ipa-split.c (execute_split_functions): Likewise.
2290 (split_function): Likewise.
2291 * ipa-utils.c (ipa_merge_profiles): Likewise.
2292 (dump_cgraph_node_set): Likewise.
2293 (ipa_reverse_postorder): Likewise.
2294 (ipa_edge_within_scc): Likewise.
2295 (ipa_get_nodes_in_cycle): Likewise.
2296 (ipa_free_postorder_info): Likewise.
2297 (ipa_reduced_postorder): Likewise.
2298 (searchc): Likewise.
2299 (recursive_call_p): Likewise.
2300 * ipa.c (ipa_cdtor_merge): Likewise.
2301 (record_cdtor_fn): Likewise.
2302 (function_and_variable_visibility): Likewise.
2303 (varpool_externally_visible_p): Likewise.
2304 (cgraph_externally_visible_p): Likewise.
2305 (comdat_can_be_unshared_p): Likewise.
2306 (comdat_can_be_unshared_p_1): Likewise.
2307 (address_taken_from_non_vtable_p): Likewise.
2308 (ipa_discover_readonly_nonaddressable_vars): Likewise.
2309 (symtab_remove_unreachable_nodes): Likewise.
2310 (walk_polymorphic_call_targets): Likewise.
2311 (process_references): Likewise.
2312 (enqueue_node): Likewise.
2313 (has_addr_references_p): Likewise.
2314 (cgraph_non_local_node_p_1): Likewise.
2315 * is-a.h (varpool_analyze_node): Likewise.
2316 * lto-cgraph.c (input_symtab): Likewise.
2317 (merge_profile_summaries): Likewise.
2318 (input_cgraph_1): Likewise.
2319 (input_edge): Likewise.
2320 (input_varpool_node): Likewise.
2321 (input_node): Likewise.
2322 (input_overwrite_node): Likewise.
2323 (compute_ltrans_boundary): Likewise.
2324 (output_refs): Likewise.
2325 (lto_output_varpool_node): Likewise.
2326 (lto_output_node): Likewise.
2327 (reachable_from_other_partition_p): Likewise.
2328 (referenced_from_other_partition_p): Likewise.
2329 (lto_output_edge): Likewise.
2330 (output_node_opt_summary): Likewise.
2331 (add_node_to): Likewise.
2332 (reachable_from_this_partition_p): Likewise.
2333 (lto_set_symtab_encoder_in_partition): Likewise.
2334 (lto_symtab_encoder_in_partition_p): Likewise.
2335 (lto_set_symtab_encoder_encode_initializer): Likewise.
2336 (lto_symtab_encoder_encode_initializer_p): Likewise.
2337 (lto_set_symtab_encoder_encode_body): Likewise.
2338 (lto_symtab_encoder_encode_body_p): Likewise.
2339 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
2341 * lto-streamer-in.c (lto_read_body): Likewise.
2342 (fixup_call_stmt_edges): Likewise.
2343 (fixup_call_stmt_edges_1): Likewise.
2344 * lto-streamer-out.c (produce_symtab): Likewise.
2345 (output_symbol_p): Likewise.
2346 (write_symbol): Likewise.
2347 (lto_output): Likewise.
2348 (copy_function): Likewise.
2349 (output_function): Likewise.
2350 * passes.c (function_called_by_processed_nodes_p): Likewise.
2351 (ipa_write_optimization_summaries): Likewise.
2352 (ipa_write_summaries): Likewise.
2353 (do_per_function_toporder): Likewise.
2354 (do_per_function): Likewise.
2355 (dump_passes): Likewise.
2356 * symtab.c (symtab_semantically_equivalent_p): Likewise.
2357 (symtab_nonoverwritable_alias): Likewise.
2358 (symtab_nonoverwritable_alias_1): Likewise.
2359 (symtab_for_node_and_aliases): Likewise.
2360 (symtab_resolve_alias): Likewise.
2361 (fixup_same_cpp_alias_visibility): Likewise.
2362 (symtab_alias_ultimate_target): Likewise.
2363 (symtab_used_from_object_file_p): Likewise.
2364 (verify_symtab_base): Likewise.
2365 (dump_symtab_base): Likewise.
2366 (symtab_node_name): Likewise.
2367 (symtab_node_asm_name): Likewise.
2368 (symtab_dissolve_same_comdat_group_list): Likewise.
2369 (symtab_add_to_same_comdat_group): Likewise.
2370 (symtab_unregister_node): Likewise.
2371 (symtab_insert_node_to_hashtable): Likewise.
2372 (symtab_register_node): Likewise.
2373 (unlink_from_assembler_name_hash): Likewise.
2374 (insert_to_assembler_name_hash): Likewise.
2375 (eq_assembler_name): Likewise.
2376 (hash_node_by_assembler_name): Likewise.
2377 (eq_node): Likewise.
2378 (hash_node): Likewise.
2379 * toplev.c (wrapup_global_declaration_2): Likewise.
2380 * trans-mem.c (ipa_tm_execute): Likewise.
2381 (ipa_tm_transform_clone): Likewise.
2382 (ipa_tm_transform_transaction): Likewise.
2383 (ipa_tm_transform_calls_redirect): Likewise.
2384 (ipa_tm_insert_gettmclone_call): Likewise.
2385 (ipa_tm_insert_irr_call): Likewise.
2386 (ipa_tm_create_version): Likewise.
2387 (ipa_tm_create_version_alias): Likewise.
2388 (ipa_tm_mark_forced_by_abi_node): Likewise.
2389 (ipa_tm_mark_force_output_node): Likewise.
2390 (ipa_tm_diagnose_tm_safe): Likewise.
2391 (ipa_tm_mayenterirr_function): Likewise.
2392 (ipa_tm_scan_irr_function): Likewise.
2393 (ipa_tm_note_irrevocable): Likewise.
2394 (ipa_tm_scan_calls_clone): Likewise.
2395 (get_cg_data): Likewise.
2396 * tree-eh.c (tree_could_trap_p): Likewise.
2397 * tree-emutls.c (ipa_lower_emutls): Likewise.
2398 (create_emultls_var): Likewise.
2399 (lower_emutls_function_body): Likewise.
2400 (gen_emutls_addr): Likewise.
2401 (emutls_decl): Likewise.
2402 (new_emutls_decl): Likewise.
2403 * tree-inline.c (tree_function_versioning): Likewise.
2404 (optimize_inline_calls): Likewise.
2405 (expand_call_inline): Likewise.
2406 (estimate_num_insns): Likewise.
2407 (copy_bb): Likewise.
2408 (delete_unreachable_blocks_update_callgraph): Likewise.
2409 * tree-nested.c (gimplify_all_functions): Likewise.
2410 (create_nesting_tree): Likewise.
2411 (check_for_nested_with_variably_modified): Likewise.
2412 * tree-pretty-print.c (dump_function_header): Likewise.
2413 * tree-profile.c (tree_profiling): Likewise.
2414 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
2415 (modify_function): Likewise.
2416 (convert_callers): Likewise.
2417 (convert_callers_for_node): Likewise.
2418 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
2419 (associate_varinfo_to_alias): Likewise.
2420 (create_variable_info_for): Likewise.
2421 (get_constraint_for_ssa_var): Likewise.
2422 * tree-vectorizer.c (increase_alignment): Likewise.
2423 * tree.c (find_decls_types_in_var): Likewise.
2424 (find_decls_types_in_node): Likewise.
2425 (free_lang_data_in_decl): Likewise.
2426 * value-prof.c (gimple_ic_transform): Likewise.
2427 (gimple_ic): Likewise.
2428 (check_ic_target): Likewise.
2429 (init_node_map): Likewise.
2430 * varasm.c (decl_binds_to_current_def_p): Likewise.
2431 (default_binds_local_p_1): Likewise.
2432 (dump_tm_clone_pairs): Likewise.
2433 (assemble_alias): Likewise.
2434 (find_decl): Likewise.
2435 (mark_decl_referenced): Likewise.
2436 * varpool.c (varpool_for_node_and_aliases): Likewise.
2437 (varpool_extra_name_alias): Likewise.
2438 (varpool_create_variable_alias): Likewise.
2439 (add_new_static_var): Likewise.
2440 (varpool_finalize_named_section_flags): Likewise.
2441 (varpool_remove_unreferenced_decls): Likewise.
2442 (enqueue_node): Likewise.
2443 (varpool_assemble_decl): Likewise.
2444 (assemble_aliases): Likewise.
2445 (varpool_analyze_node): Likewise.
2446 (cgraph_variable_initializer_availability): Likewise.
2447 (varpool_add_new_variable): Likewise.
2448 (ctor_for_folding): Likewise.
2449 (dump_varpool_node): Likewise.
2450 (varpool_remove_initializer): Likewise.
2451 (varpool_remove_node): Likewise.
2452 (varpool_node_for_decl): Likewise.
2453 (varpool_create_empty_node): Likewise.
2454 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
2455 (ix86_get_function_versions_dispatcher): Likewise.
2457 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2459 * cgraph.h (symtab_node_base): Convert to a class;
2460 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
2461 chain_next/prev from symtab_node_def.
2462 (cgraph_node): Inherit from symtab_node; add GTY option
2463 tag ("SYMTAB_FUNCTION").
2464 (varpool_node): Inherit from symtab_node; add GTY option
2465 tag ("SYMTAB_VARIABLE").
2466 (symtab_node_def): Remove.
2467 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
2468 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
2469 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
2470 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
2472 * ipa-ref.h (symtab_node_def): Drop.
2473 (symtab_node): Change underlying type from symtab_node_def to
2475 (const_symtab_node): Likwise.
2477 * is-a.h: Update examples in comment.
2479 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
2480 (assembler_name_hash): Likewise.
2482 2013-10-29 Martin Liska <marxin.liska@gmail.com>
2484 * doc/tree-ssa.texi (gimple_phi_result): Document.
2485 (gimple_phi_num_args, gimple_phi_arg): Likewise.
2486 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
2487 (PHI_RESULT, PHI_NUM_ARGS): Remove.
2488 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
2490 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2492 * expr.h: Revert change and include tree-core.h.
2493 * rtl.h: Revert change and don't include tree-core.h.
2495 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2497 * config/darwin.c: Include gimple.h.
2498 * config/i386/winnt.c: Likewise.
2500 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
2502 PR tree-optimization/19831
2503 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
2505 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2507 * tree-outof-ssa.h: Remove include files.
2508 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
2510 * tree-ssa-coalesce.c: Likewise.
2511 * cfgexpand.c: Likewise.
2512 * tree-ssa-ter.c: Likewise.
2513 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
2514 * lto-streamer.h: Likewise.
2515 * cgraphbuild.c: Add gimple.h to include list.
2516 * data-streamer-in.c: Likewise.
2517 * ipa-cp.c: Likewise.
2518 * tree-streamer.c: Likewise.
2519 * lto-compress.c: Likewise.
2520 * ipa-reference.c: Likewise.
2521 * data-streamer-out.c: Likewise.
2522 * lto-cgraph.c: Likewise.
2523 * cgraphclones.c: Likewise.
2524 * ipa-utils.c: Likewise.
2525 * data-streamer.c: Likewise.
2526 * ipa-split.c: Likewise.
2527 * lto-section-in.c: Likewise.
2528 * tree-streamer-out.c: Likewise.
2529 * ipa-prop.c: Likewise.
2530 * tree-streamer-in.c: Likewise.
2531 * symtab.c: Likewise.
2532 * opts-global.c: Likewise.
2533 * lto-opts.c: Likewise.
2534 * lto-section-out.c: Likewise.
2535 * lto-streamer.c: Likewise.
2536 * rtl.h: Add tree-core.h to include list.
2537 * expr.h: Remove tree-core.h from include list.
2538 * gimple.h: Likewise.
2539 * ipa-utils.h: Likewise.
2540 * streamer-hooks.h: Likewise.
2541 * streamer-hooks.c: Include input.h.
2543 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2545 * config/arm/arm.c (cortexa7_extra_costs): New table.
2546 (arm_cortex_a7_tune): New.
2547 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
2549 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
2551 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
2553 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2555 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
2556 to use autogenerated markers for a class-hierarchy, every class
2557 must have a GTY marker.
2558 * gengtype.h (struct type): Add linked list of subclasses to
2559 the "s" member of the union.
2560 (add_subclass): New decl.
2561 * gengtype-state.c (read_state_struct_type): Set up subclass
2563 * gengtype.c (get_ultimate_base_class): New.
2564 (add_subclass): New.
2565 (new_structure): Set up subclass linked list.
2566 (set_gc_used_type): Propagate usage information to subclasses.
2567 (output_mangled_typename): Use get_ultimate_base_class.
2568 (walk_subclasses): Use the subclass linked list, avoiding an
2569 O(N^2) when writing out all types.
2570 (walk_type): Issue an error if the base class is missing a tag,
2571 rather than generating bogus C code. Add a gcc_unreachable
2572 default case, in case people omit tags from concrete subclasses,
2573 or get the values wrong.
2574 (write_func_for_structure): Issue an error for subclasses for
2575 which the base doesn't have a "desc", since otherwise the
2576 autogenerated routines for the base would silently fail to visit
2577 any subclass fields.
2578 (write_root): Use get_ultimate_base_class, tweaking constness of
2579 tp to match that function's signature.
2581 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2583 * doc/gty.texi (GTY Options): Add note about inheritance to
2584 description of desc and tag.
2585 (Inheritance and GTY): New.
2587 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2589 * gengtype-parse.c (opts_have): Drop "static" so that
2590 we can use this from gengtype.c.
2591 * gengtype.c (set_gc_used_type): Mark any base class as used;
2592 update field traversal to visit inherited fields.
2593 (output_mangled_typename): Convert references to classes within
2594 an inheritance hierarchy to reference the ultimate base class,
2595 since only it will have gt_ functions.
2596 (get_string_option): New.
2597 (walk_subclasses): New.
2598 (walk_type): Treat GTY structs that have a "desc" as being the
2599 root of an inheritance hierarchy. Generate a switch on it
2600 within the marking function which walks all subclasses, adding
2601 cases for them via walk_subclasses. For subclasses, visit all
2602 fields of the type (including inherited ones).
2603 (write_func_for_structure): Don't write fns for subclasses, only
2604 for the ultimate base class within an inheritance hierarchy.
2605 Subclasses-marking will be handled by the base class marking functions.
2606 (write_types): Likewise.
2607 (write_local_func_for_structure): Likewise.
2608 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
2609 a "tag" option (and are thus concrete subclasses).
2610 (write_root): Use the marker function for the ultimate base class.
2611 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
2612 (opts_have): Add declaration.
2614 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
2616 * lra-spills.c (lra_final_code_change): Remove useless move insns
2617 originated from moves of pseudos.
2619 2013-10-28 Jeff Law <law@redhat.com>
2621 * lower-subreg.c (resolve_simple_move): Fix comment typo.
2623 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
2625 * df-scan.c (df_collection_rec): Adjust.
2626 (copy_defs): New constant.
2627 (copy_uses): Likewise.
2628 (copy_eq_uses): Likewise.
2629 (copy_mw): Likewise.
2630 (copy_all): Likewise.
2631 (df_insn_rescan): Adjust.
2632 (df_notes_rescan): Likewise.
2633 (df_swap_refs): Likewise.
2634 (df_sort_and_compress_refs): Likewise.
2635 (df_sort_and_compress_mws): Likewise.
2636 (df_install_refs): Likewise.
2637 (df_install_mws): Likewise.
2638 (df_refs_add_to_chains): Add flags parameter controlling which vectors
2640 (df_bb_refs_record): Adjust.
2641 (df_record_entry_block_defs): Likewise.
2642 (df_record_exit_block_defs): Likewise.
2643 (df_refs_verify): Likewise.
2644 (df_mws_verify): Likewise.
2645 (df_insn_refs_verify): Likewise.
2646 (df_bb_verify): Likewise.
2647 * ipa-pure-const.c (finish_state): Remove.
2648 (propagate): Adjust.
2649 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
2650 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
2651 var-tracking.c: Adjust.
2652 * vec.c (stack_vecs): Remove.
2653 (register_stack_vec): Likewise.
2654 (stack_vec_register_index): Likewise.
2655 (unregister_stack_vec): Likewise.
2656 * vec.h (struct va_stack): Remove.
2657 (struct vec<T, A, vl_ptr>): Specialize as
2658 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
2659 allocation strategy compatable with the vl_ptr layout.
2660 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
2661 specialization anyway.
2662 (class stack_vec): New class.
2663 (vec_stack_alloc): Remove.
2664 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
2666 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2667 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2668 Sergey Lega <sergey.s.lega@intel.com>
2669 Anna Tikhonova <anna.tikhonova@intel.com>
2670 Ilya Tocar <ilya.tocar@intel.com>
2671 Andrey Turetskiy <andrey.turetskiy@intel.com>
2672 Ilya Verbin <ilya.verbin@intel.com>
2673 Kirill Yukhin <kirill.yukhin@intel.com>
2674 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2676 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
2677 (*prefetch_avx512pf_<mode>): New.
2678 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
2679 (avx512f_maskcmp<mode>3): Ditto.
2680 (vashrv16si3): Ditto.
2682 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2683 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2684 Sergey Lega <sergey.s.lega@intel.com>
2685 Anna Tikhonova <anna.tikhonova@intel.com>
2686 Ilya Tocar <ilya.tocar@intel.com>
2687 Andrey Turetskiy <andrey.turetskiy@intel.com>
2688 Ilya Verbin <ilya.verbin@intel.com>
2689 Kirill Yukhin <kirill.yukhin@intel.com>
2690 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2692 * config/i386/i386.md (any_truncate): New.
2693 (trunsuffix): Ditto.
2694 * config/i386/predicates.md (const_8_to_9_operand): New.
2695 (const_10_to_11_operand): Ditto.
2696 (const_12_to_13_operand): Ditto.
2697 (const_14_to_15_operand): Ditto.
2698 (const_16_to_19_operand): Ditto.
2699 (const_20_to_23_operand): Ditto.
2700 (const_24_to_27_operand): Ditto.
2701 (const_28_to_31_operand): Ditto.
2702 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
2703 (cvtusi2<ssescalarmodesuffix>32): New.
2704 (cvtusi2<ssescalarmodesuffix>64): Ditto.
2705 (ufloatv16siv16sf2): Ditto.
2706 (avx512f_fix_notruncv16sfv16si): Ditto.
2707 (avx512f_ufix_notruncv16sfv16si): Ditto.
2708 (avx512f_vcvtss2usi): Ditto.
2709 (avx512f_vcvtss2usiq): Ditto.
2710 (avx512f_vcvttss2usi): Ditto.
2711 (avx512f_vcvttss2usiq): Ditto.
2712 (avx512f_vcvtsd2usi): Ditto.
2713 (avx512f_vcvtsd2usiq): Ditto.
2714 (avx512f_vcvttsd2usi): Ditto.
2715 (avx512f_vcvttsd2usiq): Ditto.
2716 (ufloatv8siv8df): Ditto.
2717 (avx512f_cvtdq2pd512_2): Ditto.
2718 (avx512f_cvtpd2dq512): Ditto.
2719 (avx512f_ufix_notruncv8dfv8si): Ditto.
2720 (avx512f_cvtpd2ps512): Ditto.
2721 (vec_unpacks_lo_v16sf): Ditto.
2722 (vec_unpacks_hi_v16sf): Ditto.
2723 (vec_unpacks_float_hi_v16si): Ditto.
2724 (vec_unpacks_float_lo_v16si): Ditto.
2725 (avx512f_unpckhps512): Ditto.
2726 (avx512f_unpcklps512): Ditto.
2727 (avx512f_movshdup512): Ditto.
2728 (avx512f_movsldup512): Ditto.
2729 (vec_extract_lo_v32hi): Ditto.
2730 (vec_extract_hi_v32hi): Ditto.
2731 (vec_extract_lo_v64qi): Ditto.
2732 (vec_extract_hi_v64qi): Ditto.
2733 (avx512f_unpckhpd512): Ditto.
2734 (avx512f_movddup512): Ditto.
2735 (avx512f_unpcklpd512): Ditto.
2736 (*avx512f_unpcklpd512): Ditto.
2737 (avx512f_shufps512_1): Ditto.
2738 (avx512f_shufpd512_1): Ditto.
2739 (avx512f_interleave_highv8di): Ditto.
2740 (avx512f_interleave_lowv8di): Ditto.
2741 (PMOV_DST_MODE): Ditto.
2742 (pmov_src_mode): Ditto.
2743 (pmov_src_lower): Ditto.
2745 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
2746 (*avx512f_<code>v8div16qi2): Ditto.
2747 (*avx512f_<code>v8div16qi2_store): Ditto.
2748 (vec_widen_umult_even_v16si): Ditto.
2749 (*vec_widen_umult_even_v16si): Ditto.
2750 (vec_widen_smult_even_v16si): Ditto.
2751 (*vec_widen_smult_even_v16si): Ditto.
2752 (avx512f_interleave_highv16si): Ditto.
2753 (avx512f_interleave_lowv16si): Ditto.
2754 (avx512f_<code>v16qiv16si2): Ditto.
2755 (avx512f_<code>v16hiv16si2): Ditto.
2756 (avx512f_<code>v8qiv8di2): Ditto.
2757 (avx512f_<code>v8hiv8di2): Ditto.
2758 (avx512f_<code>v8siv8di2): Ditto.
2759 (avx512cd_maskb_vec_dupv8di): Ditto.
2760 (avx512cd_maskw_vec_dupv16si): Ditto.
2761 (avx512f_vcvtph2ps512): Ditto.
2762 (avx512f_vcvtps2ph512): Ditto.
2763 (VEC_EXTRACT_MODE): Extened with wider modes.
2764 (VEC_PERM_AVX2): Ditto.
2765 (VEC_PERM_CONST): Ditto.
2767 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
2769 * config/arc/arc.c (arc_ccfsm_post_advance):
2770 Add comment about TYPE_RETURN.
2772 2013-10-28 Bin Cheng <bin.cheng@arm.com>
2774 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
2775 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
2776 (strip_offset): Convert offset to unsigned number.
2778 2013-10-27 Tom de Vries <tom@codesourcery.com>
2780 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
2781 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
2783 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
2785 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
2786 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
2787 (TARGET_BIG_ENDIAN): New macro.
2788 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
2789 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
2790 * config/sh/sh.c: Likewise.
2791 * config/sh/sh.h: Likewise.
2793 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
2795 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
2796 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
2798 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2800 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
2802 (cmpstr_t, cmpstrsi): Fix formatting.
2804 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2807 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
2808 addressing, do not use general_operand for memory operands.
2810 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
2813 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2814 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2816 2013-10-26 Jeff Law <law@redhat.com>
2818 * predict.c (PRED_MUDFLAP): Remove.
2819 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
2821 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
2822 (OBJS): Remove tree-nomudflap.o
2823 (GTFILES): Remove tree-mudflap.c
2824 * builtins.c (expand_builtin_alloc): Remove mudflap support.
2825 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2826 (mfwrap_spec, mflib_spec): Likewise.
2827 (cpp_unique_options, cc1_options, static_specs): Likewise.
2828 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
2829 * passes.def: Likewise.
2830 * toplev.c (compile_file, process_options): Likewise.
2831 * tree-inline.c (copy_tree_r): Likewise.
2832 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
2833 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
2834 (build_constant_desc, output_constant_def_contents): Likewise.
2835 (categorize_decl_for_section): Likewise.
2836 * tree-mudflap.c: Removed.
2837 * tree-mudflap.h: Removed.
2838 * tree-nomudflap.c: Removed.
2839 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
2840 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
2841 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2842 * config/sol2.h (MFLIB_SPEC): Likewise.
2843 * doc/install.texi: Remove mudflap references.
2844 * doc/passes.texi: Similarly.
2845 * doc/sourcebuild.texi: Similarly.
2846 * doc/invoke.texi: Remove mudlfap related options.
2848 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2850 PR rtl-optimization/58759
2851 * lra-constraints.c (lra_constraints): Remove wrong condition to
2852 remove insn setting up an equivalent pseudo.
2854 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2856 * config/rs6000/rs6000-protos.h
2857 (rs6000_secondary_memory_needed_mode): New prototype.
2858 * config/rs6000/rs6000.c: Include ira.h.
2859 (TARGET_LRA_P): Redefine.
2860 (rs6000_legitimate_offset_address_p): Call
2861 legitimate_constant_pool_address_p in strict mode for LRA.
2862 (rs6000_legitimate_address_p): Ditto.
2863 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2864 (rs6000_emit_move): Add LRA version of code to generate load/store
2866 (rs6000_secondary_memory_needed_mode): New.
2867 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2868 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2869 constants, memory, and FP registers.
2870 (rs6000_lra_p): New.
2871 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2872 * config/rs6000/rs6000.opt (mlra): New option.
2873 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2875 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2877 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2878 has_single_use () and not do the conversion if has_single_use ()
2879 returns false for the multiplication result.
2881 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2883 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2884 to "NODE", since this works on a "tree", not an
2886 (CONSTANT_CLASS_P): Likewise.
2889 (INDIRECT_REF_P): Likewise.
2890 (REFERENCE_CLASS_P): Likewise.
2891 (COMPARISON_CLASS_P): Likewise.
2892 (UNARY_CLASS_P): Likewise.
2893 (BINARY_CLASS_P): Likewise.
2894 (STATEMENT_CLASS_P): Likewise.
2895 (VL_EXP_CLASS_P): Likewise.
2896 (EXPRESSION_CLASS_P): Likewise.
2897 (IS_TYPE_OR_DECL_P): Likewise.
2899 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2901 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2902 ADDR_EXPR in the defining statement.
2904 2013-10-25 Richard Biener <rguenther@suse.de>
2906 PR tree-optimization/58626
2907 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2908 anti_dd, output_dd and input_dd.
2909 (struct rdg_edge): Remove level and relation members.
2910 (RDGE_LEVEL, RDGE_RELATION): Remove.
2911 (dot_rdg_1): Adjust.
2912 (create_rdg_edge_for_ddr): Remove.
2913 (create_rdg_edges_for_scalar): Adjust.
2914 (create_edge_for_control_dependence): Likewise.
2915 (create_rdg_edges): Split into ...
2916 (create_rdg_flow_edges): ... this
2917 (create_rdg_cd_edges): ... and this.
2919 (build_rdg): Likewise, do not compute data dependences or
2921 (pg_add_dependence_edges): New function.
2923 (distribute_loop): First apply all non-dependence based
2924 partition mergings. Then compute dependences between partitions
2925 and merge and order partitions according to them.
2927 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2929 PR rtl-optimization/58831
2930 * alias.c (init_alias_analysis): At the beginning of each iteration,
2931 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2933 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2935 * recog.c (search_ofs): New static variable moved from...
2936 (peep2_find_free_register): ...here.
2937 (peephole2_optimize): Initialize it.
2939 2013-10-25 Tobias Burnus <burnus@net-b.de>
2941 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2943 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2945 * config/i386/i386.h (TARGET_MPX): New define.
2946 (TARGET_MPX_P): Ditto.
2948 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2950 * config/i386/constraints.md (B): New.
2953 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2954 * config/i386/i386-modes.def (BND32): New.
2956 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2957 * config/i386/i386.c (isa_opts): Add mmpx.
2958 (regclass_map): Add bound registers.
2959 (dbx_register_map): Likewise.
2960 (dbx64_register_map): Likewise.
2961 (svr4_dbx_register_map): Likewise.
2963 (ix86_option_override_internal): Support MPX ISA.
2964 (ix86_conditional_register_usage): Support bound registers.
2965 (print_reg): Likewise.
2966 (ix86_code_end): Add MPX bnd prefix.
2967 (output_set_got): Likewise.
2968 (ix86_output_call_insn): Likewise.
2969 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2970 (ix86_print_operand_punct_valid_p): Likewise.
2971 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2973 (ix86_class_likely_spilled_p): Add bound regs support.
2974 (ix86_hard_regno_mode_ok): Likewise.
2975 (x86_order_regs_for_local_alloc): Likewise.
2976 (ix86_bnd_prefixed_insn_p): New.
2977 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2978 (FIXED_REGISTERS): Add bound registers.
2979 (CALL_USED_REGISTERS): Likewise.
2980 (REG_ALLOC_ORDER): Likewise.
2981 (HARD_REGNO_NREGS): Likewise.
2983 (VALID_BND_REG_MODE): New.
2984 (FIRST_BND_REG): New.
2985 (LAST_BND_REG): New.
2986 (reg_class): Add BND_REGS.
2987 (REG_CLASS_NAMES): Likewise.
2988 (REG_CLASS_CONTENTS): Likewise.
2990 (ANY_BND_REG_P): New.
2992 (HI_REGISTER_NAMES): Add bound registers.
2993 * config/i386/i386.md (UNSPEC_BNDMK): New.
2994 (UNSPEC_BNDMK_ADDR): New.
2995 (UNSPEC_BNDSTX): New.
2996 (UNSPEC_BNDLDX): New.
2997 (UNSPEC_BNDLDX_ADDR): New.
2998 (UNSPEC_BNDCL): New.
2999 (UNSPEC_BNDCU): New.
3000 (UNSPEC_BNDCN): New.
3001 (UNSPEC_MPX_FENCE): New.
3004 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
3005 (length_immediate): Likewise.
3006 (prefix_0f): Likewise.
3008 (prefix_rep): Check for bnd prefix.
3009 (length_nobnd): New.
3010 (length): Use length_nobnd if specified.
3015 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
3018 (simple_return_internal): Likewise.
3019 (simple_return_pop_internal): Likewise.
3020 (*indirect_jump): Add MPX bnd prefix.
3021 (*tablejump_1): Likewise.
3022 (simple_return_internal_long): Likewise.
3023 (simple_return_indirect_internal): Likewise.
3027 (*mov<mode>_internal_mpx): New.
3028 (<mode>_<bndcheck>): New.
3029 (*<mode>_<bndcheck>): New.
3034 * config/i386/predicates.md (lea_address_operand): Rename to...
3035 (address_no_seg_operand): ... this.
3036 (address_mpx_no_base_operand): New.
3037 (address_mpx_no_index_operand): New.
3038 (bnd_mem_operator): New.
3039 * config/i386/i386.opt (mmpx): New.
3040 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
3041 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
3043 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
3045 * mode-classes.def (MODE_POINTER_BOUNDS): New.
3046 * tree.def (POINTER_BOUNDS_TYPE): New.
3047 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
3048 (POINTER_BOUNDS_MODE): New.
3049 (make_pointer_bounds_mode): New.
3050 * machmode.h (POINTER_BOUNDS_MODE_P): New.
3051 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
3052 (layout_type): Support POINTER_BOUNDS_TYPE.
3053 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
3054 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
3055 (type_contains_placeholder_1): Likewise.
3056 * tree.h (POINTER_BOUNDS_TYPE_P): New.
3057 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
3058 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
3060 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
3062 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
3063 mode of address computation when calling memory_address_addr_space.
3065 2013-08-24 Richard Henderson <rth@twiddle.net>
3068 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
3069 instead of create_convert_operand_to.
3070 (maybe_emit_sync_lock_test_and_set): Likewise.
3071 (expand_atomic_compare_and_swap): Likewise.
3072 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
3074 2013-08-24 Sriraman Tallam <tmsriram@google.com>
3076 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
3077 Change param_is to use the struct and not the pointer to the struct.
3079 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
3081 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
3082 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
3083 Move to targhooks.c.
3084 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
3085 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
3086 Relocate from gimplify.c.
3087 * targhooks.h: Add 2 prototypes.
3088 * tree.h: Delete 2 prototypes.
3090 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
3092 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
3093 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
3095 2013-10-24 Cong Hou <congh@google.com>
3097 * convert.c (convert_to_real): Guard those unsafe math function
3098 convertions with flag_unsafe_math_optimizations. Handle sqrt()
3101 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
3104 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
3106 (cgraph_create_edge): Use the new argument.
3107 (cgraph_create_indirect_edge): Likewise.
3109 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
3111 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
3113 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
3114 changes statep->state.
3116 2013-10-24 Tobias Burnus <burnus@net-b.de>
3119 * tree-cfg.c (replace_loop_annotate): New function.
3120 (execute_build_cfg): Call it.
3121 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
3122 * internal-fn.c (expand_ANNOTATE): New function.
3123 * internal-fn.def (ANNOTATE): Define as new internal function.
3124 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
3125 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
3126 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
3127 * doc/extend.texi (Pragmas): Document #pragma ivdep.
3128 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
3130 2013-10-17 Ian Bolton <ian.bolton@arm.com>
3131 Marcus Shawcroft <marcus.shawcroft@arm.com>
3133 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
3134 Special case reload SP+C into none GENERAL_REGS.
3136 2013-10-24 Michael Matz <matz@suse.de>
3138 * gengtype.c (is_file_equal): Check that files will be same length.
3140 2013-10-25 Christian Bruel <christian.bruel@st.com>
3142 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
3143 * config/sh/t-sh (sh-mem.o): New rule.
3144 * config/sh/sh-mem.cc (expand_block_move): Moved here.
3145 (sh_expand_cmpstr): New function.
3146 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
3147 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
3148 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
3149 (rotlhi3_8): Rename.
3151 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3153 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
3154 * Makefile.in (CXXDEPMODE): Assign and change users.
3155 (CCDEPMODE): Delete.
3156 * configure: Regenerate.
3158 2013-10-23 David Malcolm <dmalcolm@redhat.com>
3160 * gengtype-parse.c (require_without_advance): New.
3161 (type): For GTY-marked types that are not GTY((user)), parse any
3162 base classes, requiring them to be single-inheritance, and not
3163 be templates. For non-GTY-marked types and GTY((user)),
3164 continue to skip over any C++ inheritance specification.
3165 * gengtype-state.c (state_writer::write_state_struct_type):
3166 Write base class of type (if any).
3167 (read_state_struct_type): Read base class of type (if any).
3168 * gengtype.c (new_structure): Add a "base_class" parameter.
3169 (create_optional_field_): Update for new parameter to new_structure.
3170 (adjust_field_rtx_def): Likewise.
3171 (adjust_field_tree_exp): Likewise.
3172 * gengtype.h (struct type): Add "base_class" field to the s
3174 (new_structure): Add "base" parameter.
3176 2013-10-23 Sriraman Tallam <tmsriram@google.com>
3179 * config/i386/i386.c (ix86_option_override_internal):
3180 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
3181 (ix86_valid_target_attribute_tree):
3182 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
3183 Change TARGET_SSE to TARGET_SSE_P (opts->...)
3185 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
3187 * tree-ssa-loop.h: Remove include files.
3188 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
3189 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
3190 * graphite-clast-to-gimple.c: Likewise.
3191 * graphite-scop-detection.c: Likewise.
3192 * graphite-sese-to-poly.c: Likewise.
3193 * ipa-inline-analysis.c: Likewise.
3194 * ipa-pure-const.c: Likewise.
3195 * loop-init.c: Likewise.
3196 * passes.c: Likewise.
3197 * predict.c: Likewise.
3198 * tree-cfg.c: Likewise.
3199 * tree-cfgcleanup.c: Likewise.
3200 * tree-chrec.c: Likewise.
3201 * tree-data-ref.c: Likewise.
3202 * tree-loop-distribution.c: Likewise.
3203 * tree-parloops.c: Likewise.
3204 * tree-predcom.c: Likewise.
3205 * tree-scalar-evolution.c: Likewise.
3206 * tree-ssa-address.c: Likewise.
3207 * tree-ssa.c: Likewise.
3208 * tree-ssa-dce.c: Likewise.
3209 * tree-ssa-loop.c: Likewise.
3210 * tree-ssa-loop-im.c: Likewise.
3211 * tree-ssa-loop-ivcanon.c: Likewise.
3212 * tree-ssa-loop-ivopts.c: Likewise.
3213 * tree-ssa-loop-manip.c: Likewise.
3214 * tree-ssa-loop-niter.c: Likewise.
3215 * tree-ssa-loop-prefetch.c: Likewise.
3216 * tree-ssa-loop-unswitch.c: Likewise.
3217 * tree-ssa-reassoc.c: Likewise.
3218 * tree-vect-data-refs.c: Likewise.
3219 * tree-vect-loop.c: Likewise.
3220 * tree-vect-loop-manip.c: Likewise.
3221 * tree-vectorizer.c: Likewise.
3222 * tree-vect-stmts.c: Likewise.
3223 * tree-vrp.c: Likewise.
3225 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3227 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
3229 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3231 PR tree-optimization/58775
3232 PR tree-optimization/58791
3233 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
3234 (insert_stmt_after): Rewritten, don't move the stmt, but really
3236 (get_stmt_uid_with_default): Remove.
3237 (build_and_add_sum): Use insert_stmt_after and
3238 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
3239 (update_range_test): Set uid on stmts added by
3240 force_gimple_operand_gsi. Don't immediately modify statements
3241 in inter-bb optimization, just update oe->op values.
3242 (optimize_range_tests): Return bool whether any changed have been made.
3243 (update_ops): New function.
3244 (struct inter_bb_range_test_entry): New type.
3245 (maybe_optimize_range_tests): Perform statement changes here.
3246 (not_dominated_by, appears_later_in_bb, get_def_stmt,
3247 ensure_ops_are_available): Remove.
3248 (find_insert_point): Rewritten.
3249 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
3250 return LHS of the (new resp. old) stmt. Don't call
3251 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
3252 instead of last, move new stmt at the right place.
3253 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
3254 (negate_value): Likewise. Set uids.
3255 (break_up_subtract_bb): Initialize uids.
3256 (reassociate_bb): Adjust rewrite_expr_tree caller.
3257 (do_reassoc): Don't call renumber_gimple_stmt_uids.
3259 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
3262 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
3263 TARGET_32BIT final condition.
3264 (mulsi3_internal2 and splitter): Same.
3266 2013-10-23 Jeff Law <law@redhat.com>
3268 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
3269 through joiner blocks with abnormal outgoing edges.
3271 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
3272 Add parameter JOINERS, to allow/disallow threading through joiner
3274 (thread_block): New. Call thread_block_1.
3275 (mark_threaded_blocks): Remove code to filter out certain cases
3276 of threading through joiner blocks.
3277 (thread_through_all_blocks): Document how we can have a dangling
3278 edge AUX field and clear it.
3280 2013-10-23 Ian Lance Taylor <iant@google.com>
3282 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
3283 (C++ Dialect Options): Likewise.
3284 (Optimize Options): Likewise.
3286 2013-10-23 Tom de Vries <tom@codesourcery.com>
3288 PR tree-optimization/58805
3289 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
3291 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3293 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
3294 sequence based on get_range_info returned range.
3296 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
3298 * tree-ssa.h: Remove all #include's
3299 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
3300 * alias.c: Move required includes from tree-ssa.h.
3302 * builtins.c: Likewise.
3303 * calls.c: Likewise.
3304 * cfgexpand.c: Likewise.
3305 * cfghooks.c: Likewise.
3306 * cfgloop.c: Likewise.
3307 * cfgloopmanip.c: Likewise.
3308 * cgraph.c: Likewise.
3309 * cgraphbuild.c: Likewise.
3310 * cgraphclones.c: Likewise.
3311 * cgraphunit.c: Likewise.
3313 * except.c: Likewise.
3315 * final.c: Likewise.
3316 * fold-const.c: Likewise.
3317 * ggc-page.c: Likewise.
3318 * gimple-builder.c: Likewise.
3319 * gimple-fold.c: Likewise.
3320 * gimple-iterator.c: Likewise.
3321 * gimple-low.c: Likewise.
3322 * gimple-pretty-print.c: Likewise.
3323 * gimple-ssa-strength-reduction.c: Likewise.
3324 * gimple-streamer-in.c: Likewise.
3325 * gimple-streamer-out.c: Likewise.
3326 * gimplify.c: Likewise.
3327 * graphite-blocking.c: Likewise.
3328 * graphite-clast-to-gimple.c: Likewise.
3329 * graphite-dependences.c: Likewise.
3330 * graphite-interchange.c: Likewise.
3331 * graphite-optimize-isl.c: Likewise.
3332 * graphite-poly.c: Likewise.
3333 * graphite-scop-detection.c: Likewise.
3334 * graphite-sese-to-poly.c: Likewise.
3335 * graphite.c: Likewise.
3336 * ipa-cp.c: Likewise.
3337 * ipa-inline-analysis.c: Likewise.
3338 * ipa-inline-transform.c: Likewise.
3339 * ipa-inline.c: Likewise.
3340 * ipa-prop.c: Likewise.
3341 * ipa-pure-const.c: Likewise.
3342 * ipa-reference.c: Likewise.
3343 * ipa-split.c: Likewise.
3344 * ipa-utils.c: Likewise.
3345 * loop-init.c: Likewise.
3346 * lto-cgraph.c: Likewise.
3347 * lto-section-in.c: Likewise.
3348 * lto-section-out.c: Likewise.
3349 * lto-streamer-in.c: Likewise.
3350 * lto-streamer-out.c: Likewise.
3351 * lto-streamer.c: Likewise.
3352 * omp-low.c: Likewise.
3353 * passes.c: Likewise.
3354 * predict.c: Likewise.
3355 * print-tree.c: Likewise.
3356 * profile.c: Likewise.
3358 * targhooks.c: Likewise.
3359 * tracer.c: Likewise.
3360 * trans-mem.c: Likewise.
3361 * tree-call-cdce.c: Likewise.
3362 * tree-cfg.c: Likewise.
3363 * tree-cfgcleanup.c: Likewise.
3364 * tree-chrec.c: Likewise.
3365 * tree-complex.c: Likewise.
3366 * tree-data-ref.c: Likewise.
3367 * tree-dfa.c: Likewise.
3368 * tree-eh.c: Likewise.
3369 * tree-emutls.c: Likewise.
3370 * tree-if-conv.c: Likewise.
3371 * tree-inline.c: Likewise.
3372 * tree-into-ssa.c: Likewise.
3373 * tree-loop-distribution.c: Likewise.
3374 * tree-mudflap.c: Likewise.
3375 * tree-nested.c: Likewise.
3376 * tree-nrv.c: Likewise.
3377 * tree-object-size.c: Likewise.
3378 * tree-outof-ssa.c: Likewise.
3379 * tree-parloops.c: Likewise.
3380 * tree-phinodes.c: Likewise.
3381 * tree-predcom.c: Likewise.
3382 * tree-pretty-print.c: Likewise.
3383 * tree-profile.c: Likewise.
3384 * tree-scalar-evolution.c: Likewise.
3385 * tree-sra.c: Likewise.
3386 * tree-ssa-address.c: Likewise.
3387 * tree-ssa-alias.c: Likewise.
3388 * tree-ssa-ccp.c: Likewise.
3389 * tree-ssa-coalesce.c: Likewise.
3390 * tree-ssa-copy.c: Likewise.
3391 * tree-ssa-copyrename.c: Likewise.
3392 * tree-ssa-dce.c: Likewise.
3393 * tree-ssa-dom.c: Likewise.
3394 * tree-ssa-dse.c: Likewise.
3395 * tree-ssa-forwprop.c: Likewise.
3396 * tree-ssa-ifcombine.c: Likewise.
3397 * tree-ssa-live.c: Likewise.
3398 * tree-ssa-loop-ch.c: Likewise.
3399 * tree-ssa-loop-im.c: Likewise.
3400 * tree-ssa-loop-ivcanon.c: Likewise.
3401 * tree-ssa-loop-ivopts.c: Likewise.
3402 * tree-ssa-loop-manip.c: Likewise.
3403 * tree-ssa-loop-niter.c: Likewise.
3404 * tree-ssa-loop-prefetch.c: Likewise.
3405 * tree-ssa-loop-unswitch.c: Likewise.
3406 * tree-ssa-loop.c: Likewise.
3407 * tree-ssa-math-opts.c: Likewise.
3408 * tree-ssa-operands.c: Likewise.
3409 * tree-ssa-phiopt.c: Likewise.
3410 * tree-ssa-phiprop.c: Likewise.
3411 * tree-ssa-pre.c: Likewise.
3412 * tree-ssa-propagate.c: Likewise.
3413 * tree-ssa-reassoc.c: Likewise.
3414 * tree-ssa-sccvn.c: Likewise.
3415 * tree-ssa-sink.c: Likewise.
3416 * tree-ssa-strlen.c: Likewise.
3417 * tree-ssa-structalias.c: Likewise.
3418 * tree-ssa-tail-merge.c: Likewise.
3419 * tree-ssa-ter.c: Likewise.
3420 * tree-ssa-threadedge.c: Likewise.
3421 * tree-ssa-threadupdate.c: Likewise.
3422 * tree-ssa-uncprop.c: Likewise.
3423 * tree-ssa-uninit.c: Likewise.
3424 * tree-ssa.c: Likewise.
3425 * tree-ssanames.c: Likewise.
3426 * tree-stdarg.c: Likewise.
3427 * tree-streamer-in.c: Likewise.
3428 * tree-switch-conversion.c: Likewise.
3429 * tree-tailcall.c: Likewise.
3430 * tree-vect-data-refs.c: Likewise.
3431 * tree-vect-generic.c: Likewise.
3432 * tree-vect-loop-manip.c: Likewise.
3433 * tree-vect-loop.c: Likewise.
3434 * tree-vect-patterns.c: Likewise.
3435 * tree-vect-slp.c: Likewise.
3436 * tree-vect-stmts.c: Likewise.
3437 * tree-vectorizer.c: Likewise.
3438 * tree-vrp.c: Likewise.
3441 * value-prof.c: Likewise.
3442 * var-tracking.c: Likewise.
3443 * varpool.c: Likewise.
3444 * vtable-verify.c: Likewise.
3446 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3448 * config/tilegx/tilegx.c: Include "tree.h".
3450 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3452 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
3453 the info, not after it.
3454 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
3455 dump_ssaname_info call.
3456 (pp_gimple_stmt_1): Adjust caller.
3457 (dump_phi_nodes): Likewise. Don't print "# " here.
3459 2013-10-22 Jan Hubicka <jh@suse.cz>
3461 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
3463 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
3464 * i386.c (expand_small_movmem_or_setmem): New function.
3465 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
3467 (alg_usable_p): Add support for value ranges; cleanup.
3468 (ix86_expand_set_or_movmem): Add support for misaligned moves.
3470 2013-10-22 Sterling Augustine <saugustine@google.com>
3472 * doc/invoke.texi: Document -ggnu-pubnames.
3473 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
3475 * dwarf2out.c: Include gdb/gdb-index.h.
3476 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
3477 debug_generate_pub_sections.
3478 (is_java, output_pubtables, output_pubname): New functions.
3479 (include_pubname_in_output): Handle debug_generate_pub_sections at
3481 (size_of_pubnames): Use new local space_for_flags based on
3482 debug_generate_pub_sections.
3483 (output_pubnames): Unify pubnames and pubtypes output logic.
3484 Genericize comments. Call output_pubname.
3485 (dwarf2out_finish): Move logic to output_pubnames and call it.
3487 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
3490 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
3491 Remove CCCmode handling.
3492 <case LTU>: Return 'c' suffix for CCCmode.
3493 <case GEU>: Return 'nc' suffix for CCCmode.
3494 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
3495 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
3496 (*sub<mode>3_cc_overflow): Ditto.
3497 (*subsi3_zext_cc_overflow): Ditto.
3499 2013-10-22 Steve Ellcey <sellcey@mips.com>
3501 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
3502 (AND (NOT OP1) (NOT OP2)).
3504 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
3506 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3507 meaning of merge-high and merge-low masks for little endian; avoid
3508 use of vector-pack masks for little endian for mismatched modes.
3510 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3512 * config/tilepro/tilepro.c: Include "tree.h".
3514 2013-10-22 Andreas Schwab <schwab@suse.de>
3516 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
3519 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
3521 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
3522 [Ed Smith-Rowland]): New entries.
3523 ([Stephen M. Webb]): Update.
3525 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
3527 * tree-ssa-ter.h: Remove duplicate copy of file contents.
3529 2013-10-21 Marek Polacek <polacek@redhat.com>
3532 * fold-const.c (fold_range_test): Return 0 if the type is not
3535 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
3537 * system.h: Move hwint.h include further down.
3538 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
3539 gcc_checking_asserts.
3540 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
3542 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3544 Fix volatile issues in optimize_bit_field_compare.
3545 * fold-const.c (optimize_bit_field_compare): Bail out if
3546 lvolatilep or rvolatilep.
3548 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3550 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
3551 and get_inner_reference returning different pmode for non-volatile
3552 bit-field members dependent on flag_strict_volatile_bitfields.
3553 * stor-layout.c (layout_decl): Remove special handling of
3554 flag_strict_volatile_bitfields.
3555 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
3556 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
3558 2013-10-21 Paulo Matos <pmatos@broadcom.com>
3560 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
3563 2013-10-21 Jeff Law <law@redhat.com>
3565 * tree-ssa-threadedge.c (thread_through_normal_block): New
3566 argument VISITED. Remove VISISTED as a local variable. When we
3567 have a threadable jump, verify the destination of the jump has not
3569 (thread_across_edge): Allocate VISITED bitmap once at function
3570 scope and use it throughout. Make sure to set appropriate bits in
3571 VISITED for E (start of jump thread path).
3572 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
3573 through a joiner if any edge on the path has a recorded jump thread.
3575 2013-10-21 Ian Lance Taylor <iant@google.com>
3577 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
3578 don't imply that attributes can solve all problems.
3579 (Directory Options): Fix typo.
3581 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3583 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
3584 extend and extend_add.
3586 2013-10-21 Richard Biener <rguenther@suse.de>
3588 PR tree-optimization/58794
3589 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
3590 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
3592 2013-10-21 Richard Biener <rguenther@suse.de>
3595 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
3596 to (T) X for sign-changing conversions (or no conversion).
3598 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
3600 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
3602 2013-10-20 Jan Hubicka <jh@suse.cz>
3604 * config/i386/i386-tune.def: Add comment; organize into categories
3606 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3608 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
3609 argument. Update function comment.
3610 (expand_set_or_movmem_via_rep): New function combining
3611 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
3612 (expand_movmem_via_rep_mov): Remove.
3613 expand_setmem_via_rep_stos): Remove.
3614 (expand_movmem_epilogue): Update calls correspondingly.
3615 (expand_setmem_epilogue_via_loop): Likewise.
3617 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
3618 (expand_set_or_movmem_prologue): New function combining
3619 expand_movmem_prologue and expand_setmem_prologue.
3620 (expand_movmem_prologue): Remove.
3621 (expand_setmem_prologue): Remove.
3622 (expand_set_or_movmem_constant_prologue): New function combining
3623 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
3624 (expand_constant_movmem_prologue): Remove.
3625 (expand_constant_setmem_prologue): Remove.
3626 (promote_duplicated_reg): Allow vector-const0 value.
3627 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
3628 and ix86_expand_setmem.
3629 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
3630 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
3632 2013-10-21 Diego Novillo <dnovillo@google.com>
3634 * asan.c: Include tree.h
3635 * bb-reorder.c: Likewise.
3636 * cfgcleanup.c: Likewise.
3637 * cfgloopmanip.c: Likewise.
3638 * data-streamer-in.c: Likewise.
3639 * data-streamer-out.c: Likewise.
3640 * data-streamer.c: Likewise.
3641 * dwarf2cfi.c: Likewise.
3642 * graphite-blocking.c: Likewise.
3643 * graphite-clast-to-gimple.c: Likewise.
3644 * graphite-dependences.c: Likewise.
3645 * graphite-interchange.c: Likewise.
3646 * graphite-optimize-isl.c: Likewise.
3647 * graphite-poly.c: Likewise.
3648 * graphite-scop-detection.c: Likewise.
3649 * graphite-sese-to-poly.c: Likewise.
3650 * graphite.c: Likewise.
3651 * ipa-devirt.c: Likewise.
3652 * ipa-profile.c: Likewise.
3655 * loop-init.c: Likewise.
3656 * loop-unroll.c: Likewise.
3657 * lower-subreg.c: Likewise.
3658 * lto/lto-object.c: Likewise.
3659 * recog.c: Likewise.
3660 * reginfo.c: Likewise.
3661 * tree-loop-distribution.c: Likewise.
3662 * tree-parloops.c: Likewise.
3663 * tree-ssa-strlen.c: Likewise.
3664 * tree-streamer.c: Likewise.
3665 * value-prof.c: Likewise.
3666 * target-globals.c: Likewise.
3667 * expr.h: Include tree-core.h instead of tree.h.
3668 * gimple.h: Likewise.
3669 * ipa-prop.h: Likewise.
3670 * ipa-utils.h: Likewise.
3671 * lto-streamer.h: Likewise.
3672 * streamer-hooks.h: Likewise.
3673 * ipa-reference.h: Include cgraph.h instead of tree.h.
3674 * cgraph.h: Include basic-block.h instead of tree.h.
3675 * tree-streamer.h: Do not include tree.h.
3676 * genattrtab.c (write_header): Generate inclusion of
3678 * genautomata.c (main): Likewise.
3679 * genemit.c: Likewise.
3680 * genopinit.c: Likewise.
3681 * genoutput.c (output_prologue): Likewise.
3682 * genpeep.c: Likewise.
3684 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3686 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
3688 (vec_unpacku_hi_v8hi): Likewise.
3689 (vec_unpacku_lo_v16qi): Likewise.
3690 (vec_unpacku_lo_v8hi): Likewise.
3692 2013-10-20 Jan Hubicka <jh@suse.cz>
3694 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
3695 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
3696 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
3698 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
3700 * config/mips/mips.h (ISA_HAS_WSBH): Define.
3701 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
3703 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
3705 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
3708 * system.h: Undef m_slot.
3710 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3712 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
3713 all elements for both endian flavors.
3715 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3718 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
3719 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
3720 and SI_REG for 64bit SYSV ABI targets.
3722 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3724 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
3725 to multi_reg_return. Clarify that we are skipping USEs of multiple
3726 return registers. Use bool type where appropriate.
3728 2013-10-18 Jan Hubicka <jh@suse.cz>
3730 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
3732 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
3733 (X86_TUNE_PUSH_MEMORY): Likewise.
3734 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
3735 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
3736 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
3738 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
3739 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
3741 (ix86_option_override_internal): Update to use tune features instead
3744 2013-10-18 Cong Hou <congh@google.com>
3746 PR tree-optimization/58508
3747 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
3748 statement that contains data refs with zero-step.
3750 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3752 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
3753 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
3754 * gimple-low.c (gimple_check_call_arg,
3755 gimple_check_call_matching_types): Move to cgraph.c.
3756 * gimple-low.h: Remove prototype.
3757 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
3758 Relocate from gimple-low.c.
3759 * cgraph.h: Add prototype. Don't include basic-block.h.
3760 * gimplify.c: Add gimple-low to include list.
3761 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
3762 * tree-eh.c: Add gimple-low to include list.
3763 * tree-nested.c: Likewise.
3764 * cfgexpand.c: Add tree-ssa-address.h to include list.
3766 * gimple-fold.c: Likewise.
3767 * gimple-ssa-strength-reduction.c: Likewise.
3768 * trans-mem.c: Likewise.
3769 * tree-mudflap.c: Likewise.
3770 * tree-ssa-loop-ivopts.c: Likewise.
3771 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
3772 (degenerate_phi_result): Move to tree-phinodes.c.
3773 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
3774 * tree-ssa-threadedge.c: Likewise.
3775 * tree-vrp.c: Likewise.
3776 * tree-phinodes.c (degenerate_phi_result): Relocate here.
3777 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
3778 * tree-phinodes.h (degenerate_phi_result): Add prototype.
3779 * tree-ssa-copy.c: Include tree-ssa-dom.h.
3780 * tree-ssa-forwprop.c: Likewise.
3781 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
3782 pass_data_cleanup_cfg_post_optimizing,
3783 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
3784 * tree-optimize.c: Delete File.
3785 * graphite.c: Include tree-cfgcleanup.h.
3786 * passes.c: Likewise.
3787 * tree-cfg.c: Likewise.
3788 * tree-profile.c: Likewise.
3789 * tree-ssa-dse.c: Likewise.
3790 * tree-ssa-loop-ivcanon.c: Likewise.
3791 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
3792 * tree-outof-ssa.c: Include sbitmap.h.
3793 * tree-ssa-live.c: Likewise.
3794 * tree-ssa-propagate.c: Likewise.
3795 * tree-ssa-structalias.c: Likewise.
3796 * tree-stdarg.c: Likewise.
3797 * Makefile.in (OBJS): Delete tree-optimize.o.
3798 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
3799 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
3800 * varasm.c: Include basic-block.h.
3801 * cfgloop.h: Include function.h instead of basic-block.h
3802 (bb_loop_depth): Move to cfgloop.c.
3803 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
3805 2013-10-18 Teresa Johnson <tejohnson@google.com>
3807 * predict.c (probably_never_executed): Compare frequency-based
3808 count to number of training runs.
3809 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
3811 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3813 * config/arm/arm.c (cortexa9_extra_costs): New table.
3814 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
3816 2013-10-18 Jeff Law <law@redhat.com>
3818 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
3820 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
3821 (register_jump_thread): Add "registered_jump_thread" debug
3823 * dbgcnt.def (registered_jump_thread): New debug counter.
3825 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3827 * config/rs6000/rs6000.c: Include cgraph.h.
3829 2013-10-18 Teresa Johnson <tejohnson@google.com>
3831 * tree-ssa-tail-merge.c (replace_block_by): Update edge
3832 weights during merging.
3834 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3836 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
3837 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
3838 * tree-ssa-operands.h: Remove prototype.
3839 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
3840 * gimple.c (virtual_operand_p): Relocate from gimple.c.
3841 * gimple.h: Add prototype.
3842 * gimple-ssa.h: Include tree-ssa-operands.h.
3843 * tree-dump.c: Add tree-cfg.h to include list.
3844 * tree-ssa-alias.c: Add ipa-reference.h to include list.
3845 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
3846 * config/i386/i386.c: Don't include tree-flow.h.
3847 * config/rs6000/rs6000.c: Likewise.
3849 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3851 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3853 2013-10-18 Richard Biener <rguenther@suse.de>
3855 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3856 or TYPE_UNSIGNED of integral types.
3857 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3858 NULL_TREE for zero-precision integral types.
3860 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3862 * config/aarch64/arm_neon.h
3863 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3865 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3868 * opth-gen.awk: Define target_flags_explicit.
3870 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3872 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3873 fields to the reg_addr array that describes the valid addressing
3874 mode for any register, general purpose registers, floating point
3875 registers, and Altivec registers.
3876 (FIRST_RELOAD_REG_CLASS): Likewise.
3877 (LAST_RELOAD_REG_CLASS): Likewise.
3878 (struct reload_reg_map_type): Likewise.
3879 (reload_reg_map_type): Likewise.
3880 (RELOAD_REG_VALID): Likewise.
3881 (RELOAD_REG_MULTIPLE): Likewise.
3882 (RELOAD_REG_INDEXED): Likewise.
3883 (RELOAD_REG_OFFSET): Likewise.
3884 (RELOAD_REG_PRE_INCDEC): Likewise.
3885 (RELOAD_REG_PRE_MODIFY): Likewise.
3886 (reg_addr): Likewise.
3887 (mode_supports_pre_incdec_p): New helper functions to say whether
3888 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3889 (mode_supports_pre_modify_p): Likewise.
3890 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3891 print the valid address mode bits for each mode.
3892 (rs6000_debug_print_mode): Likewise.
3893 (rs6000_debug_reg_global): Likewise.
3894 (rs6000_setup_reg_addr_masks): New function to set up the address
3895 mask bits for each type.
3896 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3897 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3898 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3899 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3900 PRE_MODIFY are supported.
3901 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3902 registers, instead of {src,dest}_av_p.
3903 (rs6000_print_options_internal): Tweak the debug output slightly.
3905 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3907 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3910 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3912 * tree-flow.h (struct omp_region): Move to omp-low.c.
3913 Remove omp_ prototypes and variables.
3914 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3915 (copy_var_decl): Relocate prototype from tree-flow.h.
3916 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3917 * tree.h: Move prototype to omp-low.h.
3918 * omp-low.h: New File. Relocate prototypes here.
3919 * omp-low.c (struct omp_region): Make local here.
3920 (root_omp_region): Make static.
3921 (copy_var_decl) Move to gimple.c.
3922 (new_omp_region): Make static.
3923 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3924 * tree-cfg.c: Include omp-low.h.
3925 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3926 * gimplify.c: Include omp-low.h.
3927 * tree-parloops.c: Likewise.
3929 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3931 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3932 operands are in memory, prefer to force non-matched operand 1 to
3935 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3938 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3939 restrict TImode addresses to single indirect registers if both
3940 -mquad-memory and -mvsx-timode are used.
3941 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3942 we should emit load/store quad. Remove using %y for quad memory
3945 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3946 constraints to allow load/store quad on machines where TImode is
3947 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3948 for TImode to load integer constants.
3950 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3952 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3954 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3956 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3957 handling of STACK_REG.
3959 2013-10-17 Richard Biener <rguenther@suse.de>
3961 PR tree-optimization/58143
3962 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3964 (rewrite_to_defined_overflow): Likewise.
3965 (move_computations_dom_walker::before_dom): Rewrite stmts
3966 with undefined signed overflow that are not always executed
3967 into unsigned arithmetic.
3969 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3972 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3973 explicit isa flag to be an options variable, instead of using
3974 global_options_set. Remove define from rs6000.h.
3975 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3977 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3978 Initialize rs6000_isa_flags_explicit.
3979 (rs6000_function_specific_save): Add gcc_options* parameter, so
3980 that the powerpc builds after the 2013-10-15 changes.
3981 (rs6000_function_specific_restore): Likewise.
3983 2013-10-16 DJ Delorie <dj@redhat.com>
3985 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3986 op is a REG before checking REGNO.
3987 (rl78_alloc_physical_registers): Verify pattern is a SET before
3990 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3992 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
3994 (vec_unpacks_lo_v4sf): Likewise.
3995 (vec_unpacks_float_hi_v4si): Likewise.
3996 (vec_unpacks_float_lo_v4si): Likewise.
3997 (vec_unpacku_float_hi_v4si): Likewise.
3998 (vec_unpacku_float_lo_v4si): Likewise.
4000 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4002 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
4003 (vsx_concat_v2sf): Likewise.
4005 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
4007 * config/aarch64/aarch64.md
4008 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
4010 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
4012 PR tree-optimization/58697
4013 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
4014 estimated_loop_iterations_int.
4015 (max_stmt_executions_int): Call get_max_loop_iterations_int.
4016 (get_max_loop_iterations_int): New. HWINT version of
4017 get_max_loop_iterations.
4018 * cfgloop.h: Add prototypes.
4019 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
4020 * loop-unroll.c (decide_peel_once_rolling): Call
4021 get_estimated_loop_iterations_int.
4022 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
4023 * tree-ssa-loop-niter.h: Tweak prototypes.
4025 2013-10-16 David Malcolm <dmalcolm@redhat.com>
4027 * gengtype-parse.c (struct_field_seq): Ignore access-control
4028 keywords ("public:" etc).
4030 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
4032 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
4033 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
4035 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
4037 * config/arm/arm.opt (mlra): New option.
4038 * config/arm/arm.c (arm_lra_p): New function.
4039 (TARGET_LRA_P): Define.
4041 2013-10-16 Paulo Matos <pmatos@broadcom.com>
4043 * tree-core.h (tree_code_name): Remove.
4044 * tree.h (get_tree_code_name): New prototype.
4045 * tree.c (tree_code_name): Make static.
4046 (get_tree_code_name): New function.
4047 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
4048 tree_class_check_failed, tree_range_check_failed,
4049 tree_not_class_check_failed, omp_clause_check_failed,
4050 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
4051 wrapper get_tree_code_name instead of calling tree_code_name directly.
4052 * tree-vrp.c (dump_asserts_for): Likewise.
4053 * tree-dump.c (dequeue_and_dump): Likewise.
4054 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
4055 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
4056 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
4057 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
4058 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
4059 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
4060 dump_gimple_omp_for): Likewise.
4061 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
4062 * tree-ssa-pre.c (print_pre_expr): Likewise.
4063 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
4064 * print-tree.c (print_node_brief, print_node): Likewise.
4065 * gimple.c (gimple_check_failed): Likewise.
4066 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
4067 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
4068 * config/mep/mep.c (mep_validate_vliw): Likewise.
4069 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
4070 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
4072 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
4074 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
4077 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
4079 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
4080 (MULTILIB_MATCHES): Add multilib for -march=v8.
4082 2013-10-15 Sriraman Tallam <tmsriram@google.com>
4085 * optc-save-gen.awk: Add extra parameter to the save and restore
4087 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
4088 * tree.c (build_optimization_node): New parameter. Add extra parameter
4089 to call to cl_optimization_save.
4090 (build_target_option_node): New parameter. Add extra parameter
4091 to call to cl_target_option_save.
4092 * tree.h (build_optimization_node): New parameter.
4093 (build_target_option_node): New parameter.
4094 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
4095 build_optimization_node and build_target_option_node.
4096 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
4097 (handle_pragma_push_options): Ditto.
4098 * toplev.c (process_options): Ditto.
4099 * opts.c (init_options_struct): Check for opts_set non-null.
4100 * target.def (target_option.save): New parameter.
4101 (target_option.restore): New parameter.
4102 * tm.texi: Generate.
4103 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
4104 (ix86_pragma_target_parse): Ditto.
4105 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
4107 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
4108 to build_optimization_node and build_target_option_node.
4109 (rs6000_valid_attribute_p): Ditto.
4110 (rs6000_pragma_target_parse): Ditto.
4111 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
4113 * config/i386/i386.h:
4114 TARGET_64BIT_P: New Macro
4115 TARGET_MMX_P: New Macro.
4116 TARGET_3DNOW_P: New Macro.
4117 TARGET_3DNOW_A_P: New Macro.
4118 TARGET_SSE_P: New Macro.
4119 TARGET_SSE2_P: New Macro.
4120 TARGET_SSE3_P: New Macro.
4121 TARGET_SSSE3_P: New Macro.
4122 TARGET_SSE4_1_P: New Macro.
4123 TARGET_SSE4_2_P: New Macro.
4124 TARGET_AVX_P: New Macro.
4125 TARGET_AVX2_P: New Macro.
4126 TARGET_AVX512F_P: New Macro.
4127 TARGET_AVX512PF_P: New Macro.
4128 TARGET_AVX512ER_P: New Macro.
4129 TARGET_AVX512CD_P: New Macro.
4130 TARGET_FMA_P: New Macro.
4131 TARGET_SSE4A_P: New Macro.
4132 TARGET_FMA4_P: New Macro.
4133 TARGET_XOP_P: New Macro.
4134 TARGET_LWP_P: New Macro.
4135 TARGET_ABM_P: New Macro.
4136 TARGET_BMI_P: New Macro.
4137 TARGET_BMI2_P: New Macro.
4138 TARGET_LZCNT_P: New Macro.
4139 TARGET_TBM_P: New Macro.
4140 TARGET_POPCNT_P: New Macro.
4141 TARGET_SAHF_P: New Macro.
4142 TARGET_MOVBE_P: New Macro.
4143 TARGET_CRC32_P: New Macro.
4144 TARGET_AES_P: New Macro.
4145 TARGET_PCLMUL_P: New Macro.
4146 TARGET_CMPXCHG16B_P: New Macro.
4147 TARGET_FSGSBASE_P: New Macro.
4148 TARGET_RDRND_P: New Macro.
4149 TARGET_F16C_P: New Macro.
4150 TARGET_RTM_P: New Macro.
4151 TARGET_HLE_P: New Macro.
4152 TARGET_RDSEED_P: New Macro.
4153 TARGET_PRFCHW_P: New Macro.
4154 TARGET_ADX_P: New Macro.
4155 TARGET_FXSR_P: New Macro.
4156 TARGET_XSAVE_P: New Macro.
4157 TARGET_XSAVEOPT_P: New Macro.
4158 TARGET_LP64_P: New Macro.
4159 TARGET_X32_P: New Macro.
4160 TARGET_FPMATH_DEFAULT_P: New Macro.
4161 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
4162 * config/i386/i386.c (ix86_option_override_internal): New parameters.
4164 Change ix86_tune_string to access opts->x_ix86_tune_string.
4165 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
4166 Change ix86_arch_string to access opts->x_ix86_arch_string.
4167 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
4168 Change ix86_pmode to access opts->x_ix86_pmode.
4169 Change ix86_abi to access opts->x_ix86_abi.
4170 Change ix86_cmodel to access opts->x_ix86_cmodel.
4171 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
4172 Change ix86_isa_flags_explicit to access
4173 opts->x_ix86_isa_flags_explicit.
4174 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
4175 Change ix86_regparm to access opts->x_ix86_regparm.
4176 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
4177 Change ix86_preferred_stack_boundary_arg to access
4178 opts->x_ix86_preferred_stack_boundary_arg.
4179 Change ix86_force_align_arg_pointer to access
4180 opts->x_ix86_force_align_arg_pointer.
4181 Change ix86_incoming_stack_boundar_arg to access
4182 opts->x_ix86_incoming_stack_boundar_arg.
4183 Change ix86_fpmath to access opts->x_ix86_fpmath.
4184 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
4185 Change ix86_recip_name to access opts->x_ix86_recip_name.
4186 Change ix86_stack_protector_guard to access
4187 opts->x_ix86_stack_protector_guard.
4188 Change ix86_tune_memcpy_strategy to access
4189 opts->x_ix86_tune_memcpy_strategy.
4190 Change ix86_tune_memset_strategy to access
4191 opts->x_ix86_tune_memset_strategy.
4192 Change global_options to access opts.
4193 Change global_options_set to access opts_set.
4194 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
4195 Change TARGET_MMX to TARGET_MMX_P (opts->...).
4196 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
4197 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
4198 Change TARGET_SSE to TARGET_SSE_P (opts->...).
4199 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
4200 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
4201 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
4202 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
4203 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
4204 Change TARGET_AVX to TARGET_AVX_P (opts->...).
4205 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
4206 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
4207 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
4208 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
4209 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
4210 Change TARGET_FMA to TARGET_FMA_P (opts->...).
4211 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
4212 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
4213 Change TARGET_XOP to TARGET_XOP_P (opts->...).
4214 Change TARGET_LWP to TARGET_LWP_P (opts->...).
4215 Change TARGET_ABM to TARGET_ABM_P (opts->...).
4216 Change TARGET_BMI to TARGET_BMI_P (opts->...).
4217 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
4218 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
4219 Change TARGET_TBM to TARGET_TBM_P (opts->...).
4220 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
4221 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
4222 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
4223 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
4224 Change TARGET_AES to TARGET_AES_P (opts->...).
4225 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
4226 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
4227 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
4228 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
4229 Change TARGET_F16C to TARGET_F16C_P (opts->...).
4230 Change TARGET_RTM to TARGET_RTM_P (opts->...).
4231 Change TARGET_HLE to TARGET_HLE_P (opts->...).
4232 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
4233 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
4234 Change TARGET_ADX to TARGET_ADX_P (opts->...).
4235 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
4236 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
4237 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
4238 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
4239 Change TARGET_X32 to TARGET_X32_P (opts->...).
4240 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
4241 Change TARGET_FLOAT_RETURNS_IN_80387 to
4242 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
4243 (ix86_function_specific_save): New parameter. Use opts-> fields
4244 to replace global fields.
4245 (ix86_function_specific_restore): Ditto.
4246 (ix86_valid_target_attribute_inner_p): New parameters.
4248 Fix call to ix86_handle_option and set_option.
4249 (ix86_valid_target_attribute_tree): New parameters.
4250 Change global_options to access opts.
4251 Change global_options_set to access opts_set.
4252 Fix call to ix86_valid_target_attribute_inner_p.
4253 Change ix86_tune_string to access opts->x_ix86_tune_string.
4254 Change ix86_arch_string to access opts->x_ix86_arch_string.
4255 Change ix86_fpmath to access opts->x_ix86_fpmath
4256 Fix call to ix86_option_override_internal.
4257 Fix call to ix86_add_new_builtins.
4258 Fix calls to build_optimization_node and build_target_option_node.
4259 (ix86_valid_target_attribute_p): Remove access to global_options.
4260 Use new gcc_options structure func_options.
4261 Fix call to ix86_valid_target_attribute_tree.
4262 Fix call to build_optimization_node.
4263 (get_builtin_code_for_version): Fix call to
4264 ix86_valid_target_attribute_tree.
4266 2013-10-15 David Malcolm <dmalcolm@redhat.com>
4268 * Makefile.in (PICFLAG): New.
4269 (enable_host_shared): New.
4270 (INTERNAL_CFLAGS): Use PICFLAG.
4271 (LIBIBERTY): Use pic build of libiberty.a if configured with
4272 --enable-host-shared.
4273 * configure.ac: Add --enable-host-shared, setting up new
4275 * configure: Regenerate.
4276 * doc/install.texi (--enable-shared): Add note contrasting it with ...
4277 (--enable-host-shared): New option.
4279 2013-10-15 Richard Biener <rguenther@suse.de>
4281 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
4282 for built-in functions.
4284 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
4286 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
4287 (optimize_range_tests_1): New function,
4288 extracted from optimize_range_tests.
4289 (optimize_range_tests_xor): Similarly.
4290 (optimize_range_tests_diff): New function.
4291 (optimize_range_tests): Use optimize_range_tests_1.
4293 2013-10-15 Cong Hou <congh@google.com>
4295 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
4296 requirement of the reduction pattern so that one operand of the
4297 reduction operation can come from outside of the loop.
4299 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4301 * config/arm/neon-schedgen.ml: Remove.
4302 * config/arm/cortex-a9-neon.md: Remove comment regarding
4305 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4307 * config/arm/types: Remove old neon types.
4309 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4311 * config/arm/cortex-a7.md
4312 (cortex_a7_neon_type): New.
4313 (cortex_a7_neon_mul): Update for new types.
4314 (cortex_a7_neon_mla): Likewise.
4315 (cortex_a7_neon): Likewise.
4317 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4319 * config/arm/cortex-a15-neon.md
4320 (cortex_a15_neon_type): New,
4322 (cortex_a15_neon_int_1): Remove.
4323 (cortex_a15_neon_int_2): Likewise.
4324 (cortex_a15_neon_int_3): Likewise.
4325 (cortex_a15_neon_int_4): Likewise.
4326 (cortex_a15_neon_int_5): Likewise.
4327 (cortex_a15_neon_vqneg_vqabs): Likewise.
4328 (cortex_a15_neon_vmov): Likewise.
4329 (cortex_a15_neon_vaba): Likewise.
4330 (cortex_a15_neon_vaba_qqq): Likewise.
4331 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4332 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4333 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4335 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4336 (cortex_a15_neon_mla_qqq_8_16): Likewise.
4337 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
4338 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4339 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4340 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
4341 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4342 (cortex_a15_neon_shift_1): Likewise.
4343 (cortex_a15_neon_shift_2): Likewise.
4344 (cortex_a15_neon_shift_3): Likewise.
4345 (cortex_a15_neon_vshl_ddd): Likewise.
4346 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4347 (cortex_a15_neon_vsra_vrsra): Likewise.
4348 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4349 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4350 (cortex_a15_neon_bp_3cycle): Likewise.
4351 (cortex_a15_neon_ldm_2): Likewise.
4352 (cortex_a15_neon_stm_2): Likewise.
4353 (cortex_a15_neon_mcr): Likewise.
4354 (cortex_a15_neon_mrc): Likewise.
4355 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
4356 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
4357 (cortex_a15_neon_fp_vmul_ddd): Likewise.
4358 (cortex_a15_neon_fp_vmul_qqd): Likewise.
4359 (cortex_a15_neon_fp_vmla_ddd): Likewise.
4360 (cortex_a15_neon_fp_vmla_qqq): Likewise.
4361 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4362 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4363 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4364 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4365 (cortex_a15_neon_bp_simple): Likewise.
4366 (cortex_a15_neon_bp_2cycle): Likewise.
4367 (cortex_a15_neon_bp_3cycle): Likewise.
4368 (cortex_a15_neon_vld1_1_2_regs): Likewise.
4369 (cortex_a15_neon_vld1_3_4_regs): Likewise.
4370 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4371 (cortex_a15_neon_vld2_4_regs): Likewise.
4372 (cortex_a15_neon_vld3_vld4): Likewise.
4373 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4374 (cortex_a15_neon_vst1_3_4_regs): Likewise.
4375 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
4376 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
4377 (cortex_a15_neon_vst3_vst4): Rename to...
4378 (cortex_a15_neon_vst4): This, update for new attributes.
4379 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
4380 (cortex_a15_neon_vld3_vld4_lane): Likewise.
4381 (cortex_a15_neon_vst1_vst2_lane): Likewise.
4382 (cortex_a15_neon_vst3_vst4_lane): Likewise.
4383 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
4384 (cortex_a15_neon_ldm_2): Likewise.
4385 (cortex_a15_neon_stm_2): Likewise.
4386 (cortex_a15_neon_mcr): Likewise.
4387 (cortex_a15_neon_mcr_2_mcrr): Likewise.
4388 (cortex_a15_neon_mrc): Likewise.
4389 (cortex_a15_neon_mrrc): Likewise.
4391 (cortex_a15_neon_abd): New.
4392 (cortex_a15_neon_abd_q): Likewise.
4393 (cortex_a15_neon_aba): Likewise.
4394 (cortex_a15_neon_aba_q): Likewise.
4395 (cortex_a15_neon_acc): Likewise.
4396 (cortex_a15_neon_acc_q): Likewise.
4397 (cortex_a15_neon_arith_basic): Likewise.
4398 (cortex_a15_neon_arith_complex): Likewise.
4399 (cortex_a15_neon_multiply): Likewise.
4400 (cortex_a15_neon_multiply_q): Likewise.
4401 (cortex_a15_neon_mla): Likewise.
4402 (cortex_a15_neon_mla_q): Likewise.
4403 (cortex_a15_neon_sat_mla_long): Likewise.
4404 (cortex_a15_neon_shift_acc): Likewise.
4405 (cortex_a15_neon_shift_imm_basic): Likewise.
4406 (cortex_a15_neon_shift_imm_complex): Likewise.
4407 (cortex_a15_neon_shift_reg_basic): Likewise.
4408 (cortex_a15_neon_shift_reg_basic_q): Likewise.
4409 (cortex_a15_neon_shift_reg_complex): Likewise.
4410 (cortex_a15_neon_shift_reg_complex_q): Likewise.
4411 (cortex_a15_neon_fp_negabs): Likewise
4412 (cortex_a15_neon_fp_arith): Likewise
4413 (cortex_a15_neon_fp_arith_q): Likewise
4414 (cortex_a15_neon_fp_cvt_int): Likewise
4415 (cortex_a15_neon_fp_cvt_int_q): Likewise
4416 (cortex_a15_neon_fp_cvt_16): Likewise
4417 (cortex_a15_neon_fp_mul): Likewise
4418 (cortex_a15_neon_fp_mul_q): Likewise
4419 (cortex_a15_neon_fp_mla): Likewise
4420 (cortex_a15_neon_fp_mla_q): Likewise
4421 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
4422 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
4423 (cortex_a15_neon_bitops): Likewise.
4424 (cortex_a15_neon_bitops_q): Likewise.
4425 (cortex_a15_neon_from_gp): Likewise.
4426 (cortex_a15_neon_from_gp_q): Likewise.
4427 (cortex_a15_neon_tbl3_tbl4): Likewise.
4428 (cortex_a15_neon_zip_q): Likewise.
4429 (cortex_a15_neon_to_gp): Likewise.
4430 (cortex_a15_neon_load_a): Likewise.
4431 (cortex_a15_neon_load_b): Likewise.
4432 (cortex_a15_neon_load_c): Likewise.
4433 (cortex_a15_neon_load_d): Likewise.
4434 (cortex_a15_neon_load_e): Likewise.
4435 (cortex_a15_neon_load_f): Likewise.
4436 (cortex_a15_neon_store_a): Likewise.
4437 (cortex_a15_neon_store_b): Likewise.
4438 (cortex_a15_neon_store_c): Likewise.
4439 (cortex_a15_neon_store_d): Likewise.
4440 (cortex_a15_neon_store_e): Likewise.
4441 (cortex_a15_neon_store_f): Likewise.
4442 (cortex_a15_neon_store_g): Likewise.
4443 (cortex_a15_neon_store_h): Likewise.
4444 (cortex_a15_vfp_to_from_gp): Likewise.
4446 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4448 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
4450 (cortex_a9_neon_vshl_ddd): Remove.
4451 (cortex_a9_neon_vst3_vst4): Likewise.
4452 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4454 (cortex_a9_neon_bit_ops_q): New.
4456 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
4457 (cortex_a9_neon_int_2): Likewise.
4458 (cortex_a9_neon_int_3): Likewise.
4459 (cortex_a9_neon_int_4): Likewise.
4460 (cortex_a9_neon_int_5): Likewise.
4461 (cortex_a9_neon_vqneg_vqabs): Likewise.
4462 (cortex_a9_neon_vmov): Likewise.
4463 (cortex_a9_neon_vaba): Likewise.
4464 (cortex_a9_neon_vaba_qqq): Likewise.
4465 (cortex_a9_neon_shift_1): Likewise.
4466 (cortex_a9_neon_shift_2): Likewise.
4467 (cortex_a9_neon_shift_3): Likewise.
4468 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4469 (cortex_a9_neon_vsra_vrsra): Likewise.
4470 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4471 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4472 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4474 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4475 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4476 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4478 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4479 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4480 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4481 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4482 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4483 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4484 (cortex_a9_neon_fp_vsum): Likewise.
4485 (cortex_a9_neon_fp_vmul_ddd): Likewise.
4486 (cortex_a9_neon_fp_vmul_qqd): Likewise.
4487 (cortex_a9_neon_fp_vmla_ddd): Likewise.
4488 (cortex_a9_neon_fp_vmla_qqq): Likewise.
4489 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
4490 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
4491 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4492 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4493 (cortex_a9_neon_bp_simple): Likewise.
4494 (cortex_a9_neon_bp_2cycle): Likewise.
4495 (cortex_a9_neon_bp_3cycle): Likewise.
4496 (cortex_a9_neon_ldr): Likewise.
4497 (cortex_a9_neon_str): Likewise.
4498 (cortex_a9_neon_vld1_1_2_regs): Likewise.
4499 (cortex_a9_neon_vld1_3_4_regs): Likewise.
4500 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4501 (cortex_a9_neon_vld2_4_regs): Likewise.
4502 (cortex_a9_neon_vld3_vld4): Likewise.
4503 (cortex_a9_neon_vld1_vld2_lane): Likewise.
4504 (cortex_a9_neon_vld3_vld4_lane): Likewise.
4505 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4506 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4507 (cortex_a9_neon_vst1_3_4_regs): Likewise.
4508 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
4509 (cortex_a9_neon_vst1_vst2_lane): Likewise.
4510 (cortex_a9_neon_vst3_vst4_lane): Likewise.
4511 (cortex_a9_neon_mcr): Likewise.
4512 (cortex_a9_neon_mcr_2_mcrr): Likewise.
4513 (cortex_a9_neon_mrc): Likewise.
4514 (cortex_a9_neon_mrrc): Likewise.
4516 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4518 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
4520 (cortex_a8_neon_vshl_ddd): Remove.
4521 (cortex_a8_neon_vst3_vst4): Likewise.
4522 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4524 (cortex_a8_neon_bit_ops_q): New.
4526 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
4527 (cortex_a8_neon_int_2): Likewise..
4528 (cortex_a8_neon_int_3): Likewise.
4529 (cortex_a8_neon_int_5): Likewise.
4530 (cortex_a8_neon_vqneg_vqabs): Likewise.
4531 (cortex_a8_neon_int_4): Likewise.
4532 (cortex_a8_neon_vaba): Likewise.
4533 (cortex_a8_neon_vaba_qqq): Likewise.
4534 (cortex_a8_neon_shift_1): Likewise.
4535 (cortex_a8_neon_shift_2): Likewise.
4536 (cortex_a8_neon_shift_3): Likewise.
4537 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4538 (cortex_a8_neon_vsra_vrsra): Likewise.
4539 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4540 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4541 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4543 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4544 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4545 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4547 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4548 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4549 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4550 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4551 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4552 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4553 (cortex_a8_neon_fp_vsum): Likewise.
4554 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4555 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4556 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4557 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4558 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4559 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4560 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4561 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4562 (cortex_a8_neon_bp_simple): Likewise.
4563 (cortex_a8_neon_bp_2cycle): Likewise.
4564 (cortex_a8_neon_bp_3cycle): Likewise.
4565 (cortex_a8_neon_ldr): Likewise.
4566 (cortex_a8_neon_str): Likewise.
4567 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4568 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4569 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4570 (cortex_a8_neon_vld2_4_regs): Likewise.
4571 (cortex_a8_neon_vld3_vld4): Likewise.
4572 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4573 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4574 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4575 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4576 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4577 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4578 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4579 (cortex_a8_neon_mcr): Likewise.
4580 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4581 (cortex_a8_neon_mrc): Likewise.
4582 (cortex_a8_neon_mrrc): Likewise.
4584 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4586 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
4588 * config/aarch64/aarch64-simd.md (neon_type): Remove.
4589 (aarch64_simd_dup<mode>): Add "type" attribute.
4590 (aarch64_dup_lane<mode>): Likewise.
4591 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
4592 (*aarch64_simd_mov<mode>): Likewise.
4593 (aarch64_simd_mov_from_<mode>low): Likewise.
4594 (aarch64_simd_mov_from_<mode>high): Likewise.
4595 (orn<mode>3): Likewise.
4596 (bic<mode>3): Likewise.
4597 (add<mode>3): Likewise.
4598 (sub<mode>3): Likewise.
4599 (mul<mode>3): Likewise.
4600 (*aarch64_mul3_elt<mode>): Likewise.
4601 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4602 (*aarch64_mul3_elt_to_128df): Likewise.
4603 (*aarch64_mul3_elt_to_64v2df): Likewise.
4604 (neg<mode>2): Likewise.
4605 (abs<mode>2): Likewise.
4606 (abd<mode>_3): Likewise.
4607 (aba<mode>_3): Likewise.
4608 (fabd<mode>_3): Likewise.
4609 (*fabd_scalar<mode>3): Likewise.
4610 (and<mode>3): Likewise.
4611 (ior<mode>3): Likewise.
4612 (xor<mode>3): Likewise.
4613 (one_cmpl<mode>2): Likewise.
4614 (aarch64_simd_vec_set<mode>): Likewise.
4615 (aarch64_simd_lshr<mode>): Likewise.
4616 (aarch64_simd_ashr<mode>): Likewise.
4617 (aarch64_simd_imm_shl<mode>): Likewise.
4618 (aarch64_simd_reg_sshl<mode): Likewise.
4619 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
4620 (aarch64_simd_reg_shl<mode>_signed): Likewise.
4621 (aarch64_simd_vec_setv2di): Likewise.
4622 (aarch64_simd_vec_set<mode>): Likewise.
4623 (aarch64_mla<mode>): Likewise.
4624 (*aarch64_mla_elt<mode>): Likewise.
4625 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4626 (aarch64_mls<mode>): Likewise.
4627 (*aarch64_mls_elt<mode>): Likewise.
4628 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4629 (<su><maxmin><mode>3): Likewise.
4630 (move_lo_quad_<mode>): Likewise.
4631 (aarch64_simd_move_hi_quad_<mode>): Likewise.
4632 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
4633 (vec_pack_trunc_<mode>): Likewise.
4634 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
4635 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
4636 (*aarch64_<su>mlal_lo<mode>): Likewise.
4637 (*aarch64_<su>mlal_hi<mode>): Likewise.
4638 (*aarch64_<su>mlsl_lo<mode>): Likewise.
4639 (*aarch64_<su>mlsl_hi<mode>): Likewise.
4640 (*aarch64_<su>mlal<mode>): Likewise.
4641 (*aarch64_<su>mlsl<mode>): Likewise.
4642 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
4643 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
4644 (add<mode>3): Likewise.
4645 (sub<mode>3): Likewise.
4646 (mul<mode>3): Likewise.
4647 (div<mode>3): Likewise.
4648 (neg<mode>2): Likewise.
4649 (abs<mode>2): Likewise.
4650 (fma<mode>4): Likewise.
4651 (*aarch64_fma4_elt<mode>): Likewise.
4652 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4653 (*aarch64_fma4_elt_to_128df): Likewise.
4654 (*aarch64_fma4_elt_to_64v2df): Likewise.
4655 (fnma<mode>4): Likewise.
4656 (*aarch64_fnma4_elt<mode>): Likewise.
4657 (*aarch64_fnma4_elt_<vswap_width_name><mode>
4658 (*aarch64_fnma4_elt_to_128df): Likewise.
4659 (*aarch64_fnma4_elt_to_64v2df): Likewise.
4660 (<frint_pattern><mode>2): Likewise.
4661 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
4662 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
4663 (vec_unpacks_lo_v4sf): Likewise.
4664 (aarch64_float_extend_lo_v2df): Likewise.
4665 (vec_unpacks_hi_v4sf): Likewise.
4666 (aarch64_float_truncate_lo_v2sf): Likewise.
4667 (aarch64_float_truncate_hi_v4sf): Likewise.
4668 (aarch64_vmls<mode>): Likewise.
4669 (<su><maxmin><mode>3): Likewise.
4670 (<maxmin_uns><mode>3): Likewise.
4671 (reduc_<sur>plus_<mode>): Likewise.
4672 (reduc_<sur>plus_v2di): Likewise.
4673 (reduc_<sur>plus_v2si): Likewise.
4674 (reduc_<sur>plus_<mode>): Likewise.
4675 (aarch64_addpv4sf): Likewise.
4676 (clz<mode>2): Likewise.
4677 (reduc_<maxmin_uns>_<mode>): Likewise.
4678 (reduc_<maxmin_uns>_v2di): Likewise.
4679 (reduc_<maxmin_uns>_v2si): Likewise.
4680 (reduc_<maxmin_uns>_<mode>): Likewise.
4681 (reduc_<maxmin_uns>_v4sf): Likewise.
4682 (aarch64_simd_bsl<mode>_internal): Likewise.
4683 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
4684 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
4685 (aarch64_get_lane<mode>): Likewise.
4686 (*aarch64_combinez<mode>): Likewise.
4687 (aarch64_combine<mode>): Likewise.
4688 (aarch64_simd_combine<mode>): Likewise.
4689 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
4690 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
4691 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
4692 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
4693 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
4694 (aarch64_<sur>h<addsub><mode>): Likewise.
4695 (aarch64_<sur><addsub>hn<mode>): Likewise.
4696 (aarch64_<sur><addsub>hn2<mode>): Likewise.
4697 (aarch64_pmul<mode>): Likewise.
4698 (aarch64_<su_optab><optab><mode>): Likewise.
4699 (aarch64_<sur>qadd<mode>): Likewise.
4700 (aarch64_sqmovun<mode>): Likewise.
4701 (aarch64_<sur>qmovn<mode>): Likewise.
4702 (aarch64_s<optab><mode>): Likewise.
4703 (aarch64_sq<r>dmulh<mode>): Likewise.
4704 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4705 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
4706 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4707 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
4708 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4709 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4710 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
4711 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
4712 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
4713 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
4714 (aarch64_sqdmull<mode>): Likewise.
4715 (aarch64_sqdmull_lane<mode>_internal): Likewise.
4716 (aarch64_sqdmull_n<mode>): Likewise.
4717 (aarch64_sqdmull2<mode>_internal): Likewise.
4718 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
4719 (aarch64_sqdmull2_n<mode>_internal): Likewise.
4720 (aarch64_<sur>shl<mode>): Likewise.
4721 (aarch64_<sur>q<r>shl<mode>
4722 (aarch64_<sur>shll_n<mode>): Likewise.
4723 (aarch64_<sur>shll2_n<mode>): Likewise.
4724 (aarch64_<sur>shr_n<mode>): Likewise.
4725 (aarch64_<sur>sra_n<mode>): Likewise.
4726 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
4727 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
4728 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
4729 (aarch64_cm<optab><mode>): Likewise.
4730 (aarch64_cm<optab>di): Likewise.
4731 (aarch64_cm<optab><mode>): Likewise.
4732 (aarch64_cm<optab>di): Likewise.
4733 (aarch64_cmtst<mode>): Likewise.
4734 (aarch64_cmtstdi): Likewise.
4735 (aarch64_cm<optab><mode>): Likewise.
4736 (*aarch64_fac<optab><mode>): Likewise.
4737 (aarch64_addp<mode>): Likewise.
4738 (aarch64_addpdi): Likewise.
4739 (sqrt<mode>2): Likewise.
4740 (vec_load_lanesoi<mode>): Likewise.
4741 (vec_store_lanesoi<mode>): Likewise.
4742 (vec_load_lanesci<mode>): Likewise.
4743 (vec_store_lanesci<mode>): Likewise.
4744 (vec_load_lanesxi<mode>): Likewise.
4745 (vec_store_lanesxi<mode>): Likewise.
4746 (*aarch64_mov<mode>): Likewise.
4747 (aarch64_ld2<mode>_dreg): Likewise.
4748 (aarch64_ld2<mode>_dreg): Likewise.
4749 (aarch64_ld3<mode>_dreg): Likewise.
4750 (aarch64_ld3<mode>_dreg): Likewise.
4751 (aarch64_ld4<mode>_dreg): Likewise.
4752 (aarch64_ld4<mode>_dreg): Likewise.
4753 (aarch64_tbl1<mode>): Likewise.
4754 (aarch64_tbl2v16qi): Likewise.
4755 (aarch64_combinev16qi): Likewise.
4756 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
4757 (aarch64_st2<mode>_dreg): Likewise.
4758 (aarch64_st2<mode>_dreg): Likewise.
4759 (aarch64_st3<mode>_dreg): Likewise.
4760 (aarch64_st3<mode>_dreg): Likewise.
4761 (aarch64_st4<mode>_dreg): Likewise.
4762 (aarch64_st4<mode>_dreg): Likewise.
4763 (*aarch64_simd_ld1r<mode>): Likewise.
4764 (aarch64_frecpe<mode>): Likewise.
4765 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
4766 (aarch64_frecps<mode>): Likewise.
4768 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4770 * config/arm/iterators.md (V_elem_ch): New.
4772 (VQH_type): Likewise.
4773 * config/arm/arm.md (is_neon_type): New.
4774 (conds): Use is_neon_type.
4775 (anddi3_insn): Update type attribute.
4776 (xordi3_insn): Likewise.
4777 (one_cmpldi2): Likewise.
4778 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
4779 * config/arm/neon.md (neon_mov): Update type attribute.
4780 (*movmisalign<mode>_neon_store): Likewise.
4781 (*movmisalign<mode>_neon_load): Likewise.
4782 (vec_set<mode>_internal): Likewise.
4783 (vec_set<mode>_internal): Likewise.
4784 (vec_setv2di_internal): Likewise.
4785 (vec_extract<mode>): Likewise.
4786 (vec_extract<mode>): Likewise.
4787 (vec_extractv2di): Likewise.
4788 (*add<mode>3_neon): Likewise.
4789 (adddi3_neon): Likewise.
4790 (*sub<mode>3_neon): Likewise.
4791 (subdi3_neon): Likewise.
4792 (fma<VCVTF:mode>4): Likewise.
4793 (fma<VCVTF:mode>4_intrinsic): Likewise.
4794 (*fmsub<VCVTF:mode>4): Likewise.
4795 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4796 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4797 (ior<mode>3): Likewise.
4798 (and<mode>3): Likewise.
4799 (orn<mode>3_neon): Likewise.
4800 (orndi3_neon): Likewise.
4801 (bic<mode>3_neon): Likewise.
4802 (bicdi3_neon): Likewise.
4803 (xor<mode>3): Likewise.
4804 (one_cmpl<mode>2): Likewise.
4805 (abs<mode>2): Likewise.
4806 (neg<mode>2): Likewise.
4807 (negdi2_neon): Likewise.
4808 (*umin<mode>3_neon): Likewise.
4809 (*umax<mode>3_neon): Likewise.
4810 (*smin<mode>3_neon): Likewise.
4811 (*smax<mode>3_neon): Likewise.
4812 (vashl<mode>3): Likewise.
4813 (vashr<mode>3_imm): Likewise.
4814 (vlshr<mode>3_imm): Likewise.
4815 (ashl<mode>3_signed): Likewise.
4816 (ashl<mode>3_unsigned): Likewise.
4817 (neon_load_count): Likewise.
4818 (ashldi3_neon_noclobber): Likewise.
4819 (ashldi3_neon): Likewise.
4820 (signed_shift_di3_neon): Likewise.
4821 (unsigned_shift_di3_neon): Likewise.
4822 (ashrdi3_neon_imm_noclobber): Likewise.
4823 (lshrdi3_neon_imm_noclobber): Likewise.
4824 (<shift>di3_neon): Likewise.
4825 (widen_ssum<mode>3): Likewise.
4826 (widen_usum<mode>3): Likewise.
4827 (quad_halves_<code>v4si): Likewise.
4828 (quad_halves_<code>v4sf): Likewise.
4829 (quad_halves_<code>v8hi): Likewise.
4830 (quad_halves_<code>v16qi): Likewise.
4831 (reduc_splus_v2di): Likewise.
4832 (neon_vpadd_internal<mode>): Likewise.
4833 (neon_vpsmin<mode>): Likewise.
4834 (neon_vpsmax<mode>): Likewise.
4835 (neon_vpumin<mode>): Likewise.
4836 (neon_vpumax<mode>): Likewise.
4837 (*ss_add<mode>_neon): Likewise.
4838 (*us_add<mode>_neon): Likewise.
4839 (*ss_sub<mode>_neon): Likewise.
4840 (*us_sub<mode>_neon): Likewise.
4841 (neon_vadd<mode>_unspec): Likewise.
4842 (neon_vaddl<mode>): Likewise.
4843 (neon_vaddw<mode>): Likewise.
4844 (neon_vhadd<mode>): Likewise.
4845 (neon_vqadd<mode>): Likewise.
4846 (neon_vaddhn<mode>): Likewise.
4847 (neon_vmul<mode>): Likewise.
4848 (neon_vfms<VCVTF:mode>): Likewise.
4849 (neon_vmlal<mode>): Likewise.
4850 (neon_vmls<mode>): Likewise.
4851 (neon_vmlsl<mode>): Likewise.
4852 (neon_vqdmulh<mode>): Likewise.
4853 (neon_vqdmlal<mode>): Likewise.
4854 (neon_vqdmlsl<mode>): Likewise.
4855 (neon_vmull<mode>): Likewise.
4856 (neon_vqdmull<mode>): Likewise.
4857 (neon_vsub<mode>_unspec): Likewise.
4858 (neon_vsubl<mode>): Likewise.
4859 (neon_vsubw<mode>): Likewise.
4860 (neon_vqsub<mode>): Likewise.
4861 (neon_vhsub<mode>): Likewise.
4862 (neon_vsubhn<mode>): Likewise.
4863 (neon_vceq<mode>): Likewise.
4864 (neon_vcge<mode>): Likewise.
4865 (neon_vcgeu<mode>): Likewise.
4866 (neon_vcgt<mode>): Likewise.
4867 (neon_vcgtu<mode>): Likewise.
4868 (neon_vcle<mode>): Likewise.
4869 (neon_vclt<mode>): Likewise.
4870 (neon_vcage<mode>): Likewise.
4871 (neon_vcagt<mode>): Likewise.
4872 (neon_vtst<mode>): Likewise.
4873 (neon_vabd<mode>): Likewise.
4874 (neon_vabdl<mode>): Likewise.
4875 (neon_vaba<mode>): Likewise.
4876 (neon_vabal<mode>): Likewise.
4877 (neon_vmax<mode>): Likewise.
4878 (neon_vmin<mode>): Likewise.
4879 (neon_vpaddl<mode>): Likewise.
4880 (neon_vpadal<mode>): Likewise.
4881 (neon_vpmax<mode>): Likewise.
4882 (neon_vpmin<mode>): Likewise.
4883 (neon_vrecps<mode>): Likewise.
4884 (neon_vrsqrts<mode>): Likewise.
4885 (neon_vqabs<mode>): Likewise.
4886 (neon_vqneg<mode>): Likewise.
4887 (neon_vcls<mode>): Likewise.
4888 (clz<mode>2): Likewise.
4889 (popcount<mode>2): Likewise.
4890 (neon_vrecpe<mode>): Likewise.
4891 (neon_vrsqrte<mode>): Likewise.
4892 (neon_vget_lane<mode>_sext_internal): Likewise.
4893 (neon_vget_lane<mode>_zext_internal): Likewise.
4894 (neon_vdup_n<mode>): Likewise.
4895 (neon_vdup_n<mode>): Likewise.
4896 (neon_vdup_nv2di): Likewise.
4897 (neon_vdup_lane<mode>_interal): Likewise.
4898 (*neon_vswp<mode>): Likewise.
4899 (neon_vcombine<mode>): Likewise.
4900 (float<mode><V_cvtto>2): Likewise.
4901 (floatuns<mode><V_cvtto>2): Likewise.
4902 (fix_trunc<mode><V_cvtto>2): Likewise.
4903 (fixuns_trunc<mode><V_cvtto>2
4904 (neon_vcvt<mode>): Likewise.
4905 (neon_vcvt<mode>): Likewise.
4906 (neon_vcvtv4sfv4hf): Likewise.
4907 (neon_vcvtv4hfv4sf): Likewise.
4908 (neon_vcvt_n<mode>): Likewise.
4909 (neon_vcvt_n<mode>): Likewise.
4910 (neon_vmovn<mode>): Likewise.
4911 (neon_vqmovn<mode>): Likewise.
4912 (neon_vqmovun<mode>): Likewise.
4913 (neon_vmovl<mode>): Likewise.
4914 (neon_vmul_lane<mode>): Likewise.
4915 (neon_vmul_lane<mode>): Likewise.
4916 (neon_vmull_lane<mode>): Likewise.
4917 (neon_vqdmull_lane<mode>): Likewise.
4918 (neon_vqdmulh_lane<mode>): Likewise.
4919 (neon_vqdmulh_lane<mode>): Likewise.
4920 (neon_vmla_lane<mode>): Likewise.
4921 (neon_vmla_lane<mode>): Likewise.
4922 (neon_vmlal_lane<mode>): Likewise.
4923 (neon_vqdmlal_lane<mode>): Likewise.
4924 (neon_vmls_lane<mode>): Likewise.
4925 (neon_vmls_lane<mode>): Likewise.
4926 (neon_vmlsl_lane<mode>): Likewise.
4927 (neon_vqdmlsl_lane<mode>): Likewise.
4928 (neon_vext<mode>): Likewise.
4929 (neon_vrev64<mode>): Likewise.
4930 (neon_vrev32<mode>): Likewise.
4931 (neon_vrev16<mode>): Likewise.
4932 (neon_vbsl<mode>_internal): Likewise.
4933 (neon_vshl<mode>): Likewise.
4934 (neon_vqshl<mode>): Likewise.
4935 (neon_vshr_n<mode>): Likewise.
4936 (neon_vshrn_n<mode>): Likewise.
4937 (neon_vqshrn_n<mode>): Likewise.
4938 (neon_vqshrun_n<mode>): Likewise.
4939 (neon_vshl_n<mode>): Likewise.
4940 (neon_vqshl_n<mode>): Likewise.
4941 (neon_vqshlu_n<mode>): Likewise.
4942 (neon_vshll_n<mode>): Likewise.
4943 (neon_vsra_n<mode>): Likewise.
4944 (neon_vsri_n<mode>): Likewise.
4945 (neon_vsli_n<mode>): Likewise.
4946 (neon_vtbl1v8qi): Likewise.
4947 (neon_vtbl2v8qi): Likewise.
4948 (neon_vtbl3v8qi): Likewise.
4949 (neon_vtbl4v8qi): Likewise.
4950 (neon_vtbl1v16qi): Likewise.
4951 (neon_vtbl2v16qi): Likewise.
4952 (neon_vcombinev16qi): Likewise.
4953 (neon_vtbx1v8qi): Likewise.
4954 (neon_vtbx2v8qi): Likewise.
4955 (neon_vtbx3v8qi): Likewise.
4956 (neon_vtbx4v8qi): Likewise.
4957 (*neon_vtrn<mode>_insn): Likewise.
4958 (*neon_vzip<mode>_insn): Likewise.
4959 (*neon_vuzp<mode>_insn): Likewise.
4960 (neon_vld1<mode>): Likewise.
4961 (neon_vld1_lane<mode>): Likewise.
4962 (neon_vld1_lane<mode>): Likewise.
4963 (neon_vld1_dup<mode>): Likewise.
4964 (neon_vld1_dup<mode>): Likewise.
4965 (neon_vld1_dupv2di): Likewise.
4966 (neon_vst1<mode>): Likewise.
4967 (neon_vst1_lane<mode>): Likewise.
4968 (neon_vst1_lane<mode>): Likewise.
4969 (neon_vld2<mode>): Likewise.
4970 (neon_vld2<mode>): Likewise.
4971 (neon_vld2_lane<mode>): Likewise.
4972 (neon_vld2_lane<mode>): Likewise.
4973 (neon_vld2_dup<mode>): Likewise.
4974 (neon_vst2<mode>): Likewise.
4975 (neon_vst2<mode>): Likewise.
4976 (neon_vst2_lane<mode>): Likewise.
4977 (neon_vst2_lane<mode>): Likewise.
4978 (neon_vld3<mode>): Likewise.
4979 (neon_vld3qa<mode>): Likewise.
4980 (neon_vld3qb<mode>): Likewise.
4981 (neon_vld3_lane<mode>): Likewise.
4982 (neon_vld3_lane<mode>): Likewise.
4983 (neon_vld3_dup<mode>): Likewise.
4984 (neon_vst3<mode>): Likewise.
4985 (neon_vst3qa<mode>): Likewise.
4986 (neon_vst3qb<mode>): Likewise.
4987 (neon_vst3_lane<mode>): Likewise.
4988 (neon_vst3_lane<mode>): Likewise.
4989 (neon_vld4<mode>): Likewise.
4990 (neon_vld4qa<mode>): Likewise.
4991 (neon_vld4qb<mode>): Likewise.
4992 (neon_vld4_lane<mode>): Likewise.
4993 (neon_vld4_lane<mode>): Likewise.
4994 (neon_vld4_dup<mode>): Likewise.
4995 (neon_vst4<mode>): Likewise.
4996 (neon_vst4qa<mode>): Likewise.
4997 (neon_vst4qb<mode>): Likewise.
4998 (neon_vst4_lane<mode>): Likewise.
4999 (neon_vst4_lane<mode>): Likewise.
5000 (neon_vec_unpack<US>_lo_<mode>): Likewise.
5001 (neon_vec_unpack<US>_hi_<mode>): Likewise.
5002 (neon_vec_<US>mult_lo_<mode>): Likewise.
5003 (neon_vec_<US>mult_hi_<mode>): Likewise.
5004 (neon_vec_<US>shiftl_<mode>): Likewise.
5005 (neon_unpack<US>_<mode>): Likewise.
5006 (neon_vec_<US>mult_<mode>): Likewise.
5007 (vec_pack_trunc_<mode>): Likewise.
5008 (neon_vec_pack_trunc_<mode>): Likewise.
5009 (neon_vabd<mode>_2): Likewise.
5010 (neon_vabd<mode>_3): Likewise.
5012 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5014 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
5015 (load_pair): Update type attribute.
5016 (store_pair): Update type attribute.
5017 * config/aarch64/iterators.md (q): New.
5019 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5021 * config/arm/types.md: Add new types for Neon insns.
5023 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5024 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5025 Sergey Lega <sergey.s.lega@intel.com>
5026 Anna Tikhonova <anna.tikhonova@intel.com>
5027 Ilya Tocar <ilya.tocar@intel.com>
5028 Andrey Turetskiy <andrey.turetskiy@intel.com>
5029 Ilya Verbin <ilya.verbin@intel.com>
5030 Kirill Yukhin <kirill.yukhin@intel.com>
5031 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5033 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
5034 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
5035 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
5037 (srcp14<mode>): Ditto.
5038 (rsqrt14<mode>): Ditto.
5039 (rsqrt14<mode>): Ditto.
5040 (avx512f_vmscalef<mode>): Ditto.
5041 (avx512f_scalef<mode>): Ditto.
5042 (avx512f_getexp<mode>): Ditto.
5043 (avx512f_sgetexp<mode>): Ditto.
5044 (avx512f_fixupimm<mode>): Ditto.
5045 (avx512f_sfixupimm<mode>): Ditto.
5046 (avx512f_rndscale<mode>): Ditto.
5047 (*avx512er_exp2<mode>): Ditto.
5048 (*avx512er_rcp28<mode>): Ditto.
5049 (avx512er_rsqrt28<mode>): Ditto.
5050 (avx512f_getmant<mode>): Ditto.
5051 (avx512f_getmant<mode>): Ditto.
5052 (avx512f_rndscale<mode>): Fix formatting.
5054 2013-10-15 Martin Jambor <mjambor@suse.cz>
5056 * ipa-utils.h (ipa_edge_within_scc): Declare.
5057 * ipa-cp.c (edge_within_scc): Moved...
5058 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
5060 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5061 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5062 Sergey Lega <sergey.s.lega@intel.com>
5063 Anna Tikhonova <anna.tikhonova@intel.com>
5064 Ilya Tocar <ilya.tocar@intel.com>
5065 Andrey Turetskiy <andrey.turetskiy@intel.com>
5066 Ilya Verbin <ilya.verbin@intel.com>
5067 Kirill Yukhin <kirill.yukhin@intel.com>
5068 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5070 * config/i386/predicates.md (const_8_to_15_operand): New.
5071 (const_16_to_31_operand): Ditto.
5072 * config/i386/sse.md (V8FI): New.
5074 (reduc_splus_v8df): Ditto.
5075 (reduc_splus_v16sf): Ditto.
5076 (avx512f_vextract<shuffletype>32x4_1): Ditto.
5077 (vec_extract_hi_<mode>): Ditto.
5078 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
5079 (vec_set_lo_<mode>): Ditto.
5080 (vec_set_hi_<mode>): Ditto.
5081 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
5082 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
5083 (avx512f_pshufd_1): Ditto.
5084 (avx512f_broadcast<mode>): Ditto.
5085 (avx512f_broadcast<mode>): Ditto.
5086 (define_split): Split vec_extract_lo into move.
5087 (ssequartermode): Ditto.
5088 (ssedoublemode): Extened with wider modes.
5089 (vec_extract_lo_<mode>): Ditto.
5091 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5092 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5093 Sergey Lega <sergey.s.lega@intel.com>
5094 Anna Tikhonova <anna.tikhonova@intel.com>
5095 Ilya Tocar <ilya.tocar@intel.com>
5096 Andrey Turetskiy <andrey.turetskiy@intel.com>
5097 Ilya Verbin <ilya.verbin@intel.com>
5098 Kirill Yukhin <kirill.yukhin@intel.com>
5099 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5101 * config/i386/predicates.md (register_or_constm1_operand): New.
5102 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
5103 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
5104 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
5105 UNSPEC_SCATTER_PREFETCH
5107 (avx512f_ucmp<mode>3): Ditto.
5108 (avx512f_vternlog<mode>): Ditto.
5109 (avx512f_align<mode>): Ditto.
5110 (<shift_insn><mode>3): Ditto.
5111 (avx512f_<rotate>v<mode>): Ditto.
5112 (avx512f_<rotate><mode>): Ditto.
5113 (avx512f_eq<mode>3): Ditto.
5114 (avx512f_eq<mode>3_1): Ditto.
5115 (avx512f_gt<mode>3): Ditto.
5116 (avx512f_testm<mode>3): Ditto.
5117 (avx512f_testnm<mode>3): Ditto.
5118 (avx512pf_gatherpf<mode>): Ditto.
5119 (*avx512pf_gatherpf<mode>_mask): Ditto.
5120 (*avx512pf_gatherpf<mode>): Ditto.
5121 (avx512pf_scatterpf<mode>): Ditto.
5122 (*avx512pf_scatterpf<mode>_mask): Ditto.
5123 (*avx512pf_scatterpf<mode>): Ditto.
5124 (avx512f_vec_dup_gpr<mode>): Ditto.
5125 (clz<mode>2): Ditto.
5126 (conflict<mode>): Ditto.
5127 (REDUC_SMINMAX_MODE): Extened with wider modes.
5128 (reduc_<code>_<mode>): Ditto.
5129 (vlshr<mode>3): Ditto.
5130 (vashl<mode>3): Ditto.
5132 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5133 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5134 Sergey Lega <sergey.s.lega@intel.com>
5135 Anna Tikhonova <anna.tikhonova@intel.com>
5136 Ilya Tocar <ilya.tocar@intel.com>
5137 Andrey Turetskiy <andrey.turetskiy@intel.com>
5138 Ilya Verbin <ilya.verbin@intel.com>
5139 Kirill Yukhin <kirill.yukhin@intel.com>
5140 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5142 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
5145 (avx512fmaskmode): Ditto.
5146 (bcstscalarsuff): Ditto.
5147 (avx512f_blendm<mode>): Ditto.
5148 (cmp_imm_predicate): Ditto.
5149 (avx512f_cmp<mode>3): Ditto.
5150 (avx512f_vec_dup<mode>): Ditto.
5151 (avx512f_vec_dup_mem<mode>): Ditto.
5152 (avx512f_vpermi2var<mode>3): Ditto.
5153 (avx512f_vpermt2var<mode>3): Ditto.
5154 (vec_init<mode>): Ditto.
5155 (avx512f_gathersi<mode>): Ditto.
5156 (*avx512f_gathersi<mode>): Ditto.
5157 (*avx512f_gathersi<mode>_2): Ditto.
5158 (avx512f_gatherdi<mode>): Ditto.
5159 (*avx512f_gatherdi<mode>): Ditto.
5160 (*avx512f_gatherdi<mode>_2): Ditto.
5161 (avx512f_scattersi<mode>): Ditto.
5162 (*avx512f_scattersi<mode>): Ditto.
5163 (avx512f_scatterdi<mode>): Ditto.
5164 (*avx512f_scatterdi<mode>): Ditto.
5165 (sseintprefix): Extened with wider modes.
5166 (VEC_GATHER_IDXSI): Ditto.
5167 (VEC_GATHER_IDXDI): Ditto.
5168 (VEC_GATHER_SRCDI): Ditto.
5170 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
5171 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5173 * config/arm/t-aprofile: New file.
5174 * config.gcc: Handle --with-multilib-list option.
5176 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
5178 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
5181 2013-10-15 Richard Biener <rguenther@suse.de>
5183 * tree-loop-distribution.c (build_empty_rdg): Inline into
5185 (rdg_flag_vertex): Inline into single user.
5186 (rdg_flag_vertex_and_dependent): Likewise.
5187 (build_rdg_partition_for_vertex): Remove processed bitmap.
5188 (rdg_build_partitions): Simplify.
5190 2013-10-15 Richard Biener <rguenther@suse.de>
5192 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
5193 Restructure forwarding through conversions and copies to
5194 avoid performing copy-propagation the wrong way. Adjust
5195 recursion invocations.
5196 (forward_propagate_addr_expr): Add argument stating if we
5197 are recursing from a single-use.
5198 (ssa_forward_propagate_and_combine): Adjust.
5200 2013-10-14 David Malcolm <dmalcolm@redhat.com>
5202 * dumpfile.h (gcc::dump_manager): New class, to hold state
5203 relating to dumpfile management.
5204 (get_dump_file_name): Remove in favor of method of dump_manager.
5205 (dump_initialized_p): Likewise.
5206 (dump_start): Likewise.
5207 (dump_finish): Likewise.
5208 (dump_switch_p): Likewise.
5209 (dump_register): Likewise.
5210 (get_dump_file_info): Likewise.
5211 * context.c (gcc::context::context): Construct the dump_manager
5213 * context.h (gcc::context::get_dumps): New.
5214 (gcc::context::m_dumps): New.
5215 * coverage.c (coverage_init): Port to dump_manager API.
5216 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
5217 (extra_dump_files_in_use): Likewise.
5218 (extra_dump_files_alloced): Likewise.
5219 (gcc::dump_manager::dump_manager): New.
5220 (dump_register): Convert to...
5221 (gcc::dump_manager::dump_register): ...method, replacing
5222 function-static next_dump with m_next_dump field.
5223 (get_dump_file_info): Convert to...
5224 (gcc::dump_manager::get_dump_file_info): ...method.
5225 (get_dump_file_name): Convert to...
5226 (gcc::dump_manager::get_dump_file_name): ...method.
5227 (dump_start): Convert to...
5228 (gcc::dump_manager::dump_start): ...method.
5229 (dump_finish): Convert to...
5230 (gcc::dump_manager::dump_finish): ...method.
5231 (dump_begin): Replace body with...
5232 (gcc::dump_manager::dump_begin): ...new method.
5233 (dump_phase_enabled_p): Convert to...
5234 (gcc::dump_manager::dump_phase_enabled_p): ...method.
5235 (dump_phase_enabled_p): Convert to...
5236 (gcc::dump_manager::dump_phase_enabled_p): ...method.
5237 (dump_initialized_p): Convert to...
5238 (gcc::dump_manager::dump_initialized_p): ...method.
5239 (dump_flag_name): Replace body with...
5240 (gcc::dump_manager::dump_flag_name): ...new method.
5241 (dump_enable_all): Convert to...
5242 (gcc::dump_manager::dump_enable_all): ...new method.
5243 (opt_info_enable_passes): Convert to...
5244 (gcc::dump_manager::opt_info_enable_passes): ...new method.
5245 (dump_switch_p_1): Convert to...
5246 (gcc::dump_manager::dump_switch_p_1): ...new method.
5247 (dump_switch_p): Convert to...
5248 (gcc::dump_manager::dump_switch_p): ...new method.
5249 (opt_info_switch_p): Port to dump_manager API.
5250 (enable_rtl_dump_file): Likewise.
5251 * opts-global.c (handle_common_deferred_options): Port to new
5253 * passes.c (pass_manager::finish_optimization_passes): Likewise.
5254 (pass_manager::register_one_dump_file): Likewise.
5255 (pass_manager::register_pass): Likewise.
5256 (pass_init_dump_file): Likewise.
5257 (pass_fini_dump_file): Likewise.
5258 * statistics.c (statistics_early_init): Likewise.
5260 2013-10-14 Richard Biener <rguenther@suse.de>
5262 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
5263 iterative_hash_canonical_type, gimple_canonical_type_hash,
5264 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
5265 gimple_register_canonical_type, print_gimple_types_stats,
5266 free_gimple_type_tables): Move to lto/lto.c
5267 (gt-gimple.h): Do not include.
5268 * gimple.h (gimple_register_canonical_type,
5269 print_gimple_types_stats, free_gimple_type_tables): Remove.
5270 * Makefile.in (GTFILES): Remove gimple.c.
5272 2013-10-14 Travis Snoozy <quandary@remstate.com>
5275 * config/msp430/msp430.c (msp430_option_override): Correct thinko
5276 scanning for msp430x targets.
5278 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
5281 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
5282 (registers_ok_for_ldd_peep): Move around.
5283 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
5284 * config/sparc/sparc.md (widening peepholes): Use it.
5286 2013-10-14 Richard Biener <rguenther@suse.de>
5290 * gimple.c (iterative_hash_canonical_type): Make sure to
5291 record the hash into the correct hashtable slot.
5293 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
5295 PR rtl-optimization/58662
5296 * combine.c (try_combine): Take into account death nodes on I2 when
5297 splitting a PARALLEL of two independent SETs. Fix dump message.
5299 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
5302 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
5303 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
5304 * config/sh/t-sh (sh_treg_combine.o): New entry.
5305 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
5306 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
5307 (register_sh_passes): New function. Register sh_treg_combine pass.
5308 (sh_option_override): Invoke it.
5309 (sh_canonicalize_comparison): Handle op0_preserve_value.
5310 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
5311 opportunities. Canonicalize branch condition.
5312 (nott): Allow only if pseudos can be created for non-SH2A.
5314 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
5317 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
5318 (ix86_expand_movmem): Replace copy_addr_to_reg with
5319 ix86_copy_addr_to_reg.
5320 (ix86_expand_setmem): Likewise.
5322 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
5324 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
5325 provided by ix86_fp_compare_mode instead of CCFPUmode.
5327 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
5329 * config/aarch64/arm_neon.h
5330 (vtbx<1,3>_<psu>8): Fix register constriants.
5332 2013-10-11 Jeff Law <law@redhat.com>
5334 PR tree-optimization/58640
5335 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
5336 threading paths that cross over two loop entry points.
5338 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5340 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
5341 handle vector float as well.
5342 (*vsx_le_perm_load_v4si): Likewise.
5343 (*vsx_le_perm_store_v2di): Likewise.
5344 (*vsx_le_perm_store_v4si): Likewise.
5346 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5348 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
5349 directly to circumvent subtract from splat{31} workaround.
5350 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
5352 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
5353 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
5354 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
5355 (altivec_vperm_<mode>): Convert to define_insn_and_split to
5356 separate big and little endian logic.
5357 (*altivec_vperm_<mode>_internal): New define_insn.
5358 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
5359 separate big and little endian logic.
5360 (*altivec_vperm_<mode>_uns_internal): New define_insn.
5361 (vec_permv16qi): Add little endian logic.
5363 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
5365 * doc/extend.texi (returns_nonnull): Remove arguments.
5367 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5368 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5369 Sergey Lega <sergey.s.lega@intel.com>
5370 Anna Tikhonova <anna.tikhonova@intel.com>
5371 Ilya Tocar <ilya.tocar@intel.com>
5372 Andrey Turetskiy <andrey.turetskiy@intel.com>
5373 Ilya Verbin <ilya.verbin@intel.com>
5374 Kirill Yukhin <kirill.yukhin@intel.com>
5375 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5377 * config/i386/sse.md (VI48F_256_512): New.
5378 (avx2_permvar<mode>): Change to ...
5379 (<avx2_avx512f>_permvar<mode>): This.
5381 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5382 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5383 Sergey Lega <sergey.s.lega@intel.com>
5384 Anna Tikhonova <anna.tikhonova@intel.com>
5385 Ilya Tocar <ilya.tocar@intel.com>
5386 Andrey Turetskiy <andrey.turetskiy@intel.com>
5387 Ilya Verbin <ilya.verbin@intel.com>
5388 Kirill Yukhin <kirill.yukhin@intel.com>
5389 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5391 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
5392 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
5393 * config/i386/sse.md (VI4_AVX): New.
5394 (sf2simodelower): Ditto.
5395 (sse2_cvtps2dq): Change to ...
5396 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
5398 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5399 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5400 Sergey Lega <sergey.s.lega@intel.com>
5401 Anna Tikhonova <anna.tikhonova@intel.com>
5402 Ilya Tocar <ilya.tocar@intel.com>
5403 Andrey Turetskiy <andrey.turetskiy@intel.com>
5404 Ilya Verbin <ilya.verbin@intel.com>
5405 Kirill Yukhin <kirill.yukhin@intel.com>
5406 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5408 * config/i386/sse.md (V_512): New.
5410 (vcond<V_512:mode><VF_512:mode>): Ditto.
5411 (vcond<V_512:mode><VI_512:mode>): Ditto.
5412 (vcondu<V_512:mode><VI_512:mode>): Ditto.
5414 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5415 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5416 Sergey Lega <sergey.s.lega@intel.com>
5417 Anna Tikhonova <anna.tikhonova@intel.com>
5418 Ilya Tocar <ilya.tocar@intel.com>
5419 Andrey Turetskiy <andrey.turetskiy@intel.com>
5420 Ilya Verbin <ilya.verbin@intel.com>
5421 Kirill Yukhin <kirill.yukhin@intel.com>
5422 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5424 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
5425 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
5427 (fma<mode>4): Removed condition.
5428 (fms<mode>4): Ditto.
5429 (fnma<mode>4): Ditto.
5430 (fnms<mode>4): Ditto.
5431 (fma4i_fmadd_<mode>): Ditto.
5432 (*fma_fmadd_<mode>): Ditto.
5433 (*fma_fmsub_<mode>): Ditto.
5434 (*fma_fnmadd_<mode>): Ditto.
5435 (*fma_fnmsub_<mode>): Ditto.
5436 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
5437 (*fma_fmaddsub_<mode>): Ditto.
5438 (*fma_fmsubadd_<mode>): Ditto.
5439 (*fmai_fmadd_<mode>): Ditto.
5440 (*fmai_fmsub_<mode>): Ditto.
5441 (*fmai_fnmadd_<mode>): Ditto.
5442 (*fmai_fnmsub_<mode>): Ditto.
5444 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5445 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5446 Sergey Lega <sergey.s.lega@intel.com>
5447 Anna Tikhonova <anna.tikhonova@intel.com>
5448 Ilya Tocar <ilya.tocar@intel.com>
5449 Andrey Turetskiy <andrey.turetskiy@intel.com>
5450 Ilya Verbin <ilya.verbin@intel.com>
5451 Kirill Yukhin <kirill.yukhin@intel.com>
5452 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5454 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
5455 (VI124_256): Changed to ...
5456 (VI124_256_48_512): This.
5457 (ssepackmode): Extended with wider modes.
5458 (<code><mode>3): Changed iterator.
5459 (*avx2_<code><mode>3): Ditto.
5460 (vec_pack_trunc_<mode>): Ditto.
5462 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5463 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5464 Sergey Lega <sergey.s.lega@intel.com>
5465 Anna Tikhonova <anna.tikhonova@intel.com>
5466 Ilya Tocar <ilya.tocar@intel.com>
5467 Andrey Turetskiy <andrey.turetskiy@intel.com>
5468 Ilya Verbin <ilya.verbin@intel.com>
5469 Kirill Yukhin <kirill.yukhin@intel.com>
5470 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5472 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
5473 (VI8F_256_512): Ditto.
5474 (abs<mode>2): Changed iterator.
5475 (avx2_perm<mode>): Changed to ...
5476 (<avx2_avx512f>_perm<mode>): This.
5477 (avx2_perm<mode>_1): Changed to ...
5478 (<avx2_avx512f>_perm<mode>_1): This.
5480 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5481 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5482 Sergey Lega <sergey.s.lega@intel.com>
5483 Anna Tikhonova <anna.tikhonova@intel.com>
5484 Ilya Tocar <ilya.tocar@intel.com>
5485 Andrey Turetskiy <andrey.turetskiy@intel.com>
5486 Ilya Verbin <ilya.verbin@intel.com>
5487 Kirill Yukhin <kirill.yukhin@intel.com>
5488 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5490 * config/i386/sse.md (VI48_AVX512F): New.
5491 (VI48_AVX2): Changed to ...
5492 (VI48_AVX2_48_AVX512F): This.
5493 (avx2_ashrv<mode>): Changed to ...
5494 (<avx2_avx512f>_ashrv<mode>): This.
5495 (avx2_<shift_insn>v<mode>): Changed to ...
5496 (<avx2_avx512f>_<shift_insn>v<mode>): This.
5498 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5499 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5500 Sergey Lega <sergey.s.lega@intel.com>
5501 Anna Tikhonova <anna.tikhonova@intel.com>
5502 Ilya Tocar <ilya.tocar@intel.com>
5503 Andrey Turetskiy <andrey.turetskiy@intel.com>
5504 Ilya Verbin <ilya.verbin@intel.com>
5505 Kirill Yukhin <kirill.yukhin@intel.com>
5506 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5508 * config/i386/sse.md (VI4_AVX512F): New.
5509 (VI8_AVX2_AVX512F): Ditto.
5510 (mul<mode>3): Extended with wider modes.
5511 (*<sse4_1_avx2>_mul<mode>3): Ditto.
5512 (mul<mode>3): Ditto.
5513 (vec_widen_<s>mult_odd_<mode>): Ditto.
5515 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5516 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5517 Sergey Lega <sergey.s.lega@intel.com>
5518 Anna Tikhonova <anna.tikhonova@intel.com>
5519 Ilya Tocar <ilya.tocar@intel.com>
5520 Andrey Turetskiy <andrey.turetskiy@intel.com>
5521 Ilya Verbin <ilya.verbin@intel.com>
5522 Kirill Yukhin <kirill.yukhin@intel.com>
5523 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5525 * config/i386/sse.md (VI2_AVX512F): New.
5526 (VI124_AVX512F): Ditto.
5527 (sseunpackmode): Extended with wider modes.
5528 (sseunpackfltmode): Ditto.
5529 (vec_unpacks_float_hi_<mode>): Ditto.
5530 (vec_unpacks_float_lo_<mode>): Ditto.
5531 (vec_unpacku_float_hi_<mode>): Ditto.
5532 (vec_unpacku_float_lo_<mode>): Ditto.
5533 (vec_unpacks_lo_<mode>): Ditto.
5534 (vec_unpacks_hi_<mode>): Ditto.
5535 (vec_unpacku_lo_<mode>): Ditto.
5536 (vec_unpacku_hi_<mode>): Ditto.
5538 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5539 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5540 Sergey Lega <sergey.s.lega@intel.com>
5541 Anna Tikhonova <anna.tikhonova@intel.com>
5542 Ilya Tocar <ilya.tocar@intel.com>
5543 Andrey Turetskiy <andrey.turetskiy@intel.com>
5544 Ilya Verbin <ilya.verbin@intel.com>
5545 Kirill Yukhin <kirill.yukhin@intel.com>
5546 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5548 * config/i386/i386.md (multdiv): New.
5549 (multdiv_mnemonic): Ditto.
5550 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
5551 (<sse>_vm<multdiv_mnemonic><mode>3): This.
5552 (<sse>_vmdiv<mode>3): Removed.
5554 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5555 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5556 Sergey Lega <sergey.s.lega@intel.com>
5557 Anna Tikhonova <anna.tikhonova@intel.com>
5558 Ilya Tocar <ilya.tocar@intel.com>
5559 Andrey Turetskiy <andrey.turetskiy@intel.com>
5560 Ilya Verbin <ilya.verbin@intel.com>
5561 Kirill Yukhin <kirill.yukhin@intel.com>
5562 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5564 * config/i386/sse.md (V): Extended with wider modes.
5566 (ssehalfvecmode): Ditto.
5568 (ssepackfltmode): Ditto.
5569 (avx_vec_concat<mode>): Ditto.
5570 (V_256_512): New iterator.
5571 (VF2_512_256): Ditto.
5572 (si2dfmode): New attribute.
5573 (si2dfmodelower): Ditto.
5575 (concat_tg_mode): Ditto.
5576 (floatv4siv4df2): Changed to ...
5577 (float<si2dfmodelower><mode>2): This.
5578 (avx_cvtps2pd256): Changed to ...
5579 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
5580 (vec_pack_trunc_v4df): Changed to ...
5581 (vec_pack_trunc_<mode>): This.
5582 (avx_vpermil<mode>): Changed to ...
5583 (<sse2_avx_avx512f>_vpermil<mode>): This.
5584 (<fixsuffix>fix_truncv8dfv8si2): New.
5585 (vec_pack_sfix_trunc_v8df): Ditto.
5586 (avx512f_rndscale<mode>): Ditto.
5587 (avx512f_roundpd512): Ditto.
5588 (vec_pack_ufix_trunc_<mode>): Updated iterator.
5590 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5591 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5592 Sergey Lega <sergey.s.lega@intel.com>
5593 Anna Tikhonova <anna.tikhonova@intel.com>
5594 Ilya Tocar <ilya.tocar@intel.com>
5595 Andrey Turetskiy <andrey.turetskiy@intel.com>
5596 Ilya Verbin <ilya.verbin@intel.com>
5597 Kirill Yukhin <kirill.yukhin@intel.com>
5598 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5600 * config/i386/i386.md (any_fix): New iterator.
5601 (fixsuffix): New attribute.
5602 * config/i386/sse.md (VF1): Extened with wider modes.
5606 (sseintvecmodelower): Ditto.
5607 (ssescalarmode): Ditto.
5608 (ssescalarnum): Ditto.
5610 (ssexmmmode): Ditto.
5611 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
5612 (<sse>_rcp<mode>2): Change iterator.
5613 (rsqrt<mode>2): Ditto.
5614 (<sse>_rsqrt<mode>2): Ditto.
5615 (avx2_vec_dup<mode>): Ditto.
5616 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
5617 (round<mode>2_sfix): Ditto.
5618 (avx2_pbroadcast<mode>): Ditto.
5619 (*andnot<mode>3): Handle XI mode.
5620 (*<code><mode>3): Ditto.
5621 (AVXTOSSEMODE): Removed.
5622 (avx_vpermil<mode>): Changed to ...
5623 (<sse2_avx_avx512f>_vpermil<mode>): This.
5625 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5626 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5627 Sergey Lega <sergey.s.lega@intel.com>
5628 Anna Tikhonova <anna.tikhonova@intel.com>
5629 Ilya Tocar <ilya.tocar@intel.com>
5630 Andrey Turetskiy <andrey.turetskiy@intel.com>
5631 Ilya Verbin <ilya.verbin@intel.com>
5632 Kirill Yukhin <kirill.yukhin@intel.com>
5633 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5635 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
5636 (<sse>_comi): Ditto.
5637 (<sse>_ucomi): Ditto.
5638 (sse_cvtss2siq_2): Ditto.
5639 (sse2_cvtsd2si): Ditto.
5640 (sse2_cvtsd2siq): Ditto.
5641 (sse2_cvttsd2si): Ditto.
5642 (sse2_cvttsd2siq): Ditto.
5643 (<shift_insn><mode>3): Ditto.
5644 (sse2_cvtsi2sdq): Update constraint and prefix.
5645 (sse_cvtsi2ss): Update prefix.
5646 (sse_cvtsi2ssq): Ditto.
5648 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5650 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
5651 ignore internal calls.
5653 2013-10-11 Richard Biener <rguenther@suse.de>
5655 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
5656 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
5657 TREE_OVERFLOW is set.
5659 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
5661 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
5663 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
5664 * gimple.h: Likewise.
5666 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
5668 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
5671 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
5673 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
5674 * configure: Regenerate.
5676 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5678 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
5679 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
5680 formatting fixes, use pp_colon instead of pp_character (..., ':'),
5681 similarly pp_right_paren.
5682 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
5683 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
5684 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
5685 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
5687 (omp_declare_simd_clauses_equal,
5688 omp_remove_redundant_declare_simd_attrs): New functions.
5689 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
5690 (walk_tree_1): Handle new OpenMP 4.0 clauses.
5691 * tree.h (OMP_LOOP_CHECK): Define.
5692 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
5693 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
5694 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
5695 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
5696 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
5697 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
5698 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
5699 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
5700 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
5701 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
5702 OMP_CLAUSE_SIMDLEN_EXPR): Define.
5703 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
5704 (omp_remove_redundant_declare_simd_attrs): New prototype.
5705 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
5706 GIMPLE_OMP_TEAMS): New codes.
5707 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
5708 * omp-low.c (struct omp_context): Add cancel_label and cancellable
5710 (target_nesting_level): New variable.
5711 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
5712 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
5713 for collapse > 1 static schedule unless ordered.
5714 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
5715 (determine_parallel_type): Adjust get_ws_args_for caller.
5716 (install_var_field): Handle mask & 4 for double indirection.
5717 (scan_sharing_clauses): Ignore shared clause on teams construct.
5718 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
5719 (create_omp_child_function): If inside target or declare target
5720 constructs, set "omp declare target" attribute on the child function.
5721 (find_combined_for): New function.
5722 (scan_omp_parallel): Handle combined loops.
5723 (scan_omp_target, scan_omp_teams): New functions.
5724 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
5725 restrictions and set ctx->cancellable for cancellable constructs.
5726 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
5727 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
5728 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
5729 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
5730 (omp_clause_aligned_alignment): New function.
5731 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
5732 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
5733 on teams constructs. Handle user defined reductions and new
5735 (lower_reduction_clauses): Don't set placeholder to address of ref
5736 if it has already the right type.
5737 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
5738 (expand_parallel_call): Use the new non-_start suffixed builtins,
5739 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
5740 and GOMP_parallel_end after the call.
5741 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
5742 (expand_omp_for_init_counts): Handle combined loops.
5743 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
5745 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
5746 end of cancellable loops.
5747 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
5748 Likewise. Handle collapse > 1 loops.
5749 (expand_omp_simd): Handle combined loops.
5750 (expand_omp_for): Add inner_stmt argument, adjust callers of
5751 expand_omp_for* functions, use expand_omp_for_static*chunk even
5752 for collapse > 1 unless ordered.
5753 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
5754 of cancellable sections.
5755 (expand_omp_single): Remove need_barrier variable, just rely on
5756 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
5757 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
5758 (expand_omp_atomic_load, expand_omp_atomic_store,
5759 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
5760 (expand_omp_target): New function.
5761 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
5762 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
5763 (build_omp_regions_1): Immediately close region for
5764 GF_OMP_TARGET_KIND_UPDATE.
5765 (maybe_add_implicit_barrier_cancel): New function.
5766 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
5768 (lower_omp_single): Likewise. Add clobber after the barrier.
5769 (lower_omp_taskgroup): New function.
5770 (lower_omp_for): Handle combined loops. Adjust
5771 lower_rec_input_clauses caller. Handle cancellation.
5772 (lower_depend_clauses): New function.
5773 (lower_omp_taskreg): Lower depend clauses. Adjust
5774 lower_rec_input_clauses caller. Add clobber after the call. Handle
5776 (lower_omp_target, lower_omp_teams): New functions.
5777 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
5778 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
5779 and GOMP_cancellation_point calls.
5780 (lower_omp): Fold stmts inside of target region.
5781 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
5782 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5783 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
5784 (BT_FN_VOID_OMPFN_PTR_UINT,
5785 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
5786 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
5787 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
5788 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
5789 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
5790 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
5791 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
5792 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
5793 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
5794 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
5795 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
5796 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
5797 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
5798 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
5799 BUILT_IN_GOMP_PARALLEL_END.
5800 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
5801 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5802 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
5803 GF_OMP_FOR_KIND_DISTRIBUTE.
5804 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
5805 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
5806 (dump_gimple_omp_return): Print lhs if it has any.
5807 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
5808 gimple_omp_atomic_seq_cst_p.
5809 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
5810 and GIMPLE_OMP_TEAMS.
5811 * langhooks.c (lhd_omp_mappable_type): New function.
5812 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
5813 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
5815 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
5816 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
5817 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
5818 (struct gimplify_omp_ctx): Add combined_loop field.
5819 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
5820 on stmts inside of target region.
5821 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
5822 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
5823 ORT_TARGET and ORT_TARGET_DATA.
5824 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
5826 (omp_notice_threadprivate_variable): Complain about threadprivate
5827 variables in target region.
5828 (omp_notice_variable): Complain about vars with non-mappable type
5829 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
5830 (omp_check_private): Ignore ORT_TARGET* regions.
5831 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
5832 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
5833 (find_combined_omp_for): New function.
5834 (gimplify_omp_for): Handle gimplification of combined loops.
5835 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
5837 (gimplify_omp_target_update): New function.
5838 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
5839 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
5840 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
5841 (gimplify_body): If fndecl has "omp declare target" attribute, add
5842 implicit ORT_TARGET context around it.
5843 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
5844 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
5845 * tree-nested.c (convert_nonlocal_reference_stmt,
5846 convert_local_reference_stmt, convert_gimple_call): Handle
5847 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5848 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
5849 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5850 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5851 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5852 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5853 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5854 BUILT_IN_GOMP_LOOP_END_CANCEL,
5855 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5856 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5857 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5858 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5859 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5860 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5861 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5862 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5863 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5864 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5865 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5866 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5867 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5868 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5869 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5870 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5871 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5872 gimple_build_omp_teams): New functions.
5873 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5874 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5875 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5876 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5877 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5878 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5879 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5880 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5881 GF_OMP_ATOMIC_SEQ_CST): New.
5882 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5883 gimple_build_omp_teams): New prototypes.
5884 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5885 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5886 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5887 GIMPLE_OMP_SINGLE as end of range.
5888 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5889 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5890 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5891 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5892 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5893 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5894 gimple_omp_target_kind, gimple_omp_target_set_kind,
5895 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5896 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5897 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5898 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5899 gimple_omp_teams_set_clauses): New inlines.
5900 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5901 and GIMPLE_OMP_TASKGROUP.
5902 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5903 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5904 enum omp_clause_proc_bind_kind): New.
5905 (union omp_clause_subcode): Add depend_kind, map_kind and
5906 proc_bind_kind fields.
5907 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5908 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5909 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5910 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5911 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5913 2013-10-10 Teresa Johnson <tejohnson@google.com>
5915 * predict.c (tree_estimate_probability): Add new parameter
5916 for estimate_bb_frequencies.
5917 (estimate_bb_frequencies): Add new parameter to force estimation.
5918 (rebuild_frequencies): When max frequency in function is small,
5919 recompute counts from frequencies.
5920 * predict.h (estimate_bb_frequencies): New parameter.
5922 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5924 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5925 optimizations are disabled.
5927 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5929 * coverage.c (coverage_finish): Fix leak of da_file_name.
5931 2013-10-10 Jan Hubicka <jh@suse.cz>
5933 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5934 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5937 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5940 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5941 if any labels are in FALLTHRU_BB, use a special label emitted
5942 immediately after the asm goto insn rather than label_rtx
5944 (expand_asm_stmt): Adjust caller.
5945 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5946 edge if the last insn in predecessor is a jump with single successor,
5947 but it isn't simplejump_p.
5949 2013-10-10 Richard Biener <rguenther@suse.de>
5951 PR tree-optimization/58656
5952 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5954 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5956 * gimplify.c: Include expr.h and tm_p.h for targets with special
5957 va-arg padding requirements.
5959 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5961 * tree-flow.h: Move some prototypes to gimple.h.
5962 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5963 * gimple.h: Relocate some prototypes from tree-flow.h
5964 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5966 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5967 (build_va_arg_indirect_ref): Relocate and make static.
5968 (std_gimplify_va_arg_expr): Relocate here.
5969 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5970 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5972 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5974 * doc/md.texi: Document the mnemonic attribute.
5976 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5979 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5980 (set_attr_alternative x ...) when searching for user defined
5983 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5985 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5987 2013-10-09 Easwaran Raman <eraman@google.com>
5989 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5990 * cfgexpand.c (defer_stack_allocation): ...use here
5991 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5993 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
5995 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
5996 (operand_equal_for_value_replacement): New function, extracted from
5997 value_replacement and enhanced to catch more cases.
5998 (value_replacement): Use operand_equal_for_value_replacement.
6000 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6002 * loop-doloop.c (doloop_modify, doloop_optimize): Use
6003 get_max_loop_iterations.
6005 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6007 * config/arm/aarch-common.c (arm_early_load_addr_dep):
6008 Place comment above function.
6010 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6012 * tree-flow.h: Remove all remaining prototypes, enums and structs that
6013 are not related to tree-cfg.c.
6014 * tree-ssa-address.h: New file. Relocate prototypes.
6015 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
6016 (addr_for_mem_ref): New. Combine call to get_address_description and
6017 return addr_for_mem_ref.
6018 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
6019 * tree-ssa-live.h: Adjust prototypes.
6020 * passes.c: Include tree-ssa-live.h.
6021 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
6022 * graphite.c (graphite_transform_loops): Make static.
6023 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
6024 make_pass_graphite, pass_data_graphite_transforms,
6025 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
6026 * ipa-pure-const.c (warn_function_noreturn): Make static.
6027 (execute_warn_function_noreturn, gate_warn_function_noreturn,
6028 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
6029 Relocate from tree-cfg.c
6030 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
6032 (execute_warn_function_noreturn, gate_warn_function_noreturn,
6033 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
6034 Move to ipa-pure-const.c.
6035 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
6036 Relocate from tree-optimize.c.
6037 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
6038 make_pass_fixup_cfg): Move to tree-cfg.c.
6039 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
6040 Relocate some prototypes.
6041 * tree-data-ref.h (tree_check_data_deps) Add prototype.
6042 * tree-dump.c (dump_function_to_file): Remove prototype.
6043 Add tree-flow.h to the include file.
6044 * tree-dump.h: Remove prototype.
6045 * tree-parloops.h: New File. Add prototypes.
6046 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
6047 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
6048 from tree-ssa-loop.c.
6049 * tree-predcom.c (run_tree_predictive_commoning,
6050 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
6051 Relocate here from tree-ssa-loop.c.
6052 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
6053 ssa_name_values.release ().
6054 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
6055 (ssa_name_values): Relocate from tree-flow.h.
6056 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
6057 * tree-ssa-loop.c (run_tree_predictive_commoning,
6058 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
6059 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
6060 make_pass_graphite, pass_data_graphite_transforms,
6061 make_pass_graphite_transforms, gate_tree_parallelize_loops,
6062 tree_parallelize_loops, pass_data_parallelize_loops,
6063 make_pass_parallelize_loops): Move to other files.
6064 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
6066 * tree.h: Remove prototypes from tree-address.c.
6068 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6070 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
6071 (struct int_tree_map): Move to tree-hasher.h
6072 (SCALE, LABEL, PERCENT): Move to gimple.h
6073 * tree-flow-inline.h: Delete. Move functions to other files.
6074 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
6075 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
6076 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
6077 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
6078 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
6079 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
6080 * tree-hasher.h: Don't include tree-flow.h.
6081 (struct int_tree_map): Relocate from tree-flow.h.
6082 * tree-sra.c (contains_view_convert_expr_p): Relocate from
6083 tree-flow-inline.h and make static.
6084 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
6086 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
6087 tree-flow-inline.h and make static.
6088 * tree.h (is_global_var, may_be_aliased): Relocate from
6090 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
6091 * value-prof.c: No longer include tree-flow-inline.h.
6092 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
6094 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6096 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
6097 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
6098 (enum move_pos): Move to tree-ssa-loop-im.h
6099 * cfgloop.h: Move some prototypes.
6100 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
6101 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
6102 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
6103 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
6104 (loop_containing_stmt): Relocate from tree-flow-inline.h.
6105 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
6106 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
6107 (enum move_pos): Relocate here.
6108 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
6110 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
6111 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
6112 make_pass_lim): Relocate here from tree-ssa-loop.c.
6113 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
6115 (loop_edge_to_cancel, unloop_loops): Make static.
6116 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
6117 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
6118 (tree_complete_unroll, gate_tree_complete_unroll,
6119 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
6120 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
6121 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
6122 * tree-ssa-loop-ivopts.c: Remove local prototypes.
6123 (stmt_invariant_in_loop_p): Remove unused function.
6124 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
6125 * tree-ssa-loop-manip.h: New file. Add prototypes.
6126 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
6127 (gcov_type_to_double_int): Move to cfgloop.h.
6128 (double_int_cmp, bound_index,
6129 estimate_numbers_of_iterations_loop): Make static.
6130 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
6131 (max_loop_iterations): Factor out get_max_loop_iterations.
6132 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
6134 * tree-ssa-loop-niter.h: New file. Add prototypes.
6135 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
6136 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
6137 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
6138 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
6139 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
6140 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
6141 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
6142 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
6143 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
6144 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
6145 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
6146 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
6147 pass_data_complete_unroll, make_pass_complete_unroll,
6148 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
6149 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
6150 tree-ssa-loop-ivcanon.c.
6151 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
6152 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
6153 tree-ssa-loop-prefetch.c.
6154 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
6156 (get_lsm_tmp_name): Relocate and add suffix parameter.
6157 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
6158 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
6159 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
6160 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
6161 (get_estimated_loop_iterations): Factor out accessor from
6162 estimated_loop_iterations in tree-ssa-loop-niter.c.
6163 (get_max_loop_iterations): Factor out accessor from
6164 _max_loop_iterations in tree-ssa-niter.c.
6165 * loop-unroll.c (decide_unroll_constant_iterations,
6166 decide_unroll_runtime_iterations, decide_peel_simple,
6167 decide_unroll_stupid): Use new get_* accessors.
6169 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
6171 PR tree-optimization/20318
6172 * doc/extend.texi (returns_nonnull): New function attribute.
6173 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
6175 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
6176 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
6178 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
6181 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
6182 false if both components are bitfields.
6184 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6186 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
6187 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
6188 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
6189 * config/aarch64/aarch64.h
6190 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
6191 * config/aarch64/aarch64-simd-builtins.def
6192 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
6194 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6196 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
6198 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6200 * config/aarch64/arm_neon.h (vdiv_f64): Added.
6202 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6204 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
6205 (vneg_f64): New intrinsic.
6206 (vneg_s8): Asm replaced with C.
6207 (vneg_s16): Likewise.
6208 (vneg_s32): Likewise.
6209 (vneg_s64): New intrinsic.
6210 (vnegq_f32): Asm replaced with C.
6211 (vnegq_f64): Likewise.
6212 (vnegq_s8): Likewise.
6213 (vnegq_s16): Likewise.
6214 (vnegq_s32): Likewise.
6215 (vnegq_s64): Likewise.
6217 2013-10-09 Renlin Li <Renlin.Li@arm.com>
6219 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
6221 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6223 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
6224 packed stack special handling.
6225 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
6226 back to fixed stack slots for FPRs saved due to stdarg.
6228 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6230 * config/s390/s390.c (s390_frame_info): Restructure function.
6232 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6234 * config/s390/s390.c (struct s390_frame_layout): New field
6236 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
6237 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
6238 regs_ever_clobbered to char*.
6239 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
6240 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
6241 (s390_register_info): Enable FPR save slots. Move/Copy some
6242 functionality into ...
6243 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
6244 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
6246 (s390_frame_info): Do gpr slot allocation here now. stdarg does
6247 not imply a stack frame.
6248 (s390_init_frame_layout): Remove variable clobbered_regs.
6249 (s390_update_register_info): Remove function.
6250 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
6251 cannot be used for register renaming.
6252 (s390_hard_regno_scratch_ok): New function.
6253 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
6254 (s390_initial_elimination_offset): Change offset calculation of
6255 the return address pointer.
6256 (save_gprs): Deal with only r6 being saved from the call-saved regs.
6257 (restore_gprs): Set frame related flag.
6258 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
6259 (s390_emit_prologue): Call s390_register_info instead of
6260 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
6261 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
6262 (s390_optimize_prologue): Call s390_optimize_register_info.
6263 Try to remove also FPR slot save/restore INSNs. Remove frame
6264 related flags from restore INSNs.
6266 2013-10-08 DJ Delorie <dj@redhat.com>
6268 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
6271 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
6272 avoid conflict with the MI use of %c.
6273 * config/rl78/rl78-real.md: change %c to %C throughout.
6274 * config/rl78/rl78-virt.md: Likewise.
6276 2013-10-08 Jan Hubicka <jh@suse.cz>
6278 * config/i386/i386.c (ix86_option_override_internal): Switch
6279 to SSE math for -ffast-math when target ISA supports SSE2.
6281 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
6283 * tree-flow.h: Remove some prototypes.
6284 * tree.h: Remove some protypes, add a couple.
6285 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
6286 using_eh_for_cleanups_p): Add interface routines for front ends.
6287 * tree-eh.h: New file. Add protoptyes.
6288 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
6289 (add_stmt_to_eh_lp_fn): Make static.
6290 (lower_try_finally): Use new using_eh_for_cleanups_p.
6291 * emit-rtl.c: Include tree-eh.h.
6292 * gimple.h: Include tree-eh.h.
6294 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
6296 PR tree-optimization/58480
6297 * tree-vrp.c (infer_nonnull_range): New function.
6298 (infer_value_range): Call infer_nonnull_range.
6300 2013-10-08 Dehao Chen <dehao@google.com>
6302 PR tree-optimization/58619
6303 * tree-inline.c (copy_phis_for_bb): Combine location data
6306 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6309 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
6310 RTX_FRAME_RELATED_P on INSN.
6312 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6314 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
6315 (altivec_expand_vec_perm_const): Call it.
6317 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6319 * config/rs6000/vector.md (mov<mode>): Emit permuted move
6320 sequences for LE VSX loads and stores at expand time.
6321 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
6323 * config/rs6000/rs6000.c (rs6000_const_vec): New.
6324 (rs6000_gen_le_vsx_permute): New.
6325 (rs6000_gen_le_vsx_load): New.
6326 (rs6000_gen_le_vsx_store): New.
6327 (rs6000_gen_le_vsx_move): New.
6328 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
6329 (*vsx_le_perm_load_v4si): New.
6330 (*vsx_le_perm_load_v8hi): New.
6331 (*vsx_le_perm_load_v16qi): New.
6332 (*vsx_le_perm_store_v2di): New.
6333 (*vsx_le_perm_store_v4si): New.
6334 (*vsx_le_perm_store_v8hi): New.
6335 (*vsx_le_perm_store_v16qi): New.
6336 (*vsx_xxpermdi2_le_<mode>): New.
6337 (*vsx_xxpermdi4_le_<mode>): New.
6338 (*vsx_xxpermdi8_le_V8HI): New.
6339 (*vsx_xxpermdi16_le_V16QI): New.
6340 (*vsx_lxvd2x2_le_<mode>): New.
6341 (*vsx_lxvd2x4_le_<mode>): New.
6342 (*vsx_lxvd2x8_le_V8HI): New.
6343 (*vsx_lxvd2x16_le_V16QI): New.
6344 (*vsx_stxvd2x2_le_<mode>): New.
6345 (*vsx_stxvd2x4_le_<mode>): New.
6346 (*vsx_stxvd2x8_le_V8HI): New.
6347 (*vsx_stxvd2x16_le_V16QI): New.
6349 2013-10-07 Renlin Li <Renlin.Li@arm.com>
6351 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
6353 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6355 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
6356 loop to work also for 31bit ABI.
6357 Save the stack pointer for frame_size > 0.
6359 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6361 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
6362 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
6363 constraint letters from expanders.
6364 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
6365 retry count to general_operand.
6366 ("tabort"): Give operand 0 a mode.
6367 ("tabort_1"): Add mode and constraint letter for operand 0.
6368 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
6370 2013-10-04 Jeff Law <law@redhat.com>
6372 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
6374 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
6376 (thread_across_edge): Here. Call it.
6378 2013-10-04 Cary Coutant <ccoutant@google.com>
6380 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
6381 discarding a location list expression (or a piece of one).
6383 2013-10-03 Jan Hubicka <jh@suse.cz>
6385 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
6386 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
6387 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
6389 2013-10-03 Jan Hubicka <jh@suse.cz>
6391 * config/i386/i386.c (ix86_option_override_internal): Do not enable
6392 accumulate-outgoing-args when producing unwind info.
6394 2013-10-03 Wei Mi <wmi@google.com>
6396 * lra-constraints.c (insert_move_for_subreg): New function
6397 extracted from simplify_operand_subreg.
6398 (simplify_operand_subreg): Add reload for paradoxical subreg.
6400 2013-10-03 Rong Xu <xur@google.com>
6402 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
6403 the candidate of builtin_expect such that we should fix the
6404 size/time estimation.
6405 (estimate_function_body_sizes): Do the acutally size/time fix-up
6408 2013-10-03 Rong Xu <xur@google.com>
6410 * predict.c (tree_predict_by_opcode): Get the probability
6411 for builtin_expect from param builtin_expect_probability.
6412 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
6413 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
6414 * doc/invoke.texi: Add documentation for builtin-expect-probability.
6416 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
6419 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
6420 * calls.c (alloca_call_p): Use get_callee_fndecl.
6421 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
6422 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
6424 (vrp_visit_stmt): Remove duplicated code.
6426 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
6428 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
6429 ceildf2, btruncdf2, instead of vsx_* name.
6431 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
6432 iterators to only do V2DF and V4SF here. Move the DF code to
6433 rs6000.md where it is combined with SF mode. Replace <VSv> with
6434 just 'v' since only vector operations are handled with these insns
6435 after moving the DF support to rs6000.md.
6436 (vsx_sub<mode>3): Likewise.
6437 (vsx_mul<mode>3): Likewise.
6438 (vsx_div<mode>3): Likewise.
6439 (vsx_fre<mode>2): Likewise.
6440 (vsx_neg<mode>2): Likewise.
6441 (vsx_abs<mode>2): Likewise.
6442 (vsx_nabs<mode>2): Likewise.
6443 (vsx_smax<mode>3): Likewise.
6444 (vsx_smin<mode>3): Likewise.
6445 (vsx_sqrt<mode>2): Likewise.
6446 (vsx_rsqrte<mode>2): Likewise.
6447 (vsx_fms<mode>4): Likewise.
6448 (vsx_nfma<mode>4): Likewise.
6449 (vsx_copysign<mode>3): Likewise.
6450 (vsx_btrunc<mode>2): Likewise.
6451 (vsx_floor<mode>2): Likewise.
6452 (vsx_ceil<mode>2): Likewise.
6453 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
6454 (vsx_sminsf3): Likewise.
6455 (vsx_fmadf4): Likewise.
6456 (vsx_fmsdf4): Likewise.
6457 (vsx_nfmadf4): Likewise.
6458 (vsx_nfmsdf4): Likewise.
6459 (vsx_cmpdf_internal1): Likewise.
6461 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
6462 simpler to select whether a target has SPE or traditional floating
6463 point support in iterators.
6464 (TARGET_DF_SPE): Likewise.
6465 (TARGET_SF_FPR): Likewise.
6466 (TARGET_DF_FPR): Likewise.
6467 (TARGET_SF_INSN): Macros to say whether floating point support
6468 exists for a given operation for expanders.
6469 (TARGET_DF_INSN): Likewise.
6471 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
6472 combining of SF/DF mode operations, using both traditional and VSX
6480 (abs<mode>2): Combine SF/DF modes using traditional floating point
6481 instructions. Add support for using the upper DF registers with
6482 VSX support, and SF registers with power8-vector support. Update
6483 expanders for operations supported by both the SPE and traditional
6484 floating point units.
6485 (abs<mode>2_fpr): Likewise.
6486 (nabs<mode>2): Likewise.
6487 (nabs<mode>2_fpr): Likewise.
6488 (neg<mode>2): Likewise.
6489 (neg<mode>2_fpr): Likewise.
6490 (add<mode>3): Likewise.
6491 (add<mode>3_fpr): Likewise.
6492 (sub<mode>3): Likewise.
6493 (sub<mode>3_fpr): Likewise.
6494 (mul<mode>3): Likewise.
6495 (mul<mode>3_fpr): Likewise.
6496 (div<mode>3): Likewise.
6497 (div<mode>3_fpr): Likewise.
6498 (sqrt<mode>3): Likewise.
6499 (sqrt<mode>3_fpr): Likewise.
6500 (fre<Fs>): Likewise.
6501 (rsqrt<mode>2): Likewise.
6502 (cmp<mode>_fpr): Likewise.
6503 (smax<mode>3): Likewise.
6504 (smin<mode>3): Likewise.
6505 (smax<mode>3_vsx): Likewise.
6506 (smin<mode>3_vsx): Likewise.
6507 (negsf2): Delete SF operations that are merged with DF.
6514 (fmasf4_fpr): Likewise.
6515 (fmssf4_fpr): Likewise.
6516 (nfmasf4_fpr): Likewise.
6517 (nfmssf4_fpr): Likewise.
6518 (sqrtsf2): Likewise.
6519 (rsqrtsf_internal1): Likewise.
6520 (smaxsf3): Likewise.
6521 (sminsf3): Likewise.
6522 (cmpsf_internal1): Likewise.
6523 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
6524 (negdf2): Delete DF operations that are merged with SF.
6526 (nabsdf2): Likewise.
6532 (rsqrtdf_internal1): Likewise.
6533 (fmadf4_fpr): Likewise.
6534 (fmsdf4_fpr): Likewise.
6535 (nfmadf4_fpr): Likewise.
6536 (nfmsdf4_fpr): Likewise.
6537 (sqrtdf2): Likewise.
6538 (smaxdf3): Likewise.
6539 (smindf3): Likewise.
6540 (cmpdf_internal1): Likewise.
6541 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
6542 (btrunc<mode>2): Delete separate expander, and combine with the
6543 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
6544 (btrunc<mode>2_fpr): Likewise.
6545 (ceil<mode>2): Likewise.
6546 (ceil<mode>2_fpr): Likewise.
6547 (floor<mode>2): Likewise.
6548 (floor<mode>2_fpr): Likewise.
6549 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
6550 Add support for using the upper registers with VSX and
6551 power8-vector. Move insns to be closer to the define_expands. On
6552 VSX systems, prefer the traditional form of FMA over the VSX
6553 version, since the traditional form allows the target not to
6554 overlap with the inputs.
6555 (fms<mode>4_fpr): Likewise.
6556 (nfma<mode>4_fpr): Likewise.
6557 (nfms<mode>4_fpr): Likewise.
6559 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6560 Richard Earnshaw <richard.earnshaw@arm.com>
6562 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
6563 (struct mult_cost_table): Likewise.
6564 (struct mem_cost_table): Likewise.
6565 (struct fp_cost_table): Likewise.
6566 (struct vector_cost_table): Likewise.
6567 (cpu_cost_table): Likewise.
6568 * config/arm/arm.opt (mold-rts-costs): New option.
6569 (mnew-generic-costs): Likewise.
6570 * config/arm/arm.c (generic_extra_costs): New table.
6571 (cortexa15_extra_costs): Likewise.
6572 (arm_slowmul_tune): Use NULL as new costs.
6573 (arm_fastmul_tune): Likewise.
6574 (arm_strongarm_tune): Likewise.
6575 (arm_xscale_tune): Likewise.
6576 (arm_9e_tune): Likewise.
6577 (arm_v6t2_tune): Likewise.
6578 (arm_cortex_a5_tune): Likewise.
6579 (arm_cortex_a9_tune): Likewise.
6580 (arm_v6m_tune): Likewise.
6581 (arm_fa726te_tune): Likewise.
6582 (arm_cortex_a15_tune): Use cortex15_extra_costs.
6583 (arm_cortex_tune): Use generict_extra_costs.
6584 (shifter_op_p): New function.
6585 (arm_unspec_cost): Likewise.
6586 (LIBCALL_COST): Define.
6587 (arm_new_rtx_costs): New function.
6588 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
6589 table is available. Use old costs otherwise unless mnew-generic-costs
6591 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
6592 (cpu_cost_table): Declare.
6594 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
6597 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
6598 (*subs_mul_imm_<mode>)
6599 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
6600 (*sub_<shift>_<mode>)
6601 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
6602 Remove k constraint.
6604 2013-10-03 Ian Bolton <ian.bolton@arm.com>
6606 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
6608 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
6610 2013-10-02 Teresa Johnson <tejohnson@google.com>
6612 * predict.c (probably_never_executed): New function.
6613 (probably_never_executed_bb_p): Invoke probably_never_executed.
6614 (probably_never_executed_edge_p): Ditto.
6615 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
6616 Treat profile insanities conservatively.
6618 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
6620 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
6622 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
6624 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
6625 use it. Use smaller increase for scratch. Don't increase reject
6626 for early clobber scratch.
6627 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
6628 setting eliminated regs except setting fp from hfp.
6629 (lra_eliminate): Check lra_insn_recog_data on NULL.
6631 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6634 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
6635 setting -mvsx-timode by default until the underlying problem is fixed.
6636 (RS6000_CPU, power7 defaults): Likewise.
6638 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6640 * config/x-linux (host-linux.o): Remove header dependencies.
6641 Use $(COMPILE) and $(POSTCOMPILE).
6642 * config/t-linux-android (linux-android.o): Ditto.
6644 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6646 * Makefile.in (expmed.o-warn): Remove.
6648 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6650 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
6651 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
6653 2013-10-02 Teresa Johnson <tejohnson@google.com>
6655 * dojump.c (do_jump_1): Divide probability between
6656 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
6658 2013-10-02 Tom Tromey <tromey@redhat.com>
6660 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
6662 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6664 * tree-flow.h: Remove some prototypes.
6665 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
6666 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
6667 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
6668 mark_virtual_phi_result_for_renaming): Relocate here.
6669 * tree-into-ssa.h: Add prototypes.
6670 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
6671 single_pred_before_succ_order.
6672 (blocks_in_phiopt_order): Rename and move to cfganal.c.
6673 (nonfreeing_call_p) Move to gimple.c.
6674 * cfganal.c (single_pred_before_succ_order): Move and renamed from
6676 * basic-block.h (single_pred_before_succ_order): Add prototype.
6677 * gimple.c (nonfreeing_call_p): Relocate here.
6678 * gimple.h: Add prototype.
6679 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
6680 * tree-ssa-dom.h: New file. Relocate prototypes here.
6681 * tree-ssa.h: Include tree-ssa-dom.h.
6683 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6685 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
6686 Use $(COMPILE) and $(POSTCOMPILE).
6688 * config/alpha/x-alpha (driver-alpha.o): Ditto.
6690 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6692 * tree-flow.h: Remove some prototypes.
6693 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
6694 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6695 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
6696 propagate_tree_value*): Move from here to...
6697 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
6698 propagate_tree_value*): Relocate here.
6699 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6700 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
6701 * gimple-fold.c: Remove gimple-fold.h from include list.
6702 * tree-vrp.c: Remove gimple-fold.h from include list.
6703 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
6704 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
6705 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
6706 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
6707 * sese.c: Add tree-ssa-propagate.h to include list.
6709 2013-10-02 Richard Biener <rguenther@suse.de>
6711 * tree-loop-distribution.c: Include tree-vectorizer.h for
6713 (enum partition_kind): Remove PKIND_REDUCTION.
6714 (struct partition_s): Remove has_writes member, add reduction_p member.
6715 (partition_alloc): Adjust.
6716 (partition_builtin_p): Likewise.
6717 (partition_has_writes): Remove.
6718 (partition_reduction_p): New function.
6719 (partition_merge_into): Likewise.
6720 (generate_code_for_partition): Commonize builtin partition
6722 (rdg_cannot_recompute_vertex_p): Remove.
6723 (already_processed_vertex_p): Likewise.
6724 (rdg_flag_vertex): Do not set has_writes.
6725 (classify_partition): Adjust.
6726 (rdg_build_partitions): Do not set has_writes, treat all
6727 partitions as useful.
6728 (distribute_loop): Record number of library calls generated. Adjust.
6729 (tree_loop_distribution): Report number of loops and library
6730 calls generated as opt-info.
6732 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6734 * tree-flow.h: Include new .h files. Move prototypes.
6735 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
6736 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
6737 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
6738 * tree-pretty-print.h: Add prototypes from tree-flow.h.
6739 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
6740 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
6741 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
6742 * tree.h (get_ref_base_and_extent): Move prototype out.
6743 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
6745 * gimple-low.h: New File. Add prototypes from tree-flow.h.
6746 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
6747 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
6748 * tree-scalar-evolution.c: Include tree.h.
6749 * sese.c: Include tree.h.
6750 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
6751 * dwarf2out.c: Include tree-dfa.h.
6752 * tree-chrec.c: Include tree.h.
6753 * tree-data-ref.c: Include tree.h.
6755 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
6757 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
6760 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6762 * config/t-sol2 (sol2-c.o): Remove header dependencies.
6763 Use $(COMPILE) and $(POSTCOMPILE).
6764 (sol2-cxx.o): Likewise.
6765 (sol2-stubs.o): Likewise.
6767 * config/x-solaris (host-solaris.o): Likewise.
6769 * config/sparc/t-sparc (sparc.o): Remove.
6770 (sparc-c.o): Remove header dependencies.
6771 Use $(COMPILE) and $(POSTCOMPILE).
6772 * config/sparc/x-sparc: Likewise.
6774 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
6776 * config/arc/arc-opts.h: Add 2013 to Copyright years.
6777 * config/arc/arc700.md: Likewise.
6778 * config/arc/arc-modes.def: Likewise.
6779 * config/arc/arc-simd.h: Likewise.
6780 * config/arc/t-arc-uClibc: Likewise.
6781 * config/arc/t-arc-newlib: Likewise.
6783 2013-10-02 Renlin Li <renlin.li@arm.com>
6785 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6787 (aarch64_expand_epilogue): Likewise.
6789 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6790 Yufeng Zhang <yufeng.zhang@arm.com>
6792 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
6794 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
6795 'base_in' represent a conversion and legal_cast_p_1 holds; set
6796 'base_in' with the returned value from get_unwidened.
6798 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6800 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
6801 plus_constant is not used.
6803 2013-10-01 Wei Mi <wmi@google.com>
6805 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
6806 * config/i386/i386.md: Add define_peephole2 to
6807 break partial reg stall for cvtss2sd/cvtsd2ss.
6809 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6811 * config/arc/arc.c (pass_arc_ifcvt::clone):
6812 Update for ctxt_ -> m_ctxt change.
6814 2013-10-01 Jeff Law <law@redhat.com>
6816 * tree-ssa-threadupdate.c (struct redirection_data): Delete
6817 outgoing_edge and intermediate_edge fields. Instead store the path.
6818 (redirection_data::hash): Hash on the last edge's destination index.
6819 (redirection_data::equal): Check the entire thread path.
6820 (lookup_redirectio_data): Corresponding changes.
6821 (create_edge_and_update_destination_phis): Likewise.
6822 (thread_single_edge): Likewise.
6824 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6825 Diego Novillo <dnovillo@google.com>
6827 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
6828 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
6829 (vld32wh_insn, vld32wl_insn): Delete commented-out old
6830 versions of these patterns.
6832 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
6833 (__builtin_arc_aligned): Likewise.
6835 * config/arc/arc.md: Expand adc_0 comment stating the intended
6836 purpose and why it isn't ready.
6837 Replace commented out call_value_via_label_mixed with a
6838 plain comment about bl_s.
6840 * config/arc/arc.c (stdio.h): Don't include directly.
6841 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
6842 Remove if (1) condition.
6843 (arc_encode_section_info): Fix comment.
6845 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6847 * config/arc/arc.c (arc_conditional_register_usage):
6848 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
6849 Also set reg_alloc_order for DMA config regs.
6851 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6852 Jeremy Bennett <jeremy.bennett@embecosm.com>
6854 * doc/install.texi (--with-cpu): Mention ARC.
6855 (arc-*-elf32): New paragraph.
6856 (arc-linux-uclibc): Likewise.
6857 * doc/md.texi (Machine Constraints): Add ARC part.
6858 * doc/invoke.texi: (menu): Add ARC Options.
6859 (Machine Dependent Options) <ARC Options>: Add synopsis.
6860 (node ARC Options): Add.
6861 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6862 (ARC Built-in Functions): New section defining
6863 generic ARC built-in functions.
6864 (ARC SIMD Built-in Functions): New section defining SIMD specific
6866 (Declaring Attributes of Functions): Extended
6867 description of short_call and long_call attributes for ARC and
6868 added index entries.
6870 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6871 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6872 Joern Rennecke <joern.rennecke@embecosm.com>
6873 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6874 Brendan Kehoe <brendan@zen.org>
6875 Michael Eager <eager@eagercon.com>
6876 Simon Cook <simon.cook@embecosm.com>
6877 Jeremy Bennett <jeremy.bennett@embecosm.com>
6879 * config/arc, common/config/arc: New directories.
6881 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6882 Brendan Kehoe <brendan@zen.org>
6883 Simon Cook <simon.cook@embecosm.com>
6885 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6887 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6889 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6890 * tree-ssa-coalesce.h: New. Move prototype to here.
6891 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6892 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6893 (gimple_can_coalesce_p): Move to...
6894 * gimple.c (gimple_can_coalesce_p): Here.
6896 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6898 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6899 (dump_decl_set): Move to gimple.c.
6900 * gimple.h: Don't include tree-ssa-operands.h.
6901 (dump_decl_set): Add prototype.
6902 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6903 Move to gimple-ssa.h.
6904 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6905 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6906 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6907 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6908 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6910 (dump_decl_set): Relocate here.
6911 * gimple-ssa.h: New file.
6912 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6913 Relocate from gimple.h.
6914 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6915 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6916 swap_tree_operands and remove non-ssa path.
6917 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6918 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6920 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6921 vect_is_simple_reduction_1): Use swap_ssa_operands.
6922 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6923 (enum need_phi_state): Move to tree-into-ssa.c.
6924 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6925 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6926 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6927 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6928 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6929 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6930 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6931 num_imm_uses): Move to ssa-iterators.h.
6932 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6933 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6935 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6936 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6937 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6938 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6939 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6940 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6941 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6942 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6943 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6944 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6945 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6946 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6947 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6948 (set_phi_nodes): Move to tree-phinodes.h.
6949 * tree-ssa-operands.h (enum ssa_op_iter_type,
6950 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6951 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6952 (dump_decl_set): Remove prototype.
6953 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6954 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6955 (set_phi_nodes): Relocate from tree-flow-inline.h.
6956 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6958 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6959 include list. Temporarily add gimple.h to include list.
6960 * ssa-iterators.h: New file.
6961 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6962 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6963 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6964 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6965 Relocate from tree-ssa-operands.h.
6966 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6967 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6968 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6969 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6970 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6971 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6972 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6973 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6974 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6975 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6976 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6977 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6978 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6979 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6980 Relocate from tree-flow-inline.h.
6981 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6983 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6986 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6987 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6988 Insert '\t' to output template.
6989 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6990 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6991 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6992 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
6994 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
6996 * doc/install.texi (Host/target specific installation notes for GCC):
6997 Put @anchor before @heading.
6998 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
6999 Use @email for email addresses.
7001 2013-10-01 Jeff Law <law@redhat.com>
7003 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
7004 a vec. Only delete the path if we create one without successfully
7005 registering a jump thread.
7006 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
7008 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
7009 (paths): New vector of jump threading paths.
7010 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
7011 (THREAD_PATH): New accessor macro for the entire thread path.
7012 (lookup_redirection_data): Get intermediate and final outgoing edge
7013 from the thread path.
7014 (create_edge_and_update_destination_phis): Copy the threading path.
7015 (ssa_fix_duplicate_block_edges): Get edges and block types from the
7016 jump threading path.
7017 (ssa_redirect_edges): Get edges and block types from the jump threading
7018 path. Free the path vector.
7019 (thread_block): Get edges from the jump threading path. Look at the
7020 entire path to see if we thread to a loop exit. If we cancel a jump
7021 thread request, then free the path vector.
7022 (thread_single_edge): Get edges and block types from the jump threading
7023 path. Free the path vector.
7024 (thread_through_loop_header): Get edges and block types from the jump
7025 threading path. Free the path vector.
7026 (mark_threaded_blocks): Iterate over the vector of paths and store
7027 the path on the appropriate edge. Get edges and block types from the
7028 jump threading path.
7029 (mark_threaded_blocks): Get edges and block types from the jump
7030 threading path. Free the path vector.
7031 (thread_through_all_blocks): Use the vector of paths rather than
7032 a vector of 3-edge sets.
7033 (register_jump_thread): Accept pointer to a path vector rather
7034 than the path vector itself. Store the path vector for later use.
7037 2013-10-01 Jakub Jelinek <jakub@redhat.com>
7038 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7041 * config/s390/s390.c (s390_split_branches): Modify check for table
7043 (s390_chunkify_start): Rearrange table jump insn check in order to
7044 deal with compare and branch insns correctly.
7046 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
7050 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
7051 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
7052 define_insn_and_split.
7053 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
7054 (shiftsi3_compare): New pattern.
7056 * config/arm/unspecs.md (UNSPEC_RRX): New.
7058 2013-10-01 Alan Modra <amodra@gmail.com>
7060 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
7061 casing inout operands.
7063 2013-10-01 Richard Biener <rguenther@suse.de>
7065 PR tree-optimization/58553
7066 * tree-loop-distribution.c (struct partition_s): Add niter member.
7067 (classify_partition): Populate niter member for the partition
7068 and properly identify whether the relevant store happens before
7069 or after the loop exit.
7070 (generate_memset_builtin): Use niter member from the partition.
7071 (generate_memcpy_builtin): Likewise.
7073 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
7075 * vec.h (vec_prefix, vec): Prefix member names with "m_".
7076 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
7078 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
7080 * basic-block.h (edge_list): Prefix member names with "m_".
7081 * context.h (context): Likewise.
7082 * domwalk.h (dom_walker): Likewise.
7083 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
7084 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
7085 * hash-table.h (hash_table): Likewise.
7086 * machmode.h (bit_field_mode_iterator): Likewise.
7087 * pass_manager.h (pass_list): Likewise.
7088 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
7089 * tree-pass.h (pass_data): Likewise.
7090 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
7091 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
7092 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
7093 * asan.c (pass_data_asan): Update accordingly.
7094 * cfganal.c (control_dependences::find_control_dependence): Likewise.
7095 (control_dependences::control_dependences): Likewise.
7096 (control_dependences::~control_dependences): Likewise.
7097 (control_dependences::~control_dependences): Likewise.
7098 (control_dependences::get_edges_dependent_on): Likewise.
7099 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
7100 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
7101 * context.c (gcc::context::context): Likewise.
7102 * cprop.c (pass_rtl_cprop::clone): Likewise.
7103 * domwalk.c (dom_walker::walk): Likewise.
7104 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
7105 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
7106 * mode-switching.c (pass_mode_switching::clone): Likewise.
7107 * passes.c (opt_pass::opt_pass): Likewise.
7108 (pass_manager::pass_manager): Likewise.
7109 * predict.c (pass_strip_predict_hints::clone): Likewise.
7110 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
7111 (pass_split_all_insns::clone): Likewise.
7112 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
7114 (bit_field_mode_iterator::next_mode): Likewise.
7115 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
7116 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
7117 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
7118 * tree-complex.c (pass_lower_complex::clone): Likewise.
7119 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
7120 * tree-object-size.c (pass_object_sizes::clone): Likewise.
7121 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
7122 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
7123 (pass_fold_builtins::clone): Likewise.
7124 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
7125 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
7126 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
7127 (pass_cd_dce::clone): Likewise.
7128 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
7129 (pass_phi_only_cprop::clone): Likewise.
7130 * tree-ssa-dse.c (pass_dse::clone): Likewise.
7131 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
7132 * tree-ssa-loop.c (pass_lim::clone): Likewise.
7133 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
7134 * tree-ssa-pre.c (pass_fre::clone): Likewise.
7135 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
7136 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
7137 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
7138 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
7139 * tree-vrp.c (pass_vrp::clone): Likewise.
7140 * tsan.c (pass_tsan::clone): Likewise.
7142 2013-09-30 Jakub Jelinek <jakub@redhat.com>
7145 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
7146 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
7149 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
7150 optimization, punt if sign_bit_p looked through any zero extension.
7152 2013-09-30 Teresa Johnson <tejohnson@google.com>
7154 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
7155 Update redirected out edge count in joiner case.
7156 (ssa_redirect_edges): Common the joiner and non-joiner cases
7157 so that joiner case gets profile updates.
7159 2013-09-30 Richard Biener <rguenther@suse.de>
7161 PR tree-optimization/58554
7162 * tree-loop-distribution.c (classify_partition): Require
7163 unconditionally executed stores for memcpy and memset recognition.
7164 (tree_loop_distribution): Calculate dominance info.
7166 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
7168 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
7169 (NO_PROFILE_COUNTERS): Likewise.
7170 (PROFILE_HOOK): Likewise.
7171 (FUNCTION_PROFILER): Likewise.
7172 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
7174 2013-09-30 Iain Sandoe <iain@codesourcery.com>
7176 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
7177 calls and defines in TARGET_MACHO conditional.
7178 (load_macho_picbase_di): Likewise.
7179 (reload_macho_picbase): Likewise.
7180 (reload_macho_picbase_si): Likewise.
7181 (reload_macho_picbase_di): Likewise.
7182 (nonlocal_goto_receiver): Likewise.
7184 2013-09-30 Nick Clifton <nickc@redhat.com>
7186 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
7187 that use the MSP430X ISA.
7188 (msp430_option_override): Scan -mmcu command line option for any
7189 MCU name that supports the MSP430X ISA.
7190 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
7191 -mmcu options which enable the MSP430X ISA.
7193 2013-09-30 Richard Biener <rguenther@suse.de>
7196 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
7197 before looking for setjmp-like calls.
7199 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7202 * config/darwin-protos.h (machopic_get_function_picbase): New.
7203 * config/darwin.c (machopic_get_function_picbase): New.
7204 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
7205 label for a new func. (load_macho_picbase_di): Likewise.
7206 (reload_macho_picbase): New expand.
7207 (reload_macho_picbase_si): New insn.
7208 (reload_macho_picbase_di): New insn.
7209 (nonlocal_goto_receiver): New define and split.
7210 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
7211 (unspecv enum): Add UNSPECV_NLGR.
7213 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7215 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
7216 that altivec registers are correctly sized on Darwin.
7218 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7220 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
7221 darwin-driver.o): Use COMPILE and POSTCOMPILE.
7222 * config/x-darwin (host-darwin.o): Likewise.
7223 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
7224 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
7225 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
7227 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
7229 * doc/invoke.texi: Fix usage of @tie{} command.
7231 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
7233 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
7235 2013-09-28 Jan Hubicka <jh@suse.cz>
7237 * config/i386/x86-tune.def: Add documentation for each of the options;
7240 2013-09-28 Jan Hubicka <jh@suse.cz>
7242 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
7244 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
7245 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
7246 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
7247 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
7248 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
7250 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
7252 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
7253 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
7254 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
7255 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
7256 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
7257 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
7258 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
7259 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
7260 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
7261 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
7262 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
7263 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
7264 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
7265 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
7266 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
7267 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
7268 gimple.h, godump.c, graphite-clast-to-gimple.c,
7269 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
7270 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
7271 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
7272 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
7273 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
7274 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
7275 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
7276 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
7277 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
7278 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
7279 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
7280 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
7281 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
7282 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
7283 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
7284 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
7285 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
7286 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
7287 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
7288 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
7289 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
7290 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
7291 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
7292 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
7293 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
7294 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
7295 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
7296 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
7297 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
7298 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
7299 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
7300 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
7301 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
7302 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
7303 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
7304 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
7305 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
7306 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
7307 whitespace before "(".
7309 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
7311 * expr.h (extract_bit_field): Remove packedp parameter.
7312 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
7313 from forward declaration.
7314 (store_split_bit_field): Remove packedp arg from calls to
7315 extract_fixed_bit_field.
7316 (extract_bit_field_1): Remove packedp parameter and packedp
7317 argument from recursive calls and calls to extract_fixed_bit_field.
7318 (extract_bit_field): Remove packedp parameter and corresponding
7319 arg to extract_bit_field_1.
7320 (extract_fixed_bit_field): Remove packedp parameter. Remove code
7322 (extract_split_bit_field): Remove packedp arg from call to
7323 extract_fixed_bit_field.
7324 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
7325 (copy_blkmode_from_reg): Likewise.
7326 (copy_blkmode_to_reg): Likewise.
7327 (read_complex_part): Likewise.
7328 (store_field): Likewise.
7329 (expand_expr_real_1): Likewise.
7330 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
7331 to extract_bit_field.
7332 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
7333 call to extract_bit_field.
7334 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
7335 call to extract_bit_field.
7336 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
7337 and special packedp behavior from -fstrict-volatile-bitfields
7340 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7342 * lra-eliminations.c (init_elim_table): Guard value_p.
7344 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7346 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
7347 DFmode, DImode, and SFmode in the upper VSX registers based on the
7348 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
7349 if -mpower8-vector. Combine -mvsx-timode handling with the rest
7350 of the VSX register handling.
7352 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
7354 (zero_extendsidi2_lfiwzx): Add support for loading into the
7355 Altivec registers with -mpower8-vector. Use wu/wv constraints to
7356 only do VSX memory options on Altivec registers.
7357 (extendsidi2_lfiwax): Likewise.
7358 (extendsfdf2_fpr): Likewise.
7359 (mov<mode>_hardfloat, SF/SD modes): Likewise.
7360 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
7361 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
7362 (movdi_internal64): Likewise.
7364 2013-09-27 Xinliang David Li <davidxl@google.com>
7366 * opts.c (finish_options): Adjust parameters
7367 according to vect cost model.
7368 (common_handle_option): Set dynamic vect cost
7370 targhooks.c (default_add_stmt_cost): Compute stmt cost
7372 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
7373 Use helper function.
7374 * tree-vectorizer.h (unlimited_cost_model): New function.
7375 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
7376 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
7378 (vect_enhance_data_refs_alignment): Ditto.
7379 * flag-types.h: New enum.
7380 * common/config/i386/i386-common.c (ix86_option_init_struct):
7381 No need to initialize vect_cost_model flag.
7382 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
7385 2013-09-27 Diego Novillo <dnovillo@google.com>
7387 * gimple.h (enum ssa_mode): Remove.
7389 2013-09-27 Paulo Matos <pmatos@broadcom.com>
7391 * cfgloop.h (number_of_loops): Fix typo in check for null.
7393 2013-09-27 Jakub Jelinek <jakub@redhat.com>
7396 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
7397 are children of outermost saved_cfun's loop, and set it up to
7398 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
7399 if loop != loop0 and SESE region contains bbs that belong to loop0.
7401 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
7403 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
7404 (binary_scale_code_p, get_base_term, get_index_term): New functions.
7405 (set_address_segment, set_address_base, set_address_index)
7406 (set_address_disp): Accept the argument unconditionally.
7407 (baseness): Remove must_be_base_p and must_be_index_p checks.
7408 (decompose_normal_address): Classify as much as possible in the
7411 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
7413 * cse.c (count_reg_usage): Handle INT_LIST.
7414 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
7415 * reginfo.c (reg_scan_mark_refs): Likewise.
7416 * reload1.c (eliminate_regs_1): Likewise.
7418 2013-09-27 Iain Sandoe <iain@codesourcery.com>
7421 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
7424 2013-09-27 Richard Biener <rguenther@suse.de>
7426 PR tree-optimization/58459
7427 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
7428 restriction not propagating into loops.
7430 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
7432 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
7433 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
7434 * doc/tree-ssa.texi (Walking use-def chains): Delete.
7436 2013-09-26 Richard Biener <rguenther@suse.de>
7438 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
7440 2013-09-26 Richard Biener <rguenther@suse.de>
7442 * alias.h (component_uses_parent_alias_set): Rename to ...
7443 (component_uses_parent_alias_set_from): ... this.
7444 * alias.c (component_uses_parent_alias_set): Rename to ...
7445 (component_uses_parent_alias_set_from): ... this and return
7447 (reference_alias_ptr_type_1): Use the result from
7448 component_uses_parent_alias_set_from instead of stripping
7449 components one at a time.
7450 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
7452 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7454 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
7455 Move prototypes to...
7456 * tree-ssa-ter.h: New File. Move prototypes here.
7457 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
7458 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
7459 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
7460 from is_replaceable_p.
7461 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
7462 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
7463 (process_replaceable): Use ter_is_replaceable_p.
7464 (find_replaceable_in_bb): Use ter_is_replaceable_p.
7465 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
7466 newly refactored ssa_is_replaceable_p.
7467 * cfgexpand.c: Include tree-outof-ssa.h.
7468 * ssaexpand.h: Delete.
7470 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7472 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
7473 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
7475 (create_gimple_tmp): Delete.
7476 (get_expr_type, build_assign, build_type_cast): Move to...
7477 * gimple-builder.c: New File.
7478 (get_expr_type): Relocate from gimple.c.
7479 (build_assign, build_type_cast): Change to only create ssanames.
7480 * gimple.h: Move prototypes to...
7481 * gimple-builder.h: New File. Here.
7482 * tree-ssa.h: And here.
7483 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
7484 count_uses_and_derefs): Relocate from gimple.c.
7485 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
7486 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
7487 * tree-ssa-math-opts (execute_cse_reciprocals): Use
7488 gimple_replace_ssa_lhs.
7489 * asan.c: Include gimple-builder.h.
7490 * Makefile.in: Add gimple-builder.o.
7492 2013-09-26 Richard Biener <rguenther@suse.de>
7494 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
7495 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
7496 (loe_visit_block): Use gcc_checking_assert.
7497 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
7498 gimple_assign_ssa_name_copy_p.
7499 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
7502 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
7504 * config/rs6000/t-rs6000 (rs6000.o): Remove.
7505 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
7507 2013-09-26 Richard Biener <rguenther@suse.de>
7509 PR tree-optimization/58539
7510 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
7511 the fact that debug statements are not taking part in loop-closed
7514 2013-09-26 Nick Clifton <nickc@redhat.com>
7516 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
7517 time warning message.
7518 (msp430_print_operand_raw): Delete unused letter parameter.
7519 (TARGET_PRINT_OPERAND_ADDRESS): Define.
7520 (msp430_print_operand_address): New function.
7521 (msp430_print_operand): Move address printing code from here to
7523 * config/msp430/msp430.md (movsipsi2): Add comment in generated
7525 (zero_extendpsisi2): Likewise.
7526 (extendpsisi2): New pattern.
7527 (andneghi3): New pattern.
7529 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
7531 * config/aarch64/aarch64.opt (mlra): New option.
7532 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
7533 (TARGET_LRA_P): Define.
7535 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
7537 * expr.c (expand_assignment): Remove obsolete comment.
7539 2013-09-25 Jeff Law <law@redhat.com>
7541 * tree-flow.h (thread_through_all_blocks): Prototype moved into
7542 tree-ssa-threadupdate.h.
7543 (register_jump_thread): Similarly.
7544 * tree-ssa-threadupdate.h: New header file.
7545 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
7546 * tree-vrp.c: Likewise.
7547 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
7548 (thread_around_empty_blocks): Change type of path vector argument to
7549 an edge,type pair from just an edge. Initialize both elements when
7550 appending to a jump threading path. Tweak references to elements
7552 (thread_across_edge): Similarly. Release memory for the elements
7554 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
7555 (dump_jump_thread_path): New function broken out from
7556 register_jump_thread.
7557 (register_jump_thread): Use dump_jump_thread_path. Change type of
7558 path vector entries. Search the path for NULL edges and dump
7559 the path if one is found. Tweak the conversion of path to 3-edge
7560 form to use the block copy type information embedded in the path.
7562 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7564 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
7566 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7567 Vladimir Makarov <vmakarov@redhat.com>
7569 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
7570 from the least significant bit.
7571 (strip_address_mutations): Add bitfield operations handling.
7572 (must_be_index_p): Add shifting and rotate operations handling.
7573 (set_address_base): Use must_be_base_p predicate.
7574 (set_address_index): Use must_be_index_p predicate.
7576 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
7577 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7578 Sergey Lega <sergey.s.lega@intel.com>
7579 Anna Tikhonova <anna.tikhonova@intel.com>
7580 Ilya Tocar <ilya.tocar@intel.com>
7581 Andrey Turetskiy <andrey.turetskiy@intel.com>
7582 Ilya Verbin <ilya.verbin@intel.com>
7583 Kirill Yukhin <kirill.yukhin@intel.com>
7584 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7586 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
7588 (ix86_expand_vector_move_misalign): Support new unaligned load and
7589 stores and use new names.
7590 (CODE_FOR_sse2_storedqu): Rename to ...
7591 (CODE_FOR_sse2_storedquv16qi): ... this.
7592 (CODE_FOR_sse2_loaddqu): Rename to ...
7593 (CODE_FOR_sse2_loaddquv16qi): ... this.
7594 (CODE_FOR_avx_loaddqu256): Rename to ...
7595 (CODE_FOR_avx_loaddquv32qi): ... this.
7596 (CODE_FOR_avx_storedqu256): Rename to ...
7597 (CODE_FOR_avx_storedquv32qi): ... this.
7598 * config/i386/i386.md (fpint_logic): New.
7599 * config/i386/sse.md (VMOVE): Extend for AVX512.
7603 (VI_UNALIGNED_LOADSTORE): Ditto.
7604 (sse2_avx_avx512f): Ditto.
7605 (sse2_avx2): Extend for AVX512.
7606 (sse4_1_avx2): Ditto.
7607 (avx2_avx512f): New.
7608 (sse): Extend for AVX512.
7611 (avxsizesuffix): Ditto.
7612 (sseintvecmode): Ditto.
7614 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
7615 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
7616 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7617 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
7618 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7619 (<sse2_avx_avx512f>_storedqu<mode): ... this.
7620 (<sse>_movnt<mode>): Replace constraint "x" with "v".
7621 (STORENT_MODE): Extend for AVX512.
7622 (*absneg<mode>2): Replace constraint "x" with "v".
7623 (*mul<mode>3): Ditto.
7624 (*ieee_smin<mode>3): Ditto.
7625 (*ieee_smax<mode>3): Ditto.
7626 (avx_cmp<mode>3): Replace VF with VF_128_256.
7627 (*<sse>_maskcmp<mode>3_comm): Ditto.
7628 (<sse>_maskcmp<mode>3): Ditto.
7629 (<sse>_andnot<mode>3): Extend for AVX512.
7630 (<code><mode>3, anylogic): Replace VF with VF_128_256.
7631 (<code><mode>3, fpint_logic): New.
7632 (*<code><mode>3): Extend for AVX512.
7633 (avx512flogicsuff): New.
7634 (avx512f_<logic><mode>): Ditto.
7635 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
7637 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
7638 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
7639 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
7640 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
7641 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
7642 (xop_vpermil2<mode>3): Ditto.
7643 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
7644 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
7645 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
7646 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
7648 2013-09-25 Tom Tromey <tromey@redhat.com>
7650 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
7651 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
7652 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
7653 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
7654 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
7655 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
7656 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
7657 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
7658 (GRAPHITE_HTAB_H): Remove.
7660 2013-09-25 Tom Tromey <tromey@redhat.com>
7662 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
7664 2013-09-25 Tom Tromey <tromey@redhat.com>
7666 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
7668 2013-09-25 Tom Tromey <tromey@redhat.com>
7670 * config/i386/t-i386 (i386.o): Remove.
7671 (i386-c.o): Use COMPILE and POSTCOMPILE.
7673 2013-09-25 Tom Tromey <tromey@redhat.com>
7675 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
7677 2013-09-25 Tom Tromey <tromey@redhat.com>
7679 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
7680 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
7681 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
7682 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
7683 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
7684 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
7685 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
7686 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
7687 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
7688 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
7689 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
7690 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
7691 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
7692 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
7693 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
7694 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
7695 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
7696 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
7697 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
7698 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
7699 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
7700 (tree-ssa-pre.o, tree-ssa-sccvn.o)
7701 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
7702 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
7703 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
7704 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
7705 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
7706 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
7707 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
7708 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
7709 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
7710 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
7711 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
7712 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
7713 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
7714 (graphite-clast-to-gimple.o, graphite-dependences.o)
7715 (graphite-interchange.o, graphite-poly.o)
7716 (graphite-scop-detection.o, graphite-sese-to-poly.o)
7717 (graphite-optimize-isl.o, tree-vect-loop.o)
7718 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
7719 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
7720 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
7721 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
7722 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
7723 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
7724 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
7725 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
7726 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
7727 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
7728 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
7729 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
7730 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
7731 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
7732 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
7733 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
7734 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
7735 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
7736 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
7737 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
7738 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
7739 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
7740 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
7741 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
7742 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
7743 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
7744 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
7745 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
7746 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
7747 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
7748 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
7749 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
7750 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
7751 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
7752 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
7753 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
7754 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
7755 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
7756 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
7757 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
7758 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
7759 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
7760 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
7761 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
7762 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
7763 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
7764 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
7765 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
7766 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
7767 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
7768 (gcov-dump.o): Remove.
7769 (default-c.o): Use COMPILE and POSTCOMPILE.
7770 (CFLAGS-gcc.o): New variable.
7771 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
7773 2013-09-25 Tom Tromey <tromey@redhat.com>
7775 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
7776 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
7777 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
7778 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
7779 (c-family/c-pragma.o, c-family/c-pretty-print.o)
7780 (c-family/c-semantics.o, c-family/c-ada-spec.o)
7781 (c-family/array-notation-common.o, c-family/stub-objc.o)
7782 (c-family/c-ubsan.o): Remove.
7784 2013-09-25 Tom Tromey <tromey@redhat.com>
7786 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
7788 2013-09-25 Tom Tromey <tromey@redhat.com>
7790 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
7791 to add -DENABLE_SHARED_LIBGCC.
7792 (gcc.o): Don't use subshell.
7794 2013-09-25 Tom Tromey <tromey@redhat.com>
7796 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
7797 * configure.ac: Don't invoke AM_PROG_CC_C_O.
7798 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
7799 * configure, config.in: Rebuild.
7801 2013-09-25 Tom Tromey <tromey@redhat.com>
7803 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
7804 (COMPILE, POSTCOMPILE): New variables.
7805 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
7806 (DEPFILES): New variable.
7807 Include ".Po" files.
7808 * configure.ac: Add checks for dependency checking.
7809 * configure, aclocal.m4: Regenerate.
7811 2013-09-25 Tom Tromey <tromey@redhat.com>
7813 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7814 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
7816 2013-09-25 Tom Tromey <tromey@redhat.com>
7818 * Makefile.in (generated_files): Add options.h,
7819 target-hooks-def.h, insn-opinit.h,
7820 common/common-target-hooks-def.h, pass-instances.def,
7821 c-family/c-target-hooks-def.h.
7823 2013-09-25 Jeff Law <law@redhat.com>
7825 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
7826 than foo[foo.length () - 1] to access last member in a vec.
7827 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
7829 2013-09-25 Richard Biener <rguenther@suse.de>
7832 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
7834 2013-09-25 Jan Hubicka <jh@suse.cz>
7836 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
7839 2013-09-25 Marek Polacek <polacek@redhat.com>
7842 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
7843 when determining the type name.
7845 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
7847 * config/sh/sh.md: Fix formatting.
7849 2013-09-24 Xinliang David Li <davidxl@google.com>
7851 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7852 max peel iterations parameter.
7853 * param.def: New parameter.
7854 * doc/invoke.texi: Document New parameter.
7856 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7858 * gimple-pretty-print.c: Various whitespace tweaks.
7859 * tree-core.h: Likewise.
7860 * tree-pretty-print.c: Likewise.
7861 * tree-ssa-alias.c: Likewise.
7862 * tree-ssa-copy.c: Likewise.
7863 * tree-ssanames.c: Likewise.
7864 * tree-ssanames.h: Likewise.
7865 * tree-vrp.c: Likewise.
7867 2013-09-24 Alan Modra <amodra@gmail.com>
7871 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7872 for output operands that disallow regs. Don't use EXPAND_WRITE on
7875 2013-09-24 Richard Biener <rguenther@suse.de>
7878 * tree.c (reference_alias_ptr_type): Move ...
7879 * alias.c (reference_alias_ptr_type): ... here and implement
7880 in terms of the new reference_alias_ptr_type_1.
7881 (ref_all_alias_ptr_type_p): New helper.
7882 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7883 use ref_all_alias_ptr_type_p.
7884 (get_deref_alias_set): Add flag_strict_aliasing check here.
7885 (reference_alias_ptr_type_1): New function, split out from ...
7886 (get_alias_set): ... here.
7887 (alias_ptr_types_compatible_p): New function.
7888 * alias.h (reference_alias_ptr_type): Declare.
7889 (alias_ptr_types_compatible_p): Likewise.
7890 * tree.h (reference_alias_ptr_type): Remove.
7891 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7892 to compare MEM_REF alias types.
7894 2013-09-24 Richard Biener <rguenther@suse.de>
7896 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7898 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7900 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7901 reload helper function arrays into a single array reg_addr.
7902 (reload_fpr_gpr): Likewise.
7903 (reload_gpr_vsx): Likewise.
7904 (reload_vsx_gpr): Likewise.
7905 (struct rs6000_reg_addr): Likewise.
7906 (reg_addr): Likewise.
7907 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7908 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7909 (rs6000_init_hard_regno_mode_ok): Likewise.
7910 (rs6000_secondary_reload_direct_move): Likewise.
7911 (rs6000_secondary_reload): Likewise.
7913 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7914 constraints: wu, ww, and wy. Repurpose wv constraint added during
7915 power8 changes. Put wg constraint in alphabetical order.
7917 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7918 for future work to add ISA 2.07 VSX single precision support.
7919 (-mvsx-scalar-double): Change default from -1 to 1, update
7920 documentation comment.
7921 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7922 (-mupper-regs-df): New debug switch to control whether DF values
7923 can go in the traditional Altivec registers.
7924 (-mupper-regs-sf): New debug switch to control whether SF values
7925 can go in the traditional Altivec registers.
7927 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7929 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7930 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7931 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7932 (rs6000_debug_legitimate_address_p): Print if we are running
7933 before, during, or after reload.
7934 (rs6000_secondary_reload): Add a comment.
7935 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7937 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7938 constraints. Update documentation string.
7939 (wd constraint): Likewise.
7940 (wf constraint): Likewise.
7941 (wg constraint): Likewise.
7942 (wn constraint): Likewise.
7943 (ws constraint): Likewise.
7944 (wt constraint): Likewise.
7945 (wx constraint): Likewise.
7946 (wz constraint): Likewise.
7947 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7949 (ww constraint): Likewise.
7950 (wy constraint): Likewise.
7951 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7952 in the previous submissions.
7953 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7955 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7957 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7958 (REG_BR_PROB): Say that the probability is stored in an int_list.
7959 * reg-notes.def: Update commentary to mention INT_LIST.
7960 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7961 (INT_LIST): New rtx.
7962 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7963 * rtlanal.c (int_reg_note_p): New function.
7964 (alloc_reg_note): Assert that the note does not have an int argument.
7965 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7966 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7967 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7968 rather than an INSN_LIST. Handle INT_LIST.
7969 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7970 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7971 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7972 Manipulate them as ints rather than rtxes.
7973 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7974 * regmove.c (copy_src_to_dest): Likewise.
7975 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7977 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7978 into the cases that need it.
7979 * config/arm/arm.c (arm_unwind_emit): Likewise.
7981 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7982 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7983 * loop-doloop.c (add_test, doloop_modify): Likewise.
7984 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7985 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7986 * predict.c (combine_predictions_for_insn): Likewise.
7987 * print-rtl.c (print_rtx): Handle INT_LIST.
7988 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7989 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7990 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7991 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7992 (ix86_split_fp_branch, predict_jump): Likewise.
7993 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
7994 * config/sh/sh.c (expand_cbranchsi4): Likewise.
7995 * config/spu/spu.c (ea_load_store_inline): Likewise.
7997 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
7998 value of a REG_BR_PROB note.
7999 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
8000 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
8001 * emit-rtl.c (try_split): Likewise.
8002 * predict.c (br_prob_note_reliable_p): Likewise.
8003 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
8004 * reorg.c (mostly_true_jump): Likewise.
8005 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
8006 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
8007 * config/i386/i386.c (ix86_print_operand): Likewise.
8008 * config/ia64/ia64.c (ia64_print_operand): Likewise.
8009 * config/mmix/mmix.c (mmix_print_operand): Likewise.
8010 * config/rs6000/rs6000.c (output_cbranch): Likewise.
8011 * config/s390/s390.c (s390_expand_tbegin): Likewise.
8012 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
8013 * config/sparc/sparc.c (output_cbranch): Likewise.
8014 * config/spu/spu.c (get_branch_target): Likewise.
8015 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
8016 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
8018 2013-09-23 Jan Hubicka <jh@suse.cz>
8020 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
8022 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
8023 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
8025 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
8026 (gimple_fold_call): Dump inconsistent devirtualizations; add
8027 sanity check for type based devirtualizations.
8028 * ipa-prop.c: Include ipa-utils.h
8029 (ipa_intraprocedural_devirtualization): Add sanity check.
8030 (try_make_edge_direct_virtual_call): Likewise.
8032 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
8034 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
8035 assignment statements.
8037 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
8039 * gimple-pretty-print.c (dump_ssaname_info): New function.
8040 (dump_gimple_phi): Call it.
8041 (pp_gimple_stmt_1): Likewise.
8042 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
8043 (range_info_def): Declare.
8044 * tree-pretty-print.c (pp_double_int): New function.
8045 (dump_generic_node): Call it.
8046 * tree-pretty-print.h (pp_double_int): Declare.
8047 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
8048 * tree-ssanames.h (range_info_def): New structure.
8049 (value_range_type): Move definition here.
8050 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
8052 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
8054 (set_range_info): New function.
8055 (get_range_info): Likewise.
8056 (duplicate_ssa_name_range_info): Likewise.
8057 (duplicate_ssa_name_fn): Check pointer type and call
8058 duplicate_ssa_name_range_info.
8059 * tree-ssa-copy.c (fini_copy_prop): Likewise.
8060 * tree-vrp.c (value_range_type): Remove definition, now in
8062 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
8063 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
8064 (SSA_NAME_RANGE_INFO): New macro.
8066 2013-09-23 Richard Biener <rguenther@suse.de>
8068 PR tree-optimization/58464
8069 * tree-ssa-pre.c (phi_trans_lookup): Remove.
8070 (phi_trans_add): Change to add conditionally on being not present.
8071 (phi_translate_1): Remove recursion detection here.
8072 (phi_translate): Pre-seed the cache with NULL to catch
8073 recursion here in a more generic way.
8074 (bitmap_find_leader): Adjust comment.
8075 (get_representative_for): Dump value-numbers.
8076 (create_expression_by_pieces): Likewise.
8077 (insert_into_preds_of_block): Likewise.
8079 2013-09-23 Christian Bruel <christian.bruel@st.com>
8082 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
8083 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
8085 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
8088 2013-09-20 Renlin Li <renlin.li@arm.com>
8090 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
8092 (aarch64_expand_epilogue): Likewise.
8093 (aarch64_legitimize_reload_address): Likewise.
8095 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
8097 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
8098 NULL_TREE before pushing them onto the vector. Likewise for labels.
8100 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
8102 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
8103 comparison operators when -fno-trapping-math is in effect.
8104 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
8105 comparison operators in TFmode and assert that unsupported operators
8107 (ia64_print_operand): Likewise.
8109 2013-09-21 Jan Hubicka <jh@suse.cz>
8111 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
8112 (sse_typeless_stores): Enable for core
8113 (sse_load0_by_pxor): Likewise.
8114 (four_jump_limit): Disable for core.
8115 (pad_returns): Likewise.
8116 (avoid_vector_decode): Likewise.
8117 (fuse_cmp_and_branch): Enable for cores.
8118 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
8120 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
8123 * config/pa/pa.c (pa_option_override): Disable auto increment and
8124 decrement instructions until reload is completed.
8126 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
8127 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
8128 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
8130 2013-09-20 DJ Delorie <dj@redhat.com>
8131 Nick Clifton <nickc@redhat.com>
8133 * config/rl78/rl78.c: Various whitespace and comment tweaks.
8134 (need_to_save): Save bank 0 on interrupts.
8135 (characterize_address): Strip far address wrappers.
8136 (rl78_as_legitimate_address): Likewise.
8137 (transcode_memory_rtx): Likewise.
8138 (rl78_peep_movhi_p): Disable this peephole after devirt.
8139 (rl78_propogate_register_origins): Forget all origins when a
8141 * config/rl78/rl78-virt.md: Various whitespace tweaks.
8142 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
8144 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
8145 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
8146 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
8148 * config/rl78/constraints.md: Various whitespace and paren tweaks.
8150 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
8152 * config/pa/pa.md: In "scc" insn patterns, change output template to
8153 handle const0_rtx in reg_or_0_operand operands.
8155 2013-09-20 Martin Husemann <martin@NetBSD.org>
8158 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
8159 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
8161 2013-09-20 Richard Biener <rguenther@suse.de>
8164 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
8165 remembering SSA name version and block index.
8166 (new_scev_info_str): Adjust.
8167 (hash_scev_info): Likewise. Also hash the block index.
8168 (eq_scev_info): Adjust.
8169 (find_var_scev_info): Likewise.
8170 (struct instantiate_cache_entry): Remove.
8171 (struct instantiate_cache_type): Use a htab to map name, block
8173 (instantiate_cache_type::~instantiate_cache_type): Adjust.
8174 (get_instantiated_value_entry): Likewise.
8175 (hash_idx_scev_info, eq_idx_scev_info): New functions.
8176 (instantiate_scev_name): Adjust.
8178 2013-09-20 Jeff Law <law@redhat.com>
8180 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
8182 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
8184 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
8185 Call aarch64_simd_expand_args to update op[argc].
8187 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
8189 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
8192 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
8194 * gengtype.c (file_rules): Added rule for *.cc files.
8195 (get_output_file_with_visibility): Give fatal message when no
8198 2013-09-20 Renlin Li <renlin.li@arm.com>
8200 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
8201 (aarch64_expand_epilogue): Likewise.
8202 (aarch64_legitimize_reload_address): Likewise.
8204 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
8207 * expr.c (expand_assignment): Remove misalignp code path.
8209 2013-09-20 Marek Polacek <polacek@redhat.com>
8212 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
8213 TYPE_PRECISION. Add asserts.
8215 2013-09-20 Richard Biener <rguenther@suse.de>
8217 PR tree-optimization/58453
8218 * tree-loop-distribution.c (distribute_loop): Apply the cost
8219 model for -ftree-loop-distribute-patterns, too.
8221 2013-09-20 Richard Biener <rguenther@suse.de>
8224 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
8225 make type comparison less strict.
8227 2013-09-20 Alan Modra <amodra@gmail.com>
8229 * configure: Regenerate.
8230 * aclocal.m4: Regenerate.
8232 2013-09-20 Marek Polacek <polacek@redhat.com>
8235 * doc/extend.texi: Document that attribute used is meant to be used
8236 on variables with static storage duration.
8238 2013-09-19 Jakub Jelinek <jakub@redhat.com>
8240 PR tree-optimization/58472
8241 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
8242 simd_lane_access set inv_p = false.
8243 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
8244 the simduid magic VAR_DECL.
8246 2013-09-19 Jan Hubicka <jh@suse.cz>
8248 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
8250 2013-09-17 Jeff Law <law@redhat.com>
8252 * tree-ssa-dom.c (record_temporary_equivalences): New function
8253 split out of dom_opt_dom_walker::after_dom_children.
8254 (dom_opt_dom_walker::thread_across_edge): Move common code
8255 in here from dom_opt_dom_walker::after_dom_children.
8256 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
8258 2013-09-19 Jan Hubicka <jh@suse.cz>
8260 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
8261 (TARGET_GENERIC): Use PROCESOR_GENERIC
8262 (enum processor_type): Unify generic32 and 64.
8263 * i386.md (cpu): Likewise.
8264 * x86-tune.def (use_leave): Enable for generic32.
8265 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
8266 * athlon.md: Change generic64 to generic in all occurences.
8267 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
8268 (ix86_target_macros_internal): Likewise.
8269 * driver-i386.c (host_detect_local_cpu): Likewise.
8270 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
8272 (generic_memcpy, generic_memset, generic_cost): This one.
8273 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
8274 (m_GENERIC32, m_GENERIC64): Remove.
8275 (m_GENERIC): Turn into one flag.
8276 (processor_target): Unify generic tunnings.
8277 (ix86_option_override_internal): Replace generic32/64 by generic.
8278 (ix86_issue_rate): Likewise.
8279 (ix86_adjust_cost): Likewise.
8281 2013-09-19 Jan Hubicka <jh@suse.cz>
8283 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
8284 of speculative flag.
8286 2013-09-19 Jakub Jelinek <jakub@redhat.com>
8288 * omp-low.c (expand_omp_sections): Always pass len - 1 to
8289 GOMP_sections_start, even if !exit_reachable.
8291 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
8293 * lra-constraints.c (need_for_all_save_p): Use macro
8294 HARD_REGNO_CALL_PART_CLOBBERED.
8295 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
8296 set up pseudo conflict hard regs.
8298 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8301 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
8304 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
8306 PR rtl-optimization/58438
8307 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
8308 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
8309 from previous subpasses.
8311 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
8313 * arm.c (arm_get_frame_offsets): Validate architecture supports
8314 LDRD/STRD before accepting the tuning preference.
8315 (arm_expand_prologue): Likewise.
8316 (arm_expand_epilogue): Likewise.
8318 2013-09-18 Richard Biener <rguenther@suse.de>
8320 PR tree-optimization/58417
8321 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
8322 have chrecs with symbols defined in the loop as operands.
8323 (chrec_fold_multiply): Likewise.
8324 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
8325 parameters before folding binary operations.
8326 (struct instantiate_cache_entry_hasher): Remove.
8327 (struct instantiate_cache_type): Use a pointer-map.
8328 (instantiate_cache_type::instantiate_cache_type): New function.
8329 (instantiate_cache_type::get): Likewise.
8330 (instantiate_cache_type::set): Likewise.
8331 (instantiate_cache_type::~instantiate_cache_type): Adjust.
8332 (get_instantiated_value_entry): Likewise.
8333 (global_cache): New global.
8334 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
8335 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
8336 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
8337 (instantiate_scev_name): Adjust.
8338 (instantiate_scev): Construct global instead of local cache.
8339 (resolve_mixers): Likewise.
8341 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
8342 Paolo Carlini <paolo.carlini@oracle.com>
8345 * doc/implement-cxx.texi: Fix references to the C++ standards.
8347 2013-09-18 Jakub Jelinek <jakub@redhat.com>
8349 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
8350 * tree-vect-data-refs.c (vect_analyze_data_refs): For
8351 simd_lane_access drs, update also DR_ALIGNED_TO.
8353 2013-09-18 Marek Polacek <polacek@redhat.com>
8356 * doc/extend.texi: Document no_sanitize_undefined attribute.
8357 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
8358 no_sanitize_undefined attribute.
8360 2013-09-18 Nick Clifton <nickc@redhat.com>
8362 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
8363 (ASM_DECLARE_FUNCTION_NAME): Define.
8365 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
8367 * compare-elim.c (find_comparison_dom_walker): New class
8368 (find_comparisons_in_bb): Rename to
8369 find_comparison_dom_walker::before_dom_children
8370 (find_comparisons): Adjust
8371 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
8373 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
8374 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
8375 (init_walk_dominator_tree): Remove declaration.
8376 (fini_walk_dominator_tree): Remove declaration.
8377 * fwprop.c (single_def_use_dom_walker): New class
8378 (single_def_use_enter_block): Convert to
8379 single_def_use_dom_walker::before_dom_children.
8380 (single_def_use_leave_block): Convert to
8381 single_def_use_dom_walker::after_dom_children.
8382 (build_single_def_use_links): Adjust.
8383 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
8385 (find_candidates_in_block): Convert to
8386 find_candidates_dom_walker::before_dom_children.
8387 (execute_strength_reduction): Adjust.
8388 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
8389 (sese_dom_walker): New class.
8390 (sese_dom_walker::sese_dom_walker): New constructor.
8391 (sese_dom_walker::~sese_dom_walker): New destructor.
8392 (build_sese_conditions_before): Convert to
8393 sese_dom_walker::before_dom_children.
8394 (build_sese_conditions_after): Convert to
8395 sese_dom_walker::after_dom_children.
8396 (build_poly_scop): Adjust
8397 * tree-into-ssa.c (rewrite_dom_walker): New class
8398 (rewrite_enter_block): Convert to
8399 rewrite_dom_walker::before_dom_children.
8400 (rewrite_leave_block): Convert to
8401 rewrite_dom_walker::after_dom_children.
8402 (rewrite_update_dom_walker): New class.
8403 (rewrite_update_enter_block): Convert to
8404 rewrite_update_dom_walker::before_dom_children.
8405 (rewrite_update_leave_block): Convert to
8406 rewrite_update_dom_walker::after_dom_children.
8407 (rewrite_blocks, rewrite_into_ssa): Adjust.
8408 (mark_def_dom_walker): New class.
8409 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
8410 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
8411 (mark_def_sites_blocks): Convert to
8412 mark_def_dom_walker::before_dom_children.
8413 (mark_def_site_blocks): Remove.
8414 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
8415 (tree_ssa_dominator_optimize): Adjust.
8416 (dom_thread_across_edge): Convert to method
8417 dom_opt_dom_walker::thread_across_edge.
8418 (dom_opt_enter_block): Convert to member function
8419 dom_opt_dom_walker::before_dom_children.
8420 (dom_opt_leave_block): Convert to member function
8421 dom_opt_dom_walker::after_dom_children.
8422 * tree-ssa-dse.c (dse_dom_walker): New class.
8423 (dse_enter_block): Convert to member function
8424 dse_dom_walker::before_dom_children.
8425 (tree_ssa_dse): Adjust.
8426 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
8427 (determine_invariantness_stmt): Convert to method
8428 invariantness_dom_walker::before_dom_children.
8429 (determine_invariantness): Remove
8430 (move_computations_dom_walker): New class.
8431 (move_computations_stmt): Convert to method
8432 move_computations_dom_walker::before_dom_children.
8433 (move_computations, tree_ssa_lim): Adjust.
8434 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
8435 (nt_init_block): Convert to method
8436 notrappping_dom_walker::before_dom_children.
8437 (nt_fini_block): Convert to method
8438 method nontrapping_dom_walker::after_dom_children.
8439 (get_non_trapping): Adjust.
8440 * tree-ssa-pre.c (eliminate_dom_walker): New class.
8441 (eliminate_bb): Convert to method
8442 eliminate_dom_walker::before_dom_children.
8443 (eliminate_leave_block): Convert to method
8444 eliminate_dom_walker::after_dom_children.
8445 (eliminate): Adjust.
8446 * tree-ssa-strlen.c (strlen_dom_walker): New class.
8447 (strlen_enter_block): Convert to method
8448 strlen_dom_walker::before_dom_children.
8449 (strlen_leave_block): Convert to method
8450 method strlen_dom_walker::after_dom_children.
8451 (tree_ssa_strlen): Adjust.
8452 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
8453 (tree_ssa_uncprop): Adjust.
8454 (uncprop_leave_block): Convert to method
8455 uncprop_dom_walker::after_dom_children.
8456 (uncprop_leave_block): Convert to method
8457 uncprop_dom_walker::before_dom_children.
8459 2013-09-18 Bin Cheng <bin.cheng@arm.com>
8461 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
8462 branch in same basic block. Check both src and dest of the move insn.
8464 2013-09-17 Nick Clifton <nickc@redhat.com>
8466 * config/rl78/rl78-real.md (bf): New pattern.
8468 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
8469 (rl78_print_operand): Do not put a # before a %B.
8470 * config/rl78/rl78.opt: Tweak doc strings.
8472 2013-09-17 DJ Delorie <dj@redhat.com>
8474 * config/rl78/constraints.md (Wcv): Allow up to $r31.
8475 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
8476 (rl78_option_override): Likewise, if -mallregs.
8477 (is_virtual_register): Likewise.
8478 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
8479 (REGNO_OK_FOR_BASE_P): Likewise.
8480 * config/rl78/rl78.opt (-mallregs): New.
8482 2013-09-17 Nick Clifton <nickc@redhat.com>
8484 * config/rl78/rl78.c (need_to_save): Change return type to bool.
8485 For interrupt functions: save all call clobbered registers if the
8486 interrupt handler is not a leaf function.
8487 (rl78_expand_prologue): Always recompute the frame information.
8488 For interrupt functions: only select bank 0 if one of the bank 0
8489 registers is going to be psuhed.
8491 2013-09-17 DJ Delorie <dj@redhat.com>
8493 * config/rl78/constraints.md: For each W* constraint, rename to C*
8494 and create a W* constraint that checks for an optional ES: prefix
8496 * config/rl78/rl78.md (UNS_ES_ADDR): New.
8497 (es_addr): New. Used to wrap far addresses.
8498 * config/rl78/rl78-protos.h (rl78_es_addr): New.
8499 (rl78_es_base): New.
8500 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
8501 wrapped far addresses.
8502 (rl78_print_operand_1): Unwrap far addresses before processing.
8503 (rl78_lo16): Wrap far addresses in unspecs.
8504 (rl78_es_addr): New.
8505 (rl78_es_base): New.
8506 (insn_ok_now): Check for not-yet-wrapped far addresses.
8507 (transcode_memory_rtx): Properly re-wrap far addresses.
8509 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
8511 * config/sparc/t-rtems: Add leon3 multilibs.
8513 2013-09-17 Cong Hou <congh@google.com>
8515 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
8516 when checking the dot production pattern. The type of rhs operand
8517 of multiply is now checked correctly.
8519 2013-09-17 Jeff Law <law@redhat.com>
8521 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8522 edge implied equivalences into successor phis.
8523 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
8524 here from tree-ssa-threadedge.c.
8525 (mark_threaded_blocks): When threading through a joiner, if both
8526 successors of the joiner's clone reach the same block, verify the
8527 PHI arguments are equal. If not, cancel the jump threading request.
8528 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
8529 tree-ssa-threadupdate.c
8530 (thread_across_edge): Don't check PHI argument equality when
8531 threading through joiner block here.
8533 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
8535 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
8536 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
8537 (warn_uninit, warn_uninitialized_vars,
8538 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8539 Move to tree-ssa-uninit.c.
8540 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
8541 (has_undefined_value_p): New. Pass dependant parts of
8542 ssa_undefined_value_p.
8543 (uninit_undefined_value_p): Use has_undefined_value_p.
8544 (warn_uninit, warn_uninitialized_vars,
8545 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8546 Move from tree-ssa.c.
8547 * tree-ssa.h: Adjust prototypes.
8549 2013-09-17 Jan Hubicka <jh@suse.cz>
8552 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
8553 * ipa-inline.c (can_inline_edge_p): Do not downgrade
8554 FUNCTION_NOT_OPTIMIZED.
8555 * ipa-inline-analysis.c (compute_inline_parameters): Function
8556 not optimized is not inlinable unless it is alwaysinline.
8557 (inline_analyze_function): Force calls in not optimized
8558 function not inlinable.
8560 2013-09-17 Jan Hubicka <jh@suse.cz>
8563 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
8565 * ipa.c (function_and_variable_visibility): Likewise.
8566 * ipa-profile.c (ipa_profile): Likewise.
8568 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
8571 * cgraph.c (cgraph_function_body_availability): Check for ifunc
8572 attribute, and don't inline the resolver in this case.
8574 2013-09-17 Teresa Johnson <tejohnson@google.com>
8576 * coverage.c (get_coverage_counts): Add missing newline.
8578 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8580 PR tree-optimization/58088
8581 * fold-const.c (mask_with_trailing_zeros): New function.
8582 (fold_binary_loc): Make sure we don't recurse infinitely
8583 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
8584 Use mask_with_trailing_zeros where appropriate.
8586 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
8588 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
8589 of 'prev' var to get better distance estimation.
8591 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
8593 * tree-inline.h (struct copy_body_data): Add transform_parameter.
8594 * tree-inline.c (is_parameter_of): New predicate.
8595 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
8596 a parameter has been remapped.
8597 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
8598 (optimize_inline_calls): Initialize transform_parameter.
8599 (copy_gimple_seq_and_replace_locals): Likewise.
8600 (tree_function_versioning): Likewise.
8601 (maybe_inline_call_in_expr): Likewise.
8603 2013-09-17 Nick Clifton <nickc@redhat.com>
8605 * config/msp430/msp430-protos.h: Add prototypes for new functions.
8606 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
8608 (is_attr_func): New function.
8609 (msp430_is_interrupt_func): New function.
8610 (is_naked_func): New function.
8611 (is_reentrant_func): New function.
8612 (is_critical_func): New function.
8613 (msp430_start_function): Add annotations for function attributes.
8614 (msp430_attr): New function.
8615 (msp430_attribute_table): New.
8616 (msp430_function_section): New function.
8617 (TARGET_ASM_FUNCTION_SECTION): Define.
8618 (msp430_builtin): New enum.
8619 (msp430_init_builtins): New function.
8620 (msp430_builtin_devl): New function.
8621 (msp430_expand_builtin): New function.
8622 (TARGET_INIT_BUILTINS): Define.
8623 (TARGET_EXPAND_BUILTINS): Define.
8624 (TARGET_BUILTIN_DECL): Define.
8625 (msp430_expand_prologue): Add support for naked, interrupt,
8626 critical and reentrant functions.
8627 (msp430_expand_epilogue): Likewise.
8628 (msp430_print_operand): Handle 'O' character.
8629 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
8631 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
8632 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
8633 (pushm): Use a 'n' rather than an 'i' constraint.
8634 (msp_return): Add generation of the interrupt return instruction.
8635 (disable_interrupts): New pattern.
8636 (enable_interrupts): New pattern.
8637 (push_intr_state): New pattern.
8638 (pop_intr_state): New pattern.
8639 (bic_SR): New pattern.
8640 (bis_SR): New pattern.
8641 * doc/extend.texi: Document MSP430 function attributes and builtin
8644 2013-09-17 Richard Biener <rguenther@suse.de>
8646 PR tree-optimization/58432
8647 * tree-loop-distribution.c (tree_loop_distribution): Also
8648 scan PHIs for outside loop uses and seed a partition from them.
8650 2013-09-17 Bin Cheng <bin.cheng@arm.com>
8652 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
8653 (restructure_reference): Call backtrace_base_for_ref.
8655 2013-09-17 Alan Modra <amodra@gmail.com>
8658 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
8661 2013-09-16 DJ Delorie <dj@redhat.com>
8663 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
8664 vregs location for RL78/G10.
8665 (rl78_expand_prologue): Avoid SEL on G10.
8666 (rl78_expand_epilogue): Likewise.
8667 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
8668 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
8669 __RL78_G10__ when appropriate.
8670 (ASM_SPEC): Pass -mg10 along to the assembler.
8671 * config/rl78/rl78.md (sel_rb): Disable for G10.
8672 * config/rl78/rl78.opt: Add -mg10 option.
8673 * config/rl78/t-rl78: Add -mg10 multilib.
8675 2013-09-16 Xinliang David Li <davidxl@google.com>
8677 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
8678 * omp-low.c (omp_max_vf): Ditto.
8679 (expand_omp_simd): Ditto.
8680 * tree-vectorizer.c (vectorize_loops): Ditto.
8681 (gate_vect_slp): Ditto.
8682 (gate_increase_alignment): Ditto.
8683 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
8684 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
8685 (gate_tree_vectorize): Name change.
8686 (tree_vectorize): Ditto.
8687 (pass_vectorize::gate): Call new function.
8688 (pass_vectorize::execute): Ditto.
8689 * opts.c: O3 default setting change.
8690 (finish_options): Check new flag.
8691 * doc/invoke.texi: Document new flags.
8692 * common.opt: New flags.
8694 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
8696 * doc/tm.texi.in (Cond Exec Macros): Remove node.
8697 (Condition Code): Don't reference it.
8698 * doc/tm.texi: Regenerate.
8700 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
8703 * lra-constraints.c (undo_optional_reloads): Consider all optional
8704 reload even if it did not get a hard reg.
8706 2013-09-16 Teresa Johnson <tejohnson@google.com>
8708 * dumpfile.c (dump_loc): Remove newline emission.
8709 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
8710 emission to dump_printf_loc calls where missing.
8711 (vect_mark_for_runtime_alias_test): Ditto.
8712 (vect_analyze_data_ref_dependence): Ditto.
8713 (vect_analyze_data_ref_dependences): Ditto.
8714 (vect_slp_analyze_data_ref_dependence): Ditto.
8715 (vect_slp_analyze_data_ref_dependences): Ditto.
8716 (vect_compute_data_ref_alignment): Ditto.
8717 (vect_update_misalignment_for_peel): Ditto.
8718 (vect_verify_datarefs_alignment): Ditto.
8719 (vector_alignment_reachable_p): Ditto.
8720 (vect_get_data_access_cost): Ditto.
8721 (vect_enhance_data_refs_alignment): Ditto.
8722 (vect_find_same_alignment_drs): Ditto.
8723 (vect_analyze_data_refs_alignment): Ditto.
8724 (vect_analyze_group_access): Ditto.
8725 (vect_analyze_data_ref_access): Ditto.
8726 (vect_analyze_data_ref_accesses): Ditto.
8727 (vect_prune_runtime_alias_test_list): Ditto.
8728 (vect_analyze_data_refs): Ditto.
8729 (vect_create_addr_base_for_vector_ref): Ditto.
8730 (vect_create_data_ref_ptr): Ditto.
8731 (vect_grouped_store_supported): Ditto.
8732 (vect_grouped_load_supported): Ditto.
8733 * value-prof.c (check_counter): Ditto.
8734 (check_ic_target): Ditto.
8735 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
8736 (vect_recog_widen_mult_pattern): Ditto.
8737 (vect_recog_widen_sum_pattern): Ditto.
8738 (vect_recog_over_widening_pattern): Ditto.
8739 (vect_recog_widen_shift_pattern): Ditto.
8740 (vect_recog_rotate_pattern): Ditto.
8741 (vect_recog_vector_vector_shift_pattern): Ditto.
8742 (vect_recog_divmod_pattern): Ditto.
8743 (vect_recog_mixed_size_cond_pattern): Ditto.
8744 (vect_recog_bool_pattern): Ditto.
8745 (vect_pattern_recog_1): Ditto.
8746 (vect_pattern_recog): Ditto.
8747 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
8748 (vect_is_simple_iv_evolution): Ditto.
8749 (vect_analyze_scalar_cycles_1): Ditto.
8750 (vect_get_loop_niters): Ditto.
8751 (vect_analyze_loop_1): Ditto.
8752 (vect_analyze_loop_form): Ditto.
8753 (vect_analyze_loop_operations): Ditto.
8754 (vect_analyze_loop_2): Ditto.
8755 (vect_analyze_loop): Ditto.
8756 (report_vect_op): Ditto.
8757 (vect_is_slp_reduction): Ditto.
8758 (vect_is_simple_reduction_1): Ditto.
8759 (vect_get_known_peeling_cost): Ditto.
8760 (vect_estimate_min_profitable_iters): Ditto.
8761 (vect_model_reduction_cost): Ditto.
8762 (vect_model_induction_cost): Ditto.
8763 (get_initial_def_for_induction): Ditto.
8764 (vect_create_epilog_for_reduction): Ditto.
8765 (vectorizable_reduction): Ditto.
8766 (vectorizable_induction): Ditto.
8767 (vectorizable_live_operation): Ditto.
8768 (vect_loop_kill_debug_uses): Ditto.
8769 (vect_transform_loop): Ditto.
8770 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
8771 (vect_stmt_relevant_p): Ditto.
8772 (process_use): Ditto.
8773 (vect_mark_stmts_to_be_vectorized): Ditto.
8774 (vect_model_simple_cost): Ditto.
8775 (vect_model_promotion_demotion_cost): Ditto.
8776 (vect_model_store_cost): Ditto.
8777 (vect_get_store_cost): Ditto.
8778 (vect_model_load_cost): Ditto.
8779 (vect_get_load_cost): Ditto.
8780 (vect_init_vector_1): Ditto.
8781 (vect_get_vec_def_for_operand): Ditto.
8782 (vect_finish_stmt_generation): Ditto.
8783 (vectorizable_call): Ditto.
8784 (vectorizable_conversion): Ditto.
8785 (vectorizable_assignment): Ditto.
8786 (vectorizable_shift): Ditto.
8787 (vectorizable_operation): Ditto.
8788 (vectorizable_store): Ditto.
8789 (vectorizable_load): Ditto.
8790 (vectorizable_condition): Ditto.
8791 (vect_analyze_stmt): Ditto.
8792 (vect_transform_stmt): Ditto.
8793 (vect_is_simple_use): Ditto.
8794 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
8795 (vect_can_advance_ivs_p): Ditto.
8796 (vect_update_ivs_after_vectorizer): Ditto.
8797 (vect_do_peeling_for_loop_bound): Ditto.
8798 (vect_gen_niters_for_prolog_loop): Ditto.
8799 (vect_update_inits_of_drs): Ditto.
8800 (vect_create_cond_for_alias_checks): Ditto.
8801 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
8802 (vect_build_slp_tree_1): Ditto.
8803 (vect_supported_load_permutation_p): Ditto.
8804 (vect_analyze_slp_instance): Ditto.
8805 (vect_analyze_slp): Ditto.
8806 (vect_make_slp_decision): Ditto.
8807 (vect_detect_hybrid_slp): Ditto.
8808 (vect_bb_vectorization_profitable_p): Ditto.
8809 (vect_slp_analyze_bb_1): Ditto.
8810 (vect_update_slp_costs_according_to_vf): Ditto.
8811 (vect_get_mask_element): Ditto.
8812 (vect_transform_slp_perm_load): Ditto.
8813 (vect_schedule_slp_instance): Ditto.
8814 (vect_schedule_slp): Ditto.
8815 (vect_slp_transform_bb): Ditto.
8816 * profile.c (read_profile_edge_counts): Ditto.
8817 (compute_branch_probabilities): Ditto.
8818 * coverage.c (get_coverage_counts): Ditto.
8820 2013-09-16 Diego Novillo <dnovillo@google.com>
8822 * tree-core.h: Add missing comment lines from refactoring of tree.h.
8824 2013-09-16 Jan Hubicka <jh@suse.cz>
8826 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
8827 abstract functions; for static functions check the presence of body.
8829 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8831 * config/aarch64/aarch64-simd-builtins.def (fma): New.
8832 * config/aarch64/aarch64-simd.md
8833 (aarch64_mla_elt<mode>): New.
8834 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8835 (aarch64_mls_elt<mode>): Likewise.
8836 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8837 (aarch64_fma4_elt<mode>): Likewise.
8838 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8839 (aarch64_fma4_elt_to_128v2df): Likewise.
8840 (aarch64_fma4_elt_to_64df): Likewise.
8841 (fnma<mode>4): Likewise.
8842 (aarch64_fnma4_elt<mode>): Likewise.
8843 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8844 (aarch64_fnma4_elt_to_128v2df): Likewise.
8845 (aarch64_fnma4_elt_to_64df): Likewise.
8846 * config/aarch64/iterators.md (VDQSF): New.
8847 * config/aarch64/arm_neon.h
8848 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
8849 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8851 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8853 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8854 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8855 (aarch64_mul3_elt_to_128df): Likewise.
8856 (aarch64_mul3_elt_to_64v2df): Likewise.
8857 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8859 (VMUL_CHANGE_NLANES) Likewise.
8862 * config/aarch64/arm_neon.h
8863 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8865 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8867 * config/aarch64/arm_neon.h
8868 (vcvtx_high_f32_f64): Fix parameters.
8870 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8871 Uros Bizjak <ubizjak@gmail.com>
8873 * config/alpha.c: Include tree-ssanames.h.
8875 2013-09-16 Richard Biener <rguenther@suse.de>
8877 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8878 (dot_rdg_1): Handle control_dd.
8879 (create_edge_for_control_dependence): New function.
8880 (create_rdg_edges): Add control dependences if asked for.
8881 (build_rdg): Likewise.
8882 (generate_loops_for_partition): If there are not necessary
8883 control stmts remove all their dependencies.
8884 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8885 (distribute_loop): Pass on control dependences.
8886 (tree_loop_distribution): Compute control dependences and remove
8887 restriction on number of loop nodes.
8889 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8891 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8894 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8896 * cse.c (try_const_anchors): Punt on CC modes.
8898 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8900 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8902 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8905 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8908 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8911 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8913 (darwin_objc1_section): Likewise.
8914 (darwin_file_end): Emit Image Info section when required.
8916 2013-09-14 Jan Hubicka <jh@suse.cz>
8918 * tree-into-ssa.c (gate_into_ssa): New.
8919 (pass_data_build_ssa): Use it.
8920 * cgraph.h (expand_thunk): Update prototype.
8921 * cgraphunit.c (analyze_function): Expand thunks early.
8922 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8923 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8924 set lowered flag; do not add new function.
8925 (assemble_thunks_and_aliases): Update.
8926 * tree-ssa.c (gate_init_datastructures): New gate.
8927 (pass_data_init_datastructures): Use it.
8929 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8932 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8933 xmm register set described in the psABI.
8935 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8937 * dwarf2out.c (should_emit_struct_debug): Add check
8938 for type_decl variable is not NULL.
8940 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8942 * config.gcc: Use new winnt-c.c target hooks
8943 * config/t-winnt: New file
8944 * config/winnt-c.c: New file
8945 * doc/tm.texi.in: Document new hook
8946 * doc/tm.texi: Regenerated
8948 2013-09-13 Jan Hubicka <jh@suse.cz>
8951 * ipa-inline.c (check_callers): New function.
8952 (check_caller_edge): Remove.
8953 (want_inline_function_to_all_callers_p): Also permit alises that are
8955 (inline_to_all_callers): Terminate the walk when devirtualization
8956 introduce new calls.
8958 2013-09-13 Jan Hubicka <jh@suse.cz>
8960 * ipa-inline-analysis.c (struct growth_data): Add node.
8961 (do_estimate_growth_1): Fix detection of recursion.
8963 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8965 PR tree-optimization/58392
8966 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8967 to avoid shadowing of outer loop variable. If
8968 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8969 replace_by_duplicate_decl simduid of loops that have it set and
8970 set dest_cfun->has_simduid_loops and/or
8971 dest_cfun->has_force_vect_loops.
8972 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8973 instead of maybe_lookup_decl.
8974 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8975 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8976 call. Copy over force_vect and copy and remap simduid. Set
8977 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8978 (copy_cfg_body): Remove blocks_to_copy argument. Use
8979 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8980 caller. Don't set cfun->has_simduid_loops and/or
8981 cfun->has_force_vect_loops here.
8982 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8984 (expand_call_inline, tree_function_versioning): Adjust copy_body
8987 2013-09-13 Martin Jambor <mjambor@suse.cz>
8990 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8991 the assert if the edge was a speculative one.
8993 2013-09-13 Richard Biener <rguenther@suse.de>
8995 * tree-data-ref.h (known_dependences_p): Move here ...
8996 * tree-loop-distribution.c (known_dependences_p): ... from here.
8997 (dump_rdg_component, debug_rdg_component): Remove.
8999 (generate_loops_for_partition): Use gimple_uid instead of
9000 relying on matching stmt visit order.
9001 (rdg_build_partitions): Take starting stmt vector.
9002 (ldist_gen): Merge into ...
9003 (distribute_loop): ... this function. Do not compute starting
9005 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
9007 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9009 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
9010 Set type attribute correctly. Set predicable_short_it attribute.
9011 (cmpsi_shiftsi): Remove %? from output template.
9013 2013-09-13 Richard Biener <rguenther@suse.de>
9015 * tree-loop-distribution.c (struct rdg_component,
9016 rdg_defs_used_in_other_loops_p, free_rdg_components,
9017 rdg_build_components): Remove.
9018 (stmts_from_loop): Do not record virtual PHIs.
9019 (generate_loops_for_partition): Skip virtual PHIs.
9020 (build_rdg_partition_for_component): Rename to ...
9021 (build_rdg_partition_for_vertex): ... this and adjust.
9022 (rdg_build_partitions): Take a vector of starting vertices
9023 instead of components. Remove unnecessary leftover handling.
9024 (ldist_gen): Do not build components or record other stores.
9025 (distribute_loop): Do not distribute loops containing stmts
9028 2013-09-13 Christian Bruel <christian.bruel@st.com>
9031 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
9033 2013-09-13 Kai Tietz <ktietz@redhat.com>
9035 * config.gcc: Separate cases for mingw and cygwin targets,
9036 and add 64-bit cygwin target case.
9038 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
9039 dll-export inline-functions.
9040 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
9042 2013-09-13 Jeff Law <law@redhat.com>
9046 2013-09-06 Jeff Law <law@redhat.com>
9048 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9049 edge implied equivalences into successor phis.
9051 2013-09-12 DJ Delorie <dj@redhat.com>
9053 * config/rl78/rl78-virt.md: Change from | to \; for asm line
9056 2013-09-12 Brooks Moses <bmoses@google.com>
9059 * Makefile.in: Do not install driver binaries in $(target)/bin.
9061 2013-09-12 DJ Delorie <dj@redhat.com>
9063 * config/rl78/rl78.opt (mrelax): New.
9064 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
9065 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
9067 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
9069 (rl78_expand_epilogue): Likewise.
9071 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
9074 * lra-eliminations.c (remove_reg_equal_offset_note): New.
9075 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
9076 pointer elimination with using remove_reg_equal_offset_note.
9078 2013-09-12 DJ Delorie <dj@redhat.com>
9080 * config/msp430/: New port.
9081 * config.gcc (msp430): Added.
9082 * doc/invoke.texi: Document MSP430 options.
9083 * doc/install.texi: Document msp430-elf
9084 * doc/md.texi: Document msp430-elf
9085 * doc/contrib.texi: Document msp430-elf
9087 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
9090 2013-09-12 Martin Jambor <mjambor@suse.cz>
9093 * ipa-prop.c (remove_described_reference): Give up if the edge in the
9094 reference descriptor is NULL.
9095 (ipa_edge_removal_hook): If owning a reference descriptor, set its
9098 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
9100 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
9101 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
9102 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
9103 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
9104 make_temp_ssa_name): move to tree-ssanames.h
9105 * tree-ssa-alias.h: Move prototype.
9106 * tree-ssa.h: Include tree-ssanames.h.
9107 * tree-ssanames.c (FREE_SSANAMES): Move to here.
9108 * tree-ssanames.h: New. Move items from tree-flow*.h
9109 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
9111 2013-09-12 Richard Biener <rguenther@suse.de>
9113 PR tree-optimization/58404
9114 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
9115 propagate non-invariant addresses into dereferences wrapped
9116 in component references.
9118 2013-09-12 Richard Biener <rguenther@suse.de>
9120 PR tree-optimization/58402
9121 * passes.def: Move pass_late_warn_uninitialized later.
9123 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
9125 * tree-ssa.h: New. Move content from tree-flow.h and
9127 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
9128 Move prototypes belonging to tree-ssa.c.
9129 * tree-flow-inline.h (redirect_edge_var_map_def,
9130 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
9132 * gimple.h: Adjust prototypes.
9133 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
9135 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
9136 * tree.h: Move prototype to tree-ssa.h.
9137 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
9138 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
9139 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
9140 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
9141 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
9142 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
9143 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
9144 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
9145 graphite-blocking.c, graphite-clast-to-gimple.c,
9146 graphite-dependences.c, graphite-interchange.c,
9147 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
9148 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
9149 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
9150 ipa-reference.c, ipa-split.c, ipa-utils.c,
9151 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
9152 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
9153 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
9154 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
9155 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
9156 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
9157 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
9158 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
9159 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
9160 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
9161 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
9162 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
9163 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
9164 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
9165 value-prof.c, var-tracking.c,
9166 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
9168 2013-09-12 Richard Biener <rguenther@suse.de>
9170 PR tree-optimization/58396
9171 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
9172 (build_rdg): Take a loop-nest parameter, fix memleaks.
9173 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
9175 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9177 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
9180 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
9182 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
9183 The FMA instruction names should have a 'v' prefix.
9185 2013-09-12 Richard Biener <rguenther@suse.de>
9187 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
9188 (dot_rdg): Use popen instead of system in optional code.
9189 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
9190 (already_processed_vertex_p): Adjust.
9191 (has_anti_or_output_dependence, predecessor_has_mem_write,
9192 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
9193 rdg_flag_uses): Remove.
9194 (rdg_flag_vertex): Simplify.
9195 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
9197 (build_rdg_partition_for_component): Process the first vertex
9198 of a component only.
9199 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
9201 2013-09-12 Alan Modra <amodra@gmail.com>
9203 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
9205 2013-09-11 DJ Delorie <dj@redhat.com>
9206 Nick Clifton <nickc@redhat.com>
9208 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
9209 (rl78_stack_based_mem): New.
9210 * config/rl78/constraints.md (Iv08): New.
9219 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
9221 (movsi): Change from expand to insn-and-split.
9222 (ashrsi3): Clobber AX.
9226 * config/rl78/rl78.md (CC_REG): Fix.
9227 (addsi3): Allow memory and immediate operands.
9228 (addsi3_internal): Split into...
9229 (addsi3_internal_virt): ...new, and ...
9230 (addsi3_internal_real): ...new.
9232 (subsi3_internal_virt): New.
9233 (subsi3_internal_real): New.
9234 (mulsi3): Add memory operand.
9235 (mulsi3_rl78): Likewise.
9236 (mulsi3_g13): Likewise.
9237 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
9238 (cbranchqi4_real): Add more constraint options.
9239 (cbranchhi4_real): Expand pattern.
9240 (cbranchhi4_real_signed): New.
9241 (cbranchhi4_real_inverted): New.
9242 (cbranchsi4_real_lt): New.
9243 (cbranchsi4_real_ge): New.
9244 (cbranchsi4_real_signed): New.
9245 (cbranchsi4_real): New.
9247 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
9249 (lshrsi3_virt): Likewise.
9250 (ashlsi3_virt): Likewise.
9251 (cbranchqi4_virt_signed): New.
9252 (cbranchhi4_virt_signed): New.
9253 (cbranchsi4_virt): New.
9254 * config/rl78/rl78.c: Whitespace fixes throughout.
9255 (move_elim_pass): New.
9256 (pass_data_rl78_move_elim): New.
9257 (pass_rl78_move_elim): New.
9258 (make_pass_rl78_move_elim): New.
9259 (rl78_devirt_info): Run devirt earlier.
9260 (rl78_move_elim_info): New.
9261 (rl78_asm_file_start): Register it.
9262 (rl78_split_movsi): New.
9263 (rl78_as_legitimate_address): Allow virtual base registers when
9265 (rl78_addr_space_convert): Remove spurious debug stuff.
9266 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
9267 (rl78_print_operand): More cases for not printing '#'.
9268 (rl78_expand_compare): Remove most of the logic.
9269 (content_memory): New.
9270 (clear_content_memory): New.
9271 (get_content_index): New.
9272 (get_content_name): New.
9273 (display_content_memory): New.
9274 (update_content): New.
9275 (record_content): New.
9276 (already_contains): New.
9277 (insn_ok_now): Re-recog insns with virtual registers.
9278 (add_postponed_content_update): New.
9279 (process_postponed_content_update): New.
9280 (gen_and_emit_move): New.
9281 (transcode_memory_rtx): Record new location content.
9282 Use gen_and_emit_move.
9283 (force_into_acc): New.
9284 (move_to_acc): Use gen_and_emit_move.
9285 (move_from_acc): Likewise.
9286 (move_acc_to_reg): Likewise.
9287 (move_to_x): Likewise.
9288 (move_to_hl): Likewise.
9289 (move_to_de): Likewise.
9290 (rl78_alloc_physical_registers_op1): Record location content.
9291 (has_constraint): New.
9292 (rl78_alloc_physical_registers_op2): Record location content.
9294 (rl78_alloc_physical_registers_ro1): Likewise.
9295 (rl78_alloc_physical_registers_cmp): Likewise.
9296 (rl78_alloc_physical_registers_umul): Likewise.
9297 (rl78_alloc_address_registers_macax): New.
9298 (rl78_alloc_physical_registers): Initialize and set location
9299 content memory as needed.
9300 (rl78_reorg): Make sure split2 is called.
9301 (rl78_rtx_costs): New.
9303 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
9305 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
9306 for (not (neg ...)) and (neg (not ...)) cases.
9308 2013-09-11 Richard Biener <rguenther@suse.de>
9311 * passes.def: Split critical edges before late uninit warning passes.
9312 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
9314 2013-09-11 Jakub Jelinek <jakub@redhat.com>
9316 PR tree-optimization/58385
9317 * fold-const.c (build_range_check): If both low and high are NULL,
9318 use omit_one_operand_loc to preserve exp side-effects.
9320 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9322 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
9324 2013-09-11 Richard Biener <rguenther@suse.de>
9326 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
9327 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
9328 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
9329 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
9330 stmts_from_loop, known_dependences_p, build_empty_rdg,
9331 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
9332 * tree-loop-distribution.c: ... here.
9333 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
9334 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
9335 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
9336 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
9337 * tree-loop-distribution.c: ... here.
9338 * tree-loop-distribution.c: Include gimple-pretty-print.h.
9339 (struct partition_s): Add loops member.
9340 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
9341 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
9342 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
9344 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
9345 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
9346 Sergey Lega <sergey.s.lega@intel.com>
9347 Anna Tikhonova <anna.tikhonova@intel.com>
9348 Ilya Tocar <ilya.tocar@intel.com>
9349 Andrey Turetskiy <andrey.turetskiy@intel.com>
9350 Ilya Verbin <ilya.verbin@intel.com>
9351 Kirill Yukhin <kirill.yukhin@intel.com>
9352 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9354 * config/i386/constraints.md (k): New.
9356 * config/i386/i386.c (const regclass_map): Add new mask registers.
9357 (dbx_register_map): Ditto.
9358 (dbx64_register_map): Ditto.
9359 (svr4_dbx_register_map): Ditto.
9360 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
9362 (ix86_preferred_reload_class): Disable constants for mask registers.
9363 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
9364 (ix86_hard_regno_mode_ok): Support new mask registers.
9365 (x86_order_regs_for_local_alloc): Ditto.
9366 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
9367 (FIXED_REGISTERS): Add new mask registers.
9368 (CALL_USED_REGISTERS): Ditto.
9369 (REG_ALLOC_ORDER): Ditto.
9370 (VALID_MASK_REG_MODE): New.
9371 (FIRST_MASK_REG): Ditto.
9372 (LAST_MASK_REG): Ditto.
9373 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
9374 (MAYBE_MASK_CLASS_P): New.
9375 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
9376 (REG_CLASS_CONTENTS): Ditto.
9377 (MASK_REGNO_P): New.
9378 (ANY_MASK_REG_P): Ditto.
9379 (HI_REGISTER_NAMES): Add new mask registers.
9380 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
9381 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
9383 (attribute "type"): Add mskmov, msklog.
9384 (attribute "length_immediate"): Support them.
9385 (attribute "memory"): Ditto.
9386 (attribute "prefix_0f"): Ditto.
9387 (*movhi_internal): Support new mask registers.
9388 (*movqi_internal): Ditto.
9389 (define_split): Split out clobber pattern is a logic
9390 insn on mask registers.
9391 (*k<logic><mode>): New.
9392 (*andhi_1): Extend to support mask regs.
9393 (*andqi_1): Extend to support mask regs.
9395 (define_split): Split and-not to and and not if operands
9397 (*<code><mode>_1): Separate HI mode to new pattern...
9398 (*<code>hi_1): This.
9399 (*<code>qi_1): Extend to support mask regs.
9401 (kortestzhi): Ditto.
9402 (kortestchi): Ditto.
9404 (*one_cmpl<mode>2_1): Remove HImode and handle it...
9405 (*one_cmplhi2_1): ...Here, now with mask registers support.
9406 (*one_cmplqi2_1): Support new mask registers.
9407 (HI/QImode arithmetics splitter): Don't split if mask registers
9409 (HI/QImode not splitter): Ditto.
9410 * config/i386/predicated.md (mask_reg_operand): New.
9411 (general_reg_operand): Ditto.
9413 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
9415 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
9416 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
9418 2013-09-10 Jeff Law <law@redhat.com>
9420 PR tree-optimization/58380
9421 * tree-ssa-threadupdate.c (thread_block): Recognize another case
9422 of threading through a buried loop header.
9424 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
9425 return value for single successor case.
9427 2013-09-10 Jan Hubicka <jh@suse.cz>
9429 * ipa-devirt.c (ipa_devirt): Enable with LTO.
9431 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
9434 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
9435 support conditional execution.
9436 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
9438 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
9440 * lra.c (lra): Clear lra_optional_reload_pseudos before every
9442 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
9443 Check destination too to check move insn.
9444 (undo_optional_reloads): Add check that the original peudo did not
9445 changed its allocation and the optional reload was inherited on last
9446 inheritance pass. Break loop after deciding to keep optional reload.
9447 (lra_undo_inheritance): Add check that inherited pseudo still in
9450 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
9452 * config/aarch64/aarch64.md (generic_sched): New.
9453 * config/aarch64/aarch64-generic.md (load): Make conditional
9454 on generic_sched attribute.
9455 (nonload): Likewise.
9457 2013-09-10 Jan Hubicka <jh@suse.cz>
9459 * lto-cgraph.c: Include ipa-utils.h.
9460 (compute_ltrans_boundary): Also add possible targets into the boundary.
9462 2013-09-10 Jan Hubicka <jh@suse.cz>
9464 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
9465 VAR_DECL of vtable rather than full expression.
9467 2013-09-10 Jan Hubicka <jh@suse.cz>
9468 Paolo Carlini <paolo.carlini@oracle.com>
9470 * cgraphunit.c (analyze_functions): Save input_location, set it
9471 to UNKNOWN_LOCATION and restore it at the end.
9473 2013-09-10 Martin Jambor <mjambor@suse.cz>
9475 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
9476 represented by a thunk.
9478 2013-09-10 Jeff Law <law@redhat.com>
9480 PR tree-optimization/58343
9481 * tree-ssa-threadupdate.c (thread_block): Identify and disable
9482 jump threading requests through loop headers buried in the middle
9483 of a jump threading path.
9485 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
9486 in return value/type.
9488 2013-09-10 Jakub Jelinek <jakub@redhat.com>
9490 PR rtl-optimization/58365
9491 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
9492 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
9495 2013-09-10 Richard Biener <rguenther@suse.de>
9497 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
9498 * tree-data-ref.c (create_rdg_vertices): Collect all data
9499 references, signal failure to the caller, use data-ref API.
9500 (build_rdg): Compute data references only once. Maintain lifetime
9501 of data references and data dependences from within RDG.
9502 (free_rdg): Free dependence relations.
9503 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
9504 inventing extra dependences.
9505 (distribute_loop): Update for RDG API changes.
9507 2013-09-10 Kai Tietz <ktietz@redhat.com>
9509 * doc/invoke.texi (fms-extensions): Document changed
9510 behavior for ms-abi targets.
9511 * config/i386/i386.c (ix86_option_override_internal):
9512 Set default value of option -fms-extension for ms-abi targets.
9514 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9516 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
9518 2013-09-10 Alan Modra <amodra@gmail.com>
9521 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
9523 2013-09-10 Alan Modra <amodra@gmail.com>
9525 * config/rs6000/predicates.md (add_cint_operand): New.
9526 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
9527 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
9528 using add_cint_operand.
9529 (largetoc_high_plus_aix): Likewise.
9531 2013-09-09 Jakub Jelinek <jakub@redhat.com>
9533 PR tree-optimization/58364
9534 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
9535 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
9536 the current range can't be an unconditional true or false.
9538 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
9540 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
9542 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
9544 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
9546 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
9549 * doc/invoke.texi (-Wdelete-incomplete): Document it.
9551 2013-09-09 Ian Bolton <ian.bolton@arm.com>
9553 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
9554 NO_REGS for immediate that can't be moved directly into FP_REGS.
9556 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9558 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
9559 comparison with negated operand.
9560 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
9563 2013-09-09 Richard Biener <rguenther@suse.de>
9566 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
9567 of a subloop record all its block as affecting loop-closed SSA form.
9569 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9571 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
9572 of an rtx/bitpos pair.
9573 (store_fixed_bit_field): Update accordingly.
9575 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9577 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
9579 * builtins.c (expand_errno_check): Likewise.
9580 * dwarf2cfi.c (init_return_column_size): Likewise.
9581 * except.c (sjlj_mark_call_sites): Likewise.
9582 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
9583 * lra-constraints.c (emit_inc): Likewise.
9584 * ree.c (combine_set_extension): Likewise.
9585 * regmove.c (fixup_match_2): Likewise.
9586 * reload1.c (inc_for_reload): Likewise.
9588 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9590 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
9591 (force_to_mode, simplify_shift_const_1, simplify_comparison):
9592 Use gen_int_mode with the mode of the associated simplify_* call.
9593 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
9594 * expmed.c (expand_shift_1): Likewise.
9595 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9596 * loop-iv.c (iv_number_of_iterations): Likewise.
9597 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9598 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
9600 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9602 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
9603 of the associated expand_* call.
9604 (asan_emit_stack_protection): Likewise.
9605 * builtins.c (round_trampoline_addr): Likewise.
9606 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
9607 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
9608 (emit_store_flag): Likewise.
9609 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
9610 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
9612 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9613 * ifcvt.c (noce_try_store_flag_constants): Likewise.
9614 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9615 * modulo-sched.c (generate_prolog_epilog): Likewise.
9616 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
9617 (expand_ctz, expand_ffs, expand_unop): Likewise.
9619 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9621 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
9622 of the associated gen_rtx_* call.
9623 * caller-save.c (init_caller_save): Likewise.
9624 * combine.c (find_split_point, make_extraction): Likewise.
9625 (make_compound_operation): Likewise.
9626 * dwarf2out.c (mem_loc_descriptor): Likewise.
9627 * explow.c (plus_constant, probe_stack_range): Likewise.
9628 * expmed.c (expand_mult_const): Likewise.
9629 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
9630 * reload1.c (init_reload): Likewise.
9631 * valtrack.c (cleanup_auto_inc_dec): Likewise.
9632 * var-tracking.c (adjust_mems): Likewise.
9633 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
9634 rather than gen_rtx_fmt_ee.
9636 2013-09-09 Jan Hubicka <jh@suse.cz>
9639 * value-prof.c (gimple_ic): Copy also abnormal edges.
9641 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9643 * asan.c (asan_shadow_cst): Use gen_int_mode.
9645 2013-09-08 Jan Hubicka <jh@suse.cz>
9647 * ipa-profile.c: Add toplevel comment.
9648 (ipa_propagate_frequency_1): Be more conservative when profile is read.
9649 (contains_hot_call_p): New function.
9650 (ipa_propagate_frequency): Set frequencies based on counts when
9652 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
9653 profile; do not tamper with profile after inlining if it is read.
9655 2013-09-08 Jan Hubicka <jh@suse.cz>
9657 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
9660 2013-09-08 Jan Hubicka <jh@suse.cz>
9662 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
9665 2013-09-08 Jeff Law <law@redhat.com>
9668 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
9671 2013-09-08 Andi Kleen <ak@linux.intel.com>
9673 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
9675 2013-09-08 Jan Hubicka <jh@suse.cz>
9677 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
9679 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
9681 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
9682 for non-debug insns.
9683 * lra.c (new_insn_reg): Take the containing insn as a parameter.
9684 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
9685 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
9688 2013-09-08 Jan Hubicka <jh@suse.cz>
9690 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
9691 targets and devirtualize to BUILT_IN_UNREACHABLE.
9692 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
9693 * ipa.c (walk_polymorphic_call_targets): New function.
9694 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
9695 functions; use the new timevar.
9696 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
9697 was removed from the program.
9698 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
9699 not consider it in the walk when its vtable is dead.
9700 (possible_polymorphic_call_targets_1): Pass anonymous flag to
9702 (devirt_variable_node_removal_hook): New function.
9703 (possible_polymorphic_call_targets): Also register
9704 devirt_variable_node_removal_hook.
9705 (ipa_devirt): Do not do non-speculative devirtualization.
9706 (gate_ipa_devirt): One execute if devirtualizing speculatively.
9708 2013-09-08 Jan Hubicka <jh@suse.cz>
9710 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
9711 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
9712 varpool_remove_variable_insertion_hook): Declare.
9713 * varpool.c (varpool_node_hook_list): New structure.
9714 (first_varpool_node_removal_hook,
9715 first_varpool_variable_insertion_hook): New variables.
9716 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
9717 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
9718 varpool_remove_variable_insertion_hook,
9719 varpool_call_variable_insertion_hooks): New functions.
9720 (varpool_remove_node): Use it.
9722 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
9725 * diagnostic.c (diagnostic_build_prefix): When s.file is
9726 "<built-in>" don't output line and column numbers.
9728 2013-09-06 Jan Hubicka <jh@suse.cz>
9730 * cgraphunit.c (expand_thunk): Get body before touching arguments.
9731 * lto-streamer-out.c: Stream thunks, too.
9732 * lto-streamer-in.c (input_function): Pop cfun here
9733 (lto_read_body): Instead of here.
9735 2013-09-06 Caroline Tice <cmtice@google.com>
9737 * doc/install.texi: Add documentation for the --enable-vtable-verify
9738 and the --disable-libvtv configure options.
9740 2013-09-06 Jeff Law <law@redhat.com>
9742 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9743 edge implied equivalences into successor phis.
9745 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
9747 * resource.c (mark_referenced_resources): Handle COND_EXEC.
9749 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
9751 * resource.c (mark_target_live_regs): Compute resources taking
9752 into account if a call is predicated or not.
9754 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9756 * toplev.c (output_stack_usage): Be prepared for suffixes created by
9757 the compiler in the function names.
9759 2013-09-06 Jan Hubicka <jh@suse.cz>
9762 * ipa-inline.c (has_caller_p): New function.
9763 (want_inline_function_to_all_callers_p): Use it.
9764 (sum_callers, inline_to_all_callers): Break out from ...
9765 (ipa_inline): ... here.
9767 2013-09-06 Jan Hubicka <jh@suse.cz>
9769 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
9770 only when AVX is enabled.
9772 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9774 * config/aarch64/aarch64.md
9775 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
9776 is fpsimd_<load/store>2.
9777 (load_pair<mode>): Likewise.
9778 (store_pair<mode>): Likewise.
9780 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9782 * config/arm/types.md (type): Add "mrs" type.
9783 * config/aarch64/aarch64.md
9784 (aarch64_load_tp_hard): Make type "mrs".
9786 (load_tp_hard): Make type "mrs".
9787 * config/arm/cortex-a15.md: Update with new attributes.
9788 * config/arm/cortex-a5.md: Update with new attributes.
9789 * config/arm/cortex-a53.md: Update with new attributes.
9790 * config/arm/cortex-a7.md: Update with new attributes.
9791 * config/arm/cortex-a8.md: Update with new attributes.
9792 * config/arm/cortex-a9.md: Update with new attributes.
9793 * config/arm/cortex-m4.md: Update with new attributes.
9794 * config/arm/cortex-r4.md: Update with new attributes.
9795 * config/arm/fa526.md: Update with new attributes.
9796 * config/arm/fa606te.md: Update with new attributes.
9797 * config/arm/fa626te.md: Update with new attributes.
9798 * config/arm/fa726te.md: Update with new attributes.
9800 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9802 * config/aarch64/aarch64.md
9803 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
9804 (*movtf_aarch64): Likewise.
9806 (thumb1_movdi_insn): Use "multiple" for type where more than one
9807 instruction is used for a move.
9808 (*arm32_movhf): Likewise.
9809 (*thumb_movdf_insn): Likewise.
9811 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9813 * config/arm/types.md (type): Rename fcpys to fmov.
9815 (*arm_movsi_vfp): Rename type fcpys as fmov.
9816 (*thumb2_movsi_vfp): Likewise
9817 (*movhf_vfp_neon): Likewise
9818 (*movhf_vfp): Likewise
9819 (*movsf_vfp): Likewise
9820 (*thumb2_movsf_vfp): Likewise
9821 (*movsfcc_vfp): Likewise
9822 (*thumb2_movsfcc_vfp): Likewise
9823 * config/aarch64/aarch64-simd.md
9824 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
9825 * config/aarch64/aarch64.md
9826 (*movsi_aarch64): Replace type mov_reg with fmovs.
9827 (*movdi_aarch64): Likewise
9828 (*movsf_aarch64): Likewise
9829 (*movdf_aarch64): Likewise
9831 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
9832 * config/arm/iwmmxt.md
9833 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
9834 * config/arm/arm1020e.md: Update with new attributes.
9835 * config/arm/cortex-a15-neon.md: Update with new attributes.
9836 * config/arm/cortex-a5.md: Update with new attributes.
9837 * config/arm/cortex-a53.md: Update with new attributes.
9838 * config/arm/cortex-a7.md: Update with new attributes.
9839 * config/arm/cortex-a8-neon.md: Update with new attributes.
9840 * config/arm/cortex-a9.md: Update with new attributes.
9841 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9842 * config/arm/cortex-r4f.md: Update with new attributes.
9843 * config/arm/marvell-pj4.md: Update with new attributes.
9844 * config/arm/vfp11.md: Update with new attributes.
9846 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9848 * config/aarch64/aarch64.md
9849 (*madd<mode>): Fix type attribute.
9850 (*maddsi_uxtw): Likewise.
9851 (*msub<mode>): Likewise.
9852 (*msubsi_uxtw): Likewise.
9853 (<su_optab>maddsidi4): Likewise.
9854 (<su_optab>msubsidi4): Likewise.
9856 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9858 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9859 * config/arm/arm.md (core_cycles): Remove fdiv.
9860 * config/arm/vfp.md:
9861 (*sqrtsf2_vfp): Update for attribute changes.
9862 (*sqrtdf2_vfp): Likewise.
9863 * config/aarch64/aarch64.md:
9864 (sqrt<mode>2): Update for attribute changes.
9865 * config/arm/arm1020e.md: Update with new attributes.
9866 * config/arm/cortex-a15-neon.md: Update with new attributes.
9867 * config/arm/cortex-a5.md: Update with new attributes.
9868 * config/arm/cortex-a53.md: Update with new attributes.
9869 * config/arm/cortex-a7.md: Update with new attributes.
9870 * config/arm/cortex-a8-neon.md: Update with new attributes.
9871 * config/arm/cortex-a9.md: Update with new attributes.
9872 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9873 * config/arm/cortex-r4f.md: Update with new attributes.
9874 * config/arm/marvell-pj4.md: Update with new attributes.
9875 * config/arm/vfp11.md: Update with new attributes.
9877 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9879 * config/arm/types.md
9880 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9881 * config/aarch64/aarch64.md
9882 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9884 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9885 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9886 (float<GPI:mode><GPF:mode>2): Likewise.
9888 (*truncsisf2_vfp): Update with new attributes.
9889 (*truncsidf2_vfp): Likewise.
9890 (fixuns_truncsfsi2): Likewise.
9891 (fixuns_truncdfsi2): Likewise.
9892 (*floatsisf2_vfp): Likewise.
9893 (*floatsidf2_vfp): Likewise.
9894 (floatunssisf2): Likewise.
9895 (floatunssidf2): Likewise.
9896 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9897 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9898 * config/arm/arm1020e.md: Update with new attributes.
9899 * config/arm/cortex-a15-neon.md: Update with new attributes.
9900 * config/arm/cortex-a5.md: Update with new attributes.
9901 * config/arm/cortex-a53.md: Update with new attributes.
9902 * config/arm/cortex-a7.md: Update with new attributes.
9903 * config/arm/cortex-a8-neon.md: Update with new attributes.
9904 * config/arm/cortex-a9.md: Update with new attributes.
9905 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9906 * config/arm/cortex-r4f.md: Update with new attributes.
9907 * config/arm/marvell-pj4.md: Update with new attributes.
9908 * config/arm/vfp11.md: Update with new attributes.
9910 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9912 * config/aarch64/arm_neon.h
9913 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9914 (vqtbx<1,2,3,4><q>_s8): Likewise.
9916 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9918 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9919 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9920 (add<mode>3): Add type attribute.
9921 (add<mode>3): Likewise.
9922 (usadd<mode>3): Likewise.
9923 (ssadd<mode>3): Likewise.
9924 (sub<mode>3): Likewise.
9925 (sub<mode>3): Likewise.
9926 (ussub<mode>3): Likewise.
9927 (sssub<mode>3): Likewise.
9928 (ssmulsa3): Likewise.
9929 (usmulusa3): Likewise.
9930 (arm_usatsihi): Likewise.
9932 (*movdi_vfp): Add types for all instructions.
9933 (*movdi_vfp_cortexa8): Likewise.
9934 (*movhf_vfp_neon): Likewise.
9935 (*movhf_vfp): Likewise.
9936 (*movdf_vfp): Likewise.
9937 (*thumb2_movdf_vfp): Likewise.
9938 (*thumb2_movdfcc_vfp): Likewise.
9939 * config/arm/arm.md: Add type attribute to all insn patterns.
9940 (*thumb1_adddi3): Add type attribute.
9941 (*arm_adddi3): Likewise.
9942 (*adddi_sesidi_di): Likewise.
9943 (*adddi_zesidi_di): Likewise.
9944 (*thumb1_addsi3): Likewise.
9945 (addsi3_compare0): Likewise.
9946 (*addsi3_compare0_scratch): Likewise.
9947 (*compare_negsi_si): Likewise.
9948 (cmpsi2_addneg): Likewise.
9949 (*addsi3_carryin_<optab>): Likewise.
9950 (*addsi3_carryin_alt2_<optab>): Likewise.
9951 (*addsi3_carryin_clobercc_<optab>): Likewise.
9952 (*subsi3_carryin): Likewise.
9953 (*subsi3_carryin_const): Likewise.
9954 (*subsi3_carryin_compare): Likewise.
9955 (*subsi3_carryin_compare_const): Likewise.
9956 (*arm_subdi3): Likewise.
9957 (*thumb_subdi3): Likewise.
9958 (*subdi_di_zesidi): Likewise.
9959 (*subdi_di_sesidi): Likewise.
9960 (*subdi_zesidi_di): Likewise.
9961 (*subdi_sesidi_di): Likewise.
9962 (*subdi_zesidi_ze): Likewise.
9963 (thumb1_subsi3_insn): Likewise.
9964 (*arm_subsi3_insn): Likewise.
9965 (*anddi3_insn): Likewise.
9966 (*anddi_zesidi_di): Likewise.
9967 (*anddi_sesdi_di): Likewise.
9968 (*ne_zeroextracts): Likewise.
9969 (*ne_zeroextracts): Likewise.
9970 (*ite_ne_zeroextr): Likewise.
9971 (*ite_ne_zeroextr): Likewise.
9972 (*anddi_notdi_di): Likewise.
9973 (*anddi_notzesidi): Likewise.
9974 (*anddi_notsesidi): Likewise.
9975 (andsi_notsi_si): Likewise.
9976 (thumb1_bicsi3): Likewise.
9977 (*iordi3_insn): Likewise.
9978 (*iordi_zesidi_di): Likewise.
9979 (*iordi_sesidi_di): Likewise.
9980 (*thumb1_iorsi3_insn): Likewise.
9981 (*xordi3_insn): Likewise.
9982 (*xordi_zesidi_di): Likewise.
9983 (*xordi_sesidi_di): Likewise.
9984 (*arm_xorsi3): Likewise.
9985 (*andsi_iorsi3_no): Likewise.
9986 (*smax_0): Likewise.
9987 (*smax_m1): Likewise.
9988 (*arm_smax_insn): Likewise.
9989 (*smin_0): Likewise.
9990 (*arm_smin_insn): Likewise.
9991 (*arm_umaxsi3): Likewise.
9992 (*arm_uminsi3): Likewise.
9993 (*minmax_arithsi): Likewise.
9994 (*minmax_arithsi_): Likewise.
9995 (*satsi_<SAT:code>): Likewise.
9996 (arm_ashldi3_1bit): Likewise.
9997 (arm_ashrdi3_1bit): Likewise.
9998 (arm_lshrdi3_1bit): Likewise.
9999 (*arm_negdi2): Likewise.
10000 (*thumb1_negdi2): Likewise.
10001 (*arm_negsi2): Likewise.
10002 (*thumb1_negsi2): Likewise.
10003 (*negdi_extendsid): Likewise.
10004 (*negdi_zero_extend): Likewise.
10005 (*arm_abssi2): Likewise.
10006 (*thumb1_abssi2): Likewise.
10007 (*arm_neg_abssi2): Likewise.
10008 (*thumb1_neg_abss): Likewise.
10009 (one_cmpldi2): Likewise.
10010 (extend<mode>di2): Likewise.
10011 (*compareqi_eq0): Likewise.
10012 (*arm_extendhisi2addsi): Likewise.
10013 (*arm_movdi): Likewise.
10014 (*thumb1_movdi_insn): Likewise.
10015 (*arm_movt): Likewise.
10016 (*thumb1_movsi_insn): Likewise.
10017 (pic_add_dot_plus_four): Likewise.
10018 (pic_add_dot_plus_eight): Likewise.
10019 (tls_load_dot_plus_eight): Likewise.
10020 (*thumb1_movhi_insn): Likewise.
10021 (*thumb1_movsf_insn): Likewise.
10022 (*movdf_soft_insn): Likewise.
10023 (*thumb_movdf_insn): Likewise.
10024 (cbranchsi4_insn): Likewise.
10025 (cbranchsi4_scratch): Likewise.
10026 (*negated_cbranchsi4): Likewise.
10027 (*tbit_cbranch): Likewise.
10028 (*tlobits_cbranch): Likewise.
10029 (*tstsi3_cbranch): Likewise.
10030 (*cbranchne_decr1): Likewise.
10031 (*addsi3_cbranch): Likewise.
10032 (*addsi3_cbranch_scratch): Likewise.
10033 (*arm_cmpdi_insn): Likewise.
10034 (*arm_cmpdi_unsig): Likewise.
10035 (*arm_cmpdi_zero): Likewise.
10036 (*thumb_cmpdi_zero): Likewise.
10037 (*deleted_compare): Likewise.
10038 (*mov_scc): Likewise.
10039 (*mov_negscc): Likewise.
10040 (*mov_notscc): Likewise.
10041 (*cstoresi_eq0_thumb1_insn): Likewise.
10042 (cstoresi_nltu_thumb1): Likewise.
10043 (cstoresi_ltu_thu): Likewise.
10044 (thumb1_addsi3_addgeu): Likewise.
10045 (*arm_jump): Likewise.
10046 (*thumb_jump): Likewise.
10047 (*check_arch2): Likewise.
10048 (arm_casesi_internal): Likewise.
10049 (thumb1_casesi_dispatch): Likewise.
10050 (*arm_indirect_jump): Likewise.
10051 (*thumb1_indirect_jump): Likewise.
10053 (*and_scc): Likewise.
10054 (*ior_scc): Likewise.
10055 (*compare_scc): Likewise.
10056 (*cond_move): Likewise.
10057 (*cond_arith): Likewise.
10058 (*cond_sub): Likewise.
10059 (*cmp_ite0): Likewise.
10060 (*cmp_ite1): Likewise.
10061 (*cmp_and): Likewise.
10062 (*cmp_ior): Likewise.
10063 (*ior_scc_scc): Likewise.
10064 (*ior_scc_scc_cmp): Likewise.
10065 (*and_scc_scc): Likewise.
10066 (*and_scc_scc_cmp): Likewise.
10067 (*and_scc_scc_nod): Likewise.
10068 (*negscc): Likewise.
10069 (movcond_addsi): Likewise.
10070 (movcond): Likewise.
10071 (*ifcompare_plus_move): Likewise.
10072 (*if_plus_move): Likewise.
10073 (*ifcompare_move_plus): Likewise.
10074 (*if_move_plus): Likewise.
10075 (*ifcompare_arith_arith): Likewise.
10076 (*if_arith_arith): Likewise.
10077 (*ifcompare_arith_move): Likewise.
10078 (*if_arith_move): Likewise.
10079 (*ifcompare_move_arith): Likewise.
10080 (*if_move_arith): Likewise.
10081 (*ifcompare_move_not): Likewise.
10082 (*if_move_not): Likewise.
10083 (*ifcompare_not_move): Likewise.
10084 (*if_not_move): Likewise.
10085 (*ifcompare_shift_move): Likewise.
10086 (*if_shift_move): Likewise.
10087 (*ifcompare_move_shift): Likewise.
10088 (*if_move_shift): Likewise.
10089 (*ifcompare_shift_shift): Likewise.
10090 (*ifcompare_not_arith): Likewise.
10091 (*ifcompare_arith_not): Likewise.
10092 (*if_arith_not): Likewise.
10093 (*ifcompare_neg_move): Likewise.
10094 (*if_neg_move): Likewise.
10095 (*ifcompare_move_neg): Likewise.
10096 (*if_move_neg): Likewise.
10097 (prologue_thumb1_interwork): Likewise.
10098 (*cond_move_not): Likewise.
10099 (*sign_extract_onebit): Likewise.
10100 (*not_signextract_onebit): Likewise.
10101 (stack_tie): Likewise.
10102 (align_4): Likewise.
10103 (align_8): Likewise.
10104 (consttable_end): Likewise.
10105 (consttable_1): Likewise.
10106 (consttable_2): Likewise.
10107 (consttable_4): Likewise.
10108 (consttable_8): Likewise.
10109 (consttable_16): Likewise.
10110 (*thumb1_tablejump): Likewise.
10111 (prefetch): Likewise.
10112 (force_register_use): Likewise.
10113 (thumb_eh_return): Likewise.
10114 (load_tp_hard): Likewise.
10115 (load_tp_soft): Likewise.
10116 (tlscall): Likewise.
10117 (*arm_movtas_ze): Likewise.
10118 (*arm_rev): Likewise.
10119 (*arm_revsh): Likewise.
10120 (*arm_rev16): Likewise.
10121 * config/arm/thumb2.md
10122 (*thumb2_smaxsi3): Likewise.
10123 (*thumb2_sminsi3): Likewise.
10124 (*thumb32_umaxsi3): Likewise.
10125 (*thumb2_uminsi3): Likewise.
10126 (*thumb2_negdi2): Likewise.
10127 (*thumb2_abssi2): Likewise.
10128 (*thumb2_neg_abss): Likewise.
10129 (*thumb2_movsi_insn): Likewise.
10130 (tls_load_dot_plus_four): Likewise.
10131 (*thumb2_movhi_insn): Likewise.
10132 (*thumb2_mov_scc): Likewise.
10133 (*thumb2_mov_negs): Likewise.
10134 (*thumb2_mov_negs): Likewise.
10135 (*thumb2_mov_nots): Likewise.
10136 (*thumb2_mov_nots): Likewise.
10137 (*thumb2_movsicc_): Likewise.
10138 (*thumb2_movsfcc_soft_insn): Likewise.
10139 (*thumb2_indirect_jump): Likewise.
10140 (*thumb2_and_scc): Likewise.
10141 (*thumb2_ior_scc): Likewise.
10142 (*thumb2_ior_scc_strict_it): Likewise.
10143 (*thumb2_cond_move): Likewise.
10144 (*thumb2_cond_arith): Likewise.
10145 (*thumb2_cond_ari): Likewise.
10146 (*thumb2_cond_sub): Likewise.
10147 (*thumb2_negscc): Likewise.
10148 (*thumb2_movcond): Likewise.
10149 (thumb2_casesi_internal): Likewise.
10150 (thumb2_casesi_internal_pic): Likewise.
10151 (*thumb2_alusi3_short): Likewise.
10152 (*thumb2_mov<mode>_shortim): Likewise.
10153 (*thumb2_addsi_short): Likewise.
10154 (*thumb2_subsi_short): Likewise.
10155 (thumb2_addsi3_compare0): Likewise.
10156 (*thumb2_cbz): Likewise.
10157 (*thumb2_cbnz): Likewise.
10158 (*thumb2_one_cmplsi2_short): Likewise.
10159 (*thumb2_negsi2_short): Likewise.
10160 (*orsi_notsi_si): Likewise.
10161 * config/arm/arm1020e.md: Update with new attributes.
10162 * config/arm/arm1026ejs.md: Update with new attributes.
10163 * config/arm/arm1136jfs.md: Update with new attributes.
10164 * config/arm/arm926ejs.md: Update with new attributes.
10165 * config/arm/cortex-a15.md: Update with new attributes.
10166 * config/arm/cortex-a5.md: Update with new attributes.
10167 * config/arm/cortex-a53.md: Update with new attributes.
10168 * config/arm/cortex-a7.md: Update with new attributes.
10169 * config/arm/cortex-a8.md: Update with new attributes.
10170 * config/arm/cortex-a9.md: Update with new attributes.
10171 * config/arm/cortex-m4.md: Update with new attributes.
10172 * config/arm/cortex-r4.md: Update with new attributes.
10173 * config/arm/fa526.md: Update with new attributes.
10174 * config/arm/fa606te.md: Update with new attributes.
10175 * config/arm/fa626te.md: Update with new attributes.
10176 * config/arm/fa726te.md: Update with new attributes.
10178 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10180 * config/aarch64/aarch64-simd.md
10181 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
10182 <vwx> iterator to ensure correct register choice.
10183 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
10184 (aarch64_sqdmull_n<mode>): Likewise.
10185 (aarch64_sqdmull2_n<mode>_internal): Likewise.
10186 * config/aarch64/arm_neon.h
10187 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
10188 (vml<as><q>_n_<su>16): Likewise.
10189 (vml<as>l_high_lane<q>_<su>16): Likewise.
10190 (vml<as>l_high_n_<su>16): Likewise.
10191 (vml<as>l_lane<q>_<su>16): Likewise.
10192 (vml<as>l_n_<su>16): Likewise.
10193 (vmul<q>_lane<q>_<su>16): Likewise.
10194 (vmul<q>_n_<su>16): Likewise.
10195 (vmull_lane<q>_<su>16): Likewise.
10196 (vmull_n_<su>16): Likewise.
10197 (vmull_high_lane<q>_<su>16): Likewise.
10198 (vmull_high_n_<su>16): Likewise.
10199 (vqrdmulh<q>_n_s16): Likewise.
10201 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
10203 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
10204 have the correct lane parameter.
10206 2013-09-06 Richard Biener <rguenther@suse.de>
10208 * cfganal.c (control_dependences::~control_dependences):
10209 Properly free all of the vector.
10211 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
10214 * config/i386/i386.c (ix86_conditional_register_usage):
10215 Proper initialize extended SSE registers.
10217 2013-09-06 Jan Hubicka <jh@suse.cz>
10219 PR tree-optimization/58311
10220 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
10222 2013-09-06 Jan Hubicka <jh@suse.cz>
10224 * Makefile.in (tree-sra.o): Update dependencies.
10225 * tree-sra.c: Include ipa-utils.h
10226 (scan_function): Use recursive_call_p.
10227 (has_caller_p): New function.
10228 (cgraph_for_node_and_aliases): Count also callers of aliases.
10230 2013-09-06 Jan Hubicka <jh@suse.cz>
10232 PR middle-end/58094
10233 * cgraph.h (symtab_semantically_equivalent_p): Declare.
10234 * tree-tailcall.c: Include ipa-utils.h.
10235 (find_tail_calls): Use it.
10236 * ipa-pure-const.c (check_call): Likewise.
10237 * ipa-utils.c (recursive_call_p): New function.
10238 * ipa-utils.h (recursive_call_p): Dclare.
10239 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
10240 (symtab_semantically_equivalent_p): New function.
10241 * Makefile.in (tree-tailcall.o): Update dependencies.
10243 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
10245 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
10246 non-inlinable part.
10248 2013-09-06 Richard Biener <rguenther@suse.de>
10250 * lto-streamer.h (lto_global_var_decls): Remove.
10251 * Makefile.in (OBJS): Remove lto-symtab.o.
10252 (lto-symtab.o): Remove.
10253 (GTFILES): Remove lto-symtab.c
10254 * lto-symtab.c: Move to lto/
10256 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10258 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
10259 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
10260 (UNSPEC_FPINT_RINT): New constant definitions.
10261 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
10262 definition with 2 attributes.
10263 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
10264 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
10267 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10269 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
10270 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
10271 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
10272 attribute to "z196_alone".
10273 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
10276 2013-09-06 Richard Biener <rguenther@suse.de>
10278 * basic-block.h (class control_dependences): New.
10279 * tree-ssa-dce.c (control_dependence_map): Remove.
10281 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
10282 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
10283 find_pdom, find_control_dependence, find_all_control_dependences):
10285 (mark_control_dependent_edges_necessary,
10286 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
10287 tree_dce_done, perform_tree_ssa_dce): Adjust.
10288 * cfganal.c (set_control_dependence_map_bit,
10289 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
10290 find_all_control_dependences): Move from tree-ssa-dce.c and
10291 implement as methods of control_dependences class.
10292 (control_dependences::control_dependences): New.
10293 (control_dependences::~control_dependences): Likewise.
10294 (control_dependences::get_edges_dependent_on): Likewise.
10295 (control_dependences::get_edge): Likewise.
10297 2013-09-04 Jan Hubicka <jh@suse.cz>
10299 * tree.c (types_same_for_odr): Drop overactive check.
10300 * ipa-devirt.c (hash_type_name): Likewise.
10302 2013-09-04 Jan Hubicka <jh@suse.cz>
10304 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
10305 (analyze_functions): ... here.
10307 2013-09-04 Jan Hubicka <jh@suse.cz>
10309 PR middle-end/58201
10310 * cgraphunit.c (analyze_functions): Clear AUX fields
10311 after processing; initialize assembler name has.
10313 2013-09-05 Jeff Law <law@redhat.com>
10315 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
10316 from thread_around_empty_block. Record threading path into PATH.
10317 Recurse if threading through the initial block is successful.
10318 (thread_across_edge): Corresponding changes to slightly simplify.
10320 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10322 * config/aarch64/aarch64.md
10323 (type): Remove frecpe, frecps, frecpx.
10324 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
10325 fix to be a TARGET_SIMD instruction.
10326 (aarch64_frecps): Remove.
10327 * config/aarch64/aarch64-simd.md
10328 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
10329 (aarch64_frecps<mode>): Handle all float/vector of float modes.
10331 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10332 Sofiane Naci <sofiane.naci@arm.com>
10334 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
10335 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
10336 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
10337 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
10338 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
10339 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
10340 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
10341 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
10342 into "clz, "rbit". Rename "shift" to "shift_imm".
10343 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
10344 changes. Update for attribute changes all occurrences of arlo_* and
10346 * config/arm/arm-fixed.md: Update for attribute changes
10347 all occurrences of arlo_* types.
10348 * config/arm/thumb2.md: Update for attribute changes all occurrences
10350 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
10351 (cortexa7_older_only): Likewise.
10352 (cortexa7_younger): Likewise.
10353 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
10354 (1020alu_shift_op): Likewise.
10355 (1020alu_shift_reg_op): Likewise.
10356 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
10357 (alu_shift_op): Likewise.
10358 (alu_shift_reg_op): Likewise.
10359 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
10360 (11_alu_shift_op): Likewise.
10361 (11_alu_shift_reg_op): Likewise.
10362 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
10363 (9_alu_shift_reg_op): Likewise.
10364 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
10366 (cortex_a15_alu_shift): Likewise.
10367 (cortex_a15_alu_shift_reg): Likewise.
10368 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
10370 (cortex_a5_alu_shift): Likewise.
10371 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
10373 (cortex_a53_alu_shift): Likewise.
10374 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
10376 (cortex_a7_alu_reg): Likewise.
10377 (cortex_a7_alu_shift): Likewise.
10378 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
10380 (cortex_a8_alu_shift): Likewise.
10381 (cortex_a8_alu_shift_reg): Likewise.
10382 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
10383 (cortex_a9_dp_shift): Likewise.
10384 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
10386 * config/arm/cortex-r4.md
10387 (cortex_r4_alu): Update for attribute changes.
10388 (cortex_r4_mov): Likewise.
10389 (cortex_r4_alu_shift_reg): Likewise.
10390 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
10391 (526_alu_shift_op): Likewise.
10392 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
10393 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
10394 (626te_alu_shift_op): Likewise.
10395 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
10396 (726te_alu_shift_op): Likewise.
10397 (726te_alu_shift_reg_op): Likewise.
10398 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
10399 (mp626_alu_shift_op): Likewise.
10400 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
10401 (pj4_alu_conds): Likewise.
10402 (pj4_shift): Likewise.
10403 (pj4_shift_conds): Likewise.
10404 (pj4_alu_shift): Likewise.
10405 (pj4_alu_shift_conds): Likewise.
10406 * config/aarch64/aarch64.md: Update for attribute change
10407 all occurrences of arlo_* and shift* types.
10409 2013-09-05 Mike Stump <mikestump@comcast.net>
10411 * tree.h: Move documentation for tree_function_decl to tree-core.h
10412 with the declaration.
10414 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
10417 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
10418 looking for widest mode.
10420 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
10422 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
10424 2013-09-05 Richard Biener <rguenther@suse.de>
10426 PR tree-optimization/58137
10427 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
10428 Do not create vectors of pointers.
10429 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
10430 types for the components of the vector initializer.
10431 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
10432 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
10434 2013-09-05 Martin Jambor <mjambor@suse.cz>
10436 * ipa-prop.c (remove_described_reference): Accept missing references,
10437 return false if that hppens, otherwise return true.
10438 (cgraph_node_for_jfunc): New function.
10439 (try_decrement_rdesc_refcount): Likewise.
10440 (try_make_edge_direct_simple_call): Use them.
10441 (ipa_edge_removal_hook): Remove references from rdescs.
10442 (ipa_edge_duplication_hook): Clone rdescs and their references
10443 when the new edge has the same caller as the old one.
10444 * cgraph.c (cgraph_resolve_speculation): Remove speculative
10445 reference before removing any edges.
10447 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
10449 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
10451 * thumb2.md (thumb2_storewb_parisi): New pattern.
10453 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
10455 * config/aarch64/aarch64-option-extensions.def: Add
10456 AARCH64_OPT_EXTENSION of 'crc'.
10457 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
10458 (AARCH64_ISA_CRC): Ditto.
10459 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
10460 description of the CRC extension.
10462 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
10464 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
10465 * config/rs6000/linux.h: Ditto.
10466 * alpha/linux.h: Ditto.
10467 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
10468 no_c99_libc_has_function.
10469 * config/c6x/uclinux-elf.h: Ditto.
10470 * config/lm32/uclinux-elf.h: Ditto.
10471 * config/m68k/uclinux.h: Ditto.
10472 * config/moxie/uclinux.h: Ditto.
10473 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
10474 (crisv32-*-linux*, cris-*-linux*): Ditto.
10475 * config/bfin/bfin.c: Include "tm_p.h".
10477 2013-09-05 Richard Biener <rguenther@suse.de>
10479 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
10480 check for a definition without a basic-block.
10482 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10483 Sofiane Naci <sofiane.naci@arm.com>
10485 * config/aarch64/aarch64.md
10486 (*movti_aarch64): Rename r_2_f and f_2_r.
10487 (*movsf_aarch64): Likewise.
10488 (*movdf_aarch64): Likewise.
10489 (*movtf_aarch64): Likewise.
10490 (aarch64_movdi_<mode>low): Likewise.
10491 (aarch64_movdi_<mode>high): Likewise.
10492 (aarch64_mov<mode>high_di): Likewise.
10493 (aarch64_mov<mode>low_di): Likewise.
10494 (aarch64_movtilow_tilow): Likewise.
10495 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
10496 values to config/arm/types.md
10497 (attribute "conds"): Update for attribute change.
10498 (anddi3_insn): Likewise.
10499 (iordi3_insn): Likewise.
10500 (xordi3_insn): Likewise.
10501 (one_cmpldi2): Likewise.
10502 * config/arm/types.md (type): Add Neon types.
10503 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
10504 use "type" attribute.
10505 (movmisalign<mode>_neon_store): Likewise.
10506 (movmisalign<mode>_neon_load): Likewise.
10507 (vec_set<mode>_internal): Likewise.
10508 (vec_setv2di_internal): Likewise.
10509 (vec_extract<mode>): Likewise.
10510 (vec_extractv2di): Likewise.
10511 (add<mode>3_neon): Likewise.
10512 (adddi3_neon): Likewise.
10513 (sub<mode>3_neon): Likewise.
10514 (subdi3_neon): Likewise.
10515 (mul<mode>3_neon): Likewise.
10516 (mul<mode>3add<mode>_neon): Likewise.
10517 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
10518 (fma<VCVTF:mode>4)): Likewise.
10519 (fma<VCVTF:mode>4_intrinsic): Likewise.
10520 (fmsub<VCVTF:mode>4)): Likewise.
10521 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
10522 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
10523 (ior<mode>3): Likewise.
10524 (and<mode>3): Likewise.
10525 (anddi3_neon): Likewise.
10526 (orn<mode>3_neon): Likewise.
10527 (orndi3_neon): Likewise.
10528 (bic<mode>3_neon): Likewise.
10529 (bicdi3_neon): Likewise.
10530 (xor<mode>3): Likewise.
10531 (one_cmpl<mode>2): Likewise.
10532 (abs<mode>2): Likewise.
10533 (neg<mode>2): Likewise.
10534 (umin<mode>3_neon): Likewise.
10535 (umax<mode>3_neon): Likewise.
10536 (smin<mode>3_neon): Likewise.
10537 (smax<mode>3_neon): Likewise.
10538 (vashl<mode>3): Likewise.
10539 (vashr<mode>3_imm): Likewise.
10540 (vlshr<mode>3_imm): Likewise.
10541 (ashl<mode>3_signed): Likewise.
10542 (ashl<mode>3_unsigned): Likewise.
10543 (neon_load_count): Likewise.
10544 (ashldi3_neon_noclobber): Likewise.
10545 (signed_shift_di3_neon): Likewise.
10546 (unsigned_shift_di3_neon): Likewise.
10547 (ashrdi3_neon_imm_noclobber): Likewise.
10548 (lshrdi3_neon_imm_noclobber): Likewise.
10549 (widen_ssum<mode>3): Likewise.
10550 (widen_usum<mode>3): Likewise.
10551 (quad_halves_<code>v4si): Likewise.
10552 (quad_halves_<code>v4sf): Likewise.
10553 (quad_halves_<code>v8hi): Likewise.
10554 (quad_halves_<code>v16qi): Likewise.
10555 (reduc_splus_v2di): Likewise.
10556 (neon_vpadd_internal<mode>): Likewise.
10557 (neon_vpsmin<mode>): Likewise.
10558 (neon_vpsmax<mode>): Likewise.
10559 (neon_vpumin<mode>): Likewise.
10560 (neon_vpumax<mode>): Likewise.
10561 (ss_add<mode>_neon): Likewise.
10562 (us_add<mode>_neon): Likewise.
10563 (ss_sub<mode>_neon): Likewise.
10564 (us_sub<mode>_neon): Likewise.
10565 (neon_vadd<mode>_unspec): Likewise.
10566 (neon_vaddl<mode>): Likewise.
10567 (neon_vaddw<mode>): Likewise.
10568 (neon_vhadd<mode>): Likewise.
10569 (neon_vqadd<mode>): Likewise.
10570 (neon_vaddhn<mode>): Likewise.
10571 (neon_vmul<mode>): Likewise.
10572 (neon_vmla<mode>): Likewise.
10573 (neon_vmlal<mode>): Likewise.
10574 (neon_vmls<mode>): Likewise.
10575 (neon_vmlsl<mode>): Likewise.
10576 (neon_vqdmulh<mode>): Likewise.
10577 (neon_vqdmlal<mode>): Likewise.
10578 (neon_vqdmlsl<mode>): Likewise.
10579 (neon_vmull<mode>): Likewise.
10580 (neon_vqdmull<mode>): Likewise.
10581 (neon_vsub<mode>_unspec): Likewise.
10582 (neon_vsubl<mode>): Likewise.
10583 (neon_vsubw<mode>): Likewise.
10584 (neon_vqsub<mode>): Likewise.
10585 (neon_vhsub<mode>): Likewise.
10586 (neon_vsubhn<mode>): Likewise.
10587 (neon_vceq<mode>): Likewise.
10588 (neon_vcge<mode>): Likewise.
10589 (neon_vcgeu<mode>): Likewise.
10590 (neon_vcgt<mode>): Likewise.
10591 (neon_vcgtu<mode>): Likewise.
10592 (neon_vcle<mode>): Likewise.
10593 (neon_vclt<mode>): Likewise.
10594 (neon_vcage<mode>): Likewise.
10595 (neon_vcagt<mode>): Likewise.
10596 (neon_vtst<mode>): Likewise.
10597 (neon_vabd<mode>): Likewise.
10598 (neon_vabdl<mode>): Likewise.
10599 (neon_vaba<mode>): Likewise.
10600 (neon_vabal<mode>): Likewise.
10601 (neon_vmax<mode>): Likewise.
10602 (neon_vmin<mode>): Likewise.
10603 (neon_vpaddl<mode>): Likewise.
10604 (neon_vpadal<mode>): Likewise.
10605 (neon_vpmax<mode>): Likewise.
10606 (neon_vpmin<mode>): Likewise.
10607 (neon_vrecps<mode>): Likewise.
10608 (neon_vrsqrts<mode>): Likewise.
10609 (neon_vqabs<mode>): Likewise.
10610 (neon_vqneg<mode>): Likewise.
10611 (neon_vcls<mode>): Likewise.
10612 (clz<mode>2): Likewise.
10613 (popcount<mode>2): Likewise.
10614 (neon_vrecpe): Likewise.
10615 (neon_vrsqrte): Likewise.
10616 (neon_vget_lane<mode>_sext_internal): Likewise.
10617 (neon_vget_lane<mode>_zext_internal): Likewise.
10618 (neon_vdup_n<mode>): Likewise.
10619 (neon_vdup_nv2di): Likewise.
10620 (neon_vdpu_lane<mode>_internal): Likewise.
10621 (neon_vswp<mode>): Likewise.
10622 (float<mode><V_cvtto>2): Likewise.
10623 (floatuns<mode><V_cvtto>2): Likewise.
10624 (fix_trunc<mode><V_cvtto>)2): Likewise
10625 (fixuns_trunc<mode><V_cvtto)2): Likewise.
10626 (neon_vcvt<mode>): Likewise.
10627 (neon_vcvtv4sfv4hf): Likewise.
10628 (neon_vcvtv4hfv4sf): Likewise.
10629 (neon_vcvt_n<mode>): Likewise.
10630 (neon_vmovn<mode>): Likewise.
10631 (neon_vqmovn<mode>): Likewise.
10632 (neon_vqmovun<mode>): Likewise.
10633 (neon_vmovl<mode>): Likewise.
10634 (neon_vmul_lane<mode>): Likewise.
10635 (neon_vmull_lane<mode>): Likewise.
10636 (neon_vqdmull_lane<mode>): Likewise.
10637 (neon_vqdmulh_lane<mode>): Likewise.
10638 (neon_vmla_lane<mode>): Likewise.
10639 (neon_vmlal_lane<mode>): Likewise.
10640 (neon_vqdmlal_lane<mode>): Likewise.
10641 (neon_vmls_lane<mode>): Likewise.
10642 (neon_vmlsl_lane<mode>): Likewise.
10643 (neon_vqdmlsl_lane<mode>): Likewise.
10644 (neon_vext<mode>): Likewise.
10645 (neon_vrev64<mode>): Likewise.
10646 (neon_vrev32<mode>): Likewise.
10647 (neon_vrev16<mode>): Likewise.
10648 (neon_vbsl<mode>_internal): Likewise.
10649 (neon_vshl<mode>): Likewise.
10650 (neon_vqshl<mode>): Likewise.
10651 (neon_vshr_n<mode>): Likewise.
10652 (neon_vshrn_n<mode>): Likewise.
10653 (neon_vqshrn_n<mode>): Likewise.
10654 (neon_vqshrun_n<mode>): Likewise.
10655 (neon_vshl_n<mode>): Likewise.
10656 (neon_vqshl_n<mode>): Likewise.
10657 (neon_vqshlu_n<mode>): Likewise.
10658 (neon_vshll_n<mode>): Likewise.
10659 (neon_vsra_n<mode>): Likewise.
10660 (neon_vsri_n<mode>): Likewise.
10661 (neon_vsli_n<mode>): Likewise.
10662 (neon_vtbl1v8qi): Likewise.
10663 (neon_vtbl2v8qi): Likewise.
10664 (neon_vtbl3v8qi): Likewise.
10665 (neon_vtbl4v8qi): Likewise.
10666 (neon_vtbx1v8qi): Likewise.
10667 (neon_vtbx2v8qi): Likewise.
10668 (neon_vtbx3v8qi): Likewise.
10669 (neon_vtbx4v8qi): Likewise.
10670 (neon_vtrn<mode>_internal): Likewise.
10671 (neon_vzip<mode>_internal): Likewise.
10672 (neon_vuzp<mode>_internal): Likewise.
10673 (neon_vld1<mode>): Likewise.
10674 (neon_vld1_lane<mode>): Likewise.
10675 (neon_vld1_dup<mode>): Likewise.
10676 (neon_vld1_dupv2di): Likewise.
10677 (neon_vst1<mode>): Likewise.
10678 (neon_vst1_lane<mode>): Likewise.
10679 (neon_vld2<mode>): Likewise.
10680 (neon_vld2_lane<mode>): Likewise.
10681 (neon_vld2_dup<mode>): Likewise.
10682 (neon_vst2<mode>): Likewise.
10683 (neon_vst2_lane<mode>): Likewise.
10684 (neon_vld3<mode>): Likewise.
10685 (neon_vld3qa<mode>): Likewise.
10686 (neon_vld3qb<mode>): Likewise.
10687 (neon_vld3_lane<mode>): Likewise.
10688 (neon_vld3_dup<mode>): Likewise.
10689 (neon_vst3<mode>): Likewise.
10690 (neon_vst3qa<mode>): Likewise.
10691 (neon_vst3qb<mode>): Likewise.
10692 (neon_vst3_lane<mode>): Likewise.
10693 (neon_vld4<mode>): Likewise.
10694 (neon_vld4qa<mode>): Likewise.
10695 (neon_vld4qb<mode>): Likewise.
10696 (neon_vld4_lane<mode>): Likewise.
10697 (neon_vld4_dup<mode>): Likewise.
10698 (neon_vst4<mode>): Likewise.
10699 (neon_vst4qa<mode>): Likewise.
10700 (neon_vst4qb<mode>): Likewise.
10701 (neon_vst4_lane<mode>): Likewise.
10702 (neon_vec_unpack<US>_lo_<mode>): Likewise.
10703 (neon_vec_unpack<US>_hi_<mode>): Likewise.
10704 (neon_vec_<US>mult_lo_<mode>): Likewise.
10705 (neon_vec_<US>mult_hi_<mode>): Likewise.
10706 (neon_vec_<US>shiftl_<mode>): Likewise.
10707 (neon_unpack<US>_<mode>): Likewise.
10708 (neon_vec_<US>mult_<mode>): Likewise.
10709 (vec_pack_trunc_<mode>): Likewise.
10710 (neon_vec_pack_trunk_<mode>): Likewise.
10711 (neon_vabd<mode>_2): Likewise.
10712 (neon_vabd<mode>_3): Likewise.
10713 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
10714 (thumb2_movsi_vfp): Likewise.
10715 (movdi_vfp): Likewise.
10716 (movdi_vfp_cortexa8): Likewise.
10717 (movhf_vfp_neon): Likewise.
10718 (movhf_vfp): Likewiwse.
10719 (movsf_vfp): Likewiwse.
10720 (thumb2_movsf_vfp): Likewiwse.
10721 (movdf_vfp): Likewise.
10722 (thumb2_movdf_vfp): Likewise.
10723 (movsfcc_vfp): Likewise.
10724 (thumb2_movsfcc_vfp): Likewise.
10725 (movdfcc_vfp): Likewise.
10726 (thumb2_movdfcc_vfp): Likewise.
10727 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
10728 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
10729 (v10_v2c): Likewise.
10730 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
10732 (cortex_a15_neon_int_2): Likewise.
10733 (cortex_a15_neon_int_3): Likewise.
10734 (cortex_a15_neon_int_4): Likewise.
10735 (cortex_a15_neon_int_5): Likewise.
10736 (cortex_a15_neon_vqneg_vqabs): Likewise.
10737 (cortex_a15_neon_vmov): Likewise.
10738 (cortex_a15_neon_vaba): Likewise.
10739 (cortex_a15_neon_vaba_qqq): Likewise.
10740 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10741 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10742 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
10743 scalar_64_32_long_scalar): Likewise.
10744 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10745 (cortex_a15_neon_mla_qqq_8_16): Likewise.
10746 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10747 lotype_qdd_64_32_long): Likewise.
10748 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10749 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10750 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
10751 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10752 (cortex_a15_neon_shift_1): Likewise.
10753 (cortex_a15_neon_shift_2): Likewise.
10754 (cortex_a15_neon_shift_3): Likewise.
10755 (cortex_a15_neon_vshl_ddd): Likewise.
10756 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10757 (cortex_a15_neon_vsra_vrsra): Likewise.
10758 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
10759 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
10760 (cortex_a15_neon_fp_vmul_ddd): Likewise.
10761 (cortex_a15_neon_fp_vmul_qqd): Likewise.
10762 (cortex_a15_neon_fp_vmla_ddd): Likewise.
10763 (cortex_a15_neon_fp_vmla_qqq): Likewise.
10764 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
10765 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
10766 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10767 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10768 (cortex_a15_neon_bp_simple): Likewise.
10769 (cortex_a15_neon_bp_2cycle): Likewise.
10770 (cortex_a15_neon_bp_3cycle): Likewise.
10771 (cortex_a15_neon_vld1_1_2_regs): Likewise.
10772 (cortex_a15_neon_vld1_3_4_regs): Likewise.
10773 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10774 (cortex_a15_neon_vld2_4_regs): Likewise.
10775 (cortex_a15_neon_vld3_vld4): Likewise.
10776 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10777 (cortex_a15_neon_vst1_3_4_regs): Likewise.
10778 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
10779 (cortex_a15_neon_vst3_vst4): Likewise.
10780 (cortex_a15_neon_vld1_vld2_lane): Likewise.
10781 (cortex_a15_neon_vld3_vld4_lane" 10
10782 (cortex_a15_neon_vst1_vst2_lane): Likewise.
10783 (cortex_a15_neon_vst3_vst4_lane): Likewise.
10784 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
10785 (cortex_a15_neon_ldm_2): Likewise.0
10786 (cortex_a15_neon_stm_2): Likewise.
10787 (cortex_a15_neon_mcr): Likewise.
10788 (cortex_a15_neon_mcr_2_mcrr): Likewise.
10789 (cortex_a15_neon_mrc): Likewise.
10790 (cortex_a15_neon_mrrc): Likewise.
10791 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10793 (cortex_a15_alu_shift): Likewise.
10794 (cortex_a15_alu_shift_reg): Likewise.
10795 (cortex_a15_mult32): Likewise.
10796 (cortex_a15_mult64): Likewise.
10797 (cortex_a15_block): Likewise.
10798 (cortex_a15_branch): Likewise.
10799 (cortex_a15_load1): Likewise.
10800 (cortex_a15_load3): Likewise.
10801 (cortex_a15_store1): Likewise.
10802 (cortex_a15_store3): Likewise.
10803 (cortex_a15_call): Likewise.
10804 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
10805 (cortex_a5_f2r): Likewise.
10806 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
10808 (cortex_a53_f2r): Likewise.
10809 * config/arm/cortex-a7.md
10810 (cortex_a7_branch): Update for attribute change.
10811 (cortex_a7_call): Likewise.
10812 (cortex_a7_alu_imm): Likewise.
10813 (cortex_a7_alu_reg): Likewise.
10814 (cortex_a7_alu_shift): Likewise.
10815 (cortex_a7_mul): Likewise.
10816 (cortex_a7_load1): Likewise.
10817 (cortex_a7_store1): Likewise.
10818 (cortex_a7_load2): Likewise.
10819 (cortex_a7_store2): Likewise.
10820 (cortex_a7_load3): Likewise.
10821 (cortex_a7_store3): Likewise.
10822 (cortex_a7_load4): Likewise.
10823 (cortex_a7_store4): Likewise.
10824 (cortex_a7_fpalu): Likewise.
10825 (cortex_a7_fconst): Likewise.
10826 (cortex_a7_fpmuls): Likewise.
10827 (cortex_a7_neon_mul): Likewise.
10828 (cortex_a7_fpmacs): Likewise.
10829 (cortex_a7_neon_mla: Likewise.
10830 (cortex_a7_fpmuld: Likewise.
10831 (cortex_a7_fpmacd: Likewise.
10832 (cortex_a7_fpfmad: Likewise.
10833 (cortex_a7_fdivs: Likewise.
10834 (cortex_a7_fdivd: Likewise.
10835 (cortex_a7_r2f: Likewise.
10836 (cortex_a7_f2r: Likewise.
10837 (cortex_a7_f_flags: Likewise.
10838 (cortex_a7_f_loads: Likewise.
10839 (cortex_a7_f_loadd: Likewise.
10840 (cortex_a7_f_stores: Likewise.
10841 (cortex_a7_f_stored: Likewise.
10842 (cortex_a7_neon): Likewise.
10843 * config/arm/cortex-a8-neon.md
10844 (cortex_a8_neon_mrc): Update for attribute change.
10845 (cortex_a8_neon_mrrc): Likewise.
10846 (cortex_a8_neon_int_1): Likewise.
10847 (cortex_a8_neon_int_2): Likewise.
10848 (cortex_a8_neon_int_3): Likewise.
10849 (cortex_a8_neon_int_4): Likewise.
10850 (cortex_a8_neon_int_5): Likewise.
10851 (cortex_a8_neon_vqneg_vqabs): Likewise.
10852 (cortex_a8_neon_vmov): Likewise.
10853 (cortex_a8_neon_vaba): Likewise.
10854 (cortex_a8_neon_vaba_qqq): Likewise.
10855 (cortex_a8_neon_vsma): Likewise.
10856 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10857 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10858 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10859 long_scalar): Likewise.
10860 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10861 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10862 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10863 long_scalar_qdd_64_32_long): Likewise.
10864 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10865 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10866 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10867 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10868 (cortex_a8_neon_shift_1): Likewise.
10869 (cortex_a8_neon_shift_2): Likewise.
10870 (cortex_a8_neon_shift_3): Likewise.
10871 (cortex_a8_neon_vshl_ddd): Likewise.
10872 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10873 (cortex_a8_neon_vsra_vrsra): Likewise.
10874 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10875 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10876 (cortex_a8_neon_fp_vsum): Likewise.
10877 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10878 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10879 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10880 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10881 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10882 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10883 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10884 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10885 (cortex_a8_neon_bp_simple): Likewise.
10886 (cortex_a8_neon_bp_2cycle): Likewise.
10887 (cortex_a8_neon_bp_3cycle): Likewise.
10888 (cortex_a8_neon_ldr): Likewise.
10889 (cortex_a8_neon_str): Likewise.
10890 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10891 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10892 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10893 (cortex_a8_neon_vld2_4_regs): Likewise.
10894 (cortex_a8_neon_vld3_vld4): Likewise.
10895 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10896 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10897 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10898 (cortex_a8_neon_vst3_vst4): Likewise.
10899 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10900 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10901 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10902 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10903 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10904 (cortex_a8_neon_mcr): Likewise.
10905 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10906 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10907 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10909 (ca9_neon_mrrc): Likewise.
10910 (cortex_a9_neon_int_1): Likewise.
10911 (cortex_a9_neon_int_2): Likewise.
10912 (cortex_a9_neon_int_3): Likewise.
10913 (cortex_a9_neon_int_4): Likewise.
10914 (cortex_a9_neon_int_5): Likewise.
10915 (cortex_a9_neon_vqneg_vqabs): Likewise.
10916 (cortex_a9_neon_vmov): Likewise.
10917 (cortex_a9_neon_vaba): Likewise.
10918 (cortex_a9_neon_vaba_qqq): Likewise.
10919 (cortex_a9_neon_vsma): Likewise.
10920 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10921 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10922 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10923 long_scalar): Likewise.
10924 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10925 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10926 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10927 long_scalar_qdd_64_32_long): Likewise.
10928 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10929 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10930 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10931 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10932 (cortex_a9_neon_shift_1): Likewise.
10933 (cortex_a9_neon_shift_2): Likewise.
10934 (cortex_a9_neon_shift_3): Likewise.
10935 (cortex_a9_neon_vshl_ddd): Likewise.
10936 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10937 (cortex_a9_neon_vsra_vrsra): Likewise.
10938 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10939 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10940 (cortex_a9_neon_fp_vsum): Likewise.
10941 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10942 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10943 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10944 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10945 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10946 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10947 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10948 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10949 (cortex_a9_neon_bp_simple): Likewise.
10950 (cortex_a9_neon_bp_2cycle): Likewise.
10951 (cortex_a9_neon_bp_3cycle): Likewise.
10952 (cortex_a9_neon_ldr): Likewise.
10953 (cortex_a9_neon_str): Likewise.
10954 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10955 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10956 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10957 (cortex_a9_neon_vld2_4_regs): Likewise.
10958 (cortex_a9_neon_vld3_vld4): Likewise.
10959 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10960 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10961 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10962 (cortex_a9_neon_vst3_vst4): Likewise.
10963 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10964 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10965 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10966 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10967 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10968 (cortex_a9_neon_mcr): Likewise.
10969 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10970 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10971 (cortex_a9_fps): Likewise.
10972 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10974 (cortex_m4_fmuls): Likewise.
10975 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10977 (cortex_r4_mrc): Likewise.
10978 * config/arm/iterators.md: Update comment referring to neon_type.
10979 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10980 (iwmmxt_movsi_insn): Likewise.
10981 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10983 (pj4_core_to_vfp): Likewise.
10984 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10986 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10987 (vfp_fstore): Likewise.
10988 * doc/md.texi: Change references to neon_type to refer to type.
10990 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10992 * tree.h (DECL_BUILT_IN): Fix typo in comment.
10994 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
10996 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
10997 lglobl if not weak.
10999 2013-09-04 Easwaran Raman <eraman@google.com>
11001 PR middle-end/57370
11002 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
11003 (build_and_add_sum): Use it.
11004 (appears_later_in_bb): Simplify code.
11006 2013-09-04 Teresa Johnson <tejohnson@google.com>
11008 * dumpfile.c (dump_finish): Don't close stderr/stdout.
11010 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
11012 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
11014 2013-09-04 Jan Hubicka <jh@suse.cz>
11016 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
11017 * ipa-devirt.c: Include diganostic.h
11018 (odr_type_d): Add types and types_set.
11019 (hash_type_name): Work for types with vtables during LTO.
11020 (odr_hasher::remove): Fix comment; destroy types_set.
11021 (add_type_duplicate): New function,
11022 (get_odr_type): Use it.
11023 (dump_type_inheritance_graph): Dump type duplicates.
11024 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
11026 * tree.c (types_same_for_odr): Give exact answers on types with
11029 2013-09-04 Dodji Seketeli <dodji@redhat.com>
11031 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
11032 explaining their differences.
11034 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
11036 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
11038 2013-09-03 Jeff Law <law@redhat.com>
11040 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
11041 when not threading through a joiner block. Pass joiner/no joiner
11042 state to register_jump_thread.
11043 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
11044 state from argument rather than implying on path length.
11045 Dump the entire jump thread path into debugging dump.
11046 * tree-flow.h (register_jump_thread): Update prototype.
11048 2013-08-29 Xinliang David Li <davidxl@google.com>
11050 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
11051 Remove a trivial gcc_assert.
11053 2013-08-29 Xinliang David Li <davidxl@google.com>
11055 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
11056 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
11057 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
11058 Delay base decl alignment adjustment.
11059 * tree-vectorizer.c (vect_destroy_datarefs): New function.
11060 * tree-vectorizer.h: New data structure.
11061 (set_dr_misalignment): New function.
11062 (dr_misalignment): Ditto.
11063 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
11064 (vectorizable_load): Ditto.
11065 (ensure_base_align): New function.
11066 (vectorize_loops): Add dbg_cnt support.
11067 (execute_vect_slp): Ditto.
11068 * dbgcnt.def: New debug counter.
11069 * Makefile: New dependency.
11071 2013-09-03 Meador Inge <meadori@codesourcery.com>
11075 2013-08-30 Meador Inge <meadori@codesourcery.com>
11077 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
11079 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
11081 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
11082 function descriptor.
11084 2013-09-03 Richard Biener <rguenther@suse.de>
11086 * tree-affine.c (add_elt_to_tree): Fix association issue,
11087 avoid useless converts and make sure to always return a
11088 properly typed result.
11090 2013-09-03 Richard Biener <rguenther@suse.de>
11092 PR middle-end/57656
11093 * fold-const.c (negate_expr_p): Fix division case.
11094 (negate_expr): Likewise.
11096 2013-09-03 Richard Biener <rguenther@suse.de>
11099 * tree-streamer-out.c: Include tm.h.
11100 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
11102 2013-09-03 Jan Hubicka <jh@suse.cz>
11104 * tree-profile.c (tree_profiling): Cleanup CFG when done.
11106 2013-09-03 Alan Modra <amodra@gmail.com>
11108 * config.gcc (powerpc*-*-linux*): Add support for little-endian
11109 multilibs to big-endian target and vice versa.
11110 * config/rs6000/t-linux64: Use := assignment on all vars.
11111 (MULTILIB_EXTRA_OPTS): Remove fPIC.
11112 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
11113 * config/rs6000/t-linux64le: New file.
11114 * config/rs6000/t-linux64bele: New file.
11115 * config/rs6000/t-linux64lebe: New file.
11117 2013-09-02 Jan Hubicka <jh@suse.cz>
11119 * ipa-inline-transform.c (inline_transform): Do not
11120 optimize_inline_calls when not optimizing.
11122 2013-09-02 Jan Hubicka <jh@suse.cz>
11124 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
11125 duplicated nodes for assembler names.
11126 * symtab.c (symtab_unregister_node): Do not attempt to unlink
11127 hard registers from assembler name hash.
11129 2013-09-02 Jan Hubicka <jh@suse.cz>
11131 * ipa-split.c (execute_split_functions): Split externally visible
11132 functions called once.
11134 2013-09-02 Martin Jambor <mjambor@suse.cz>
11137 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
11138 linked list. When finding the correct duplicate, also consider also
11139 the caller in additon to its inlined_to node.
11141 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
11143 * config/aarch64/aarch64-simd-builtins.def
11144 (dup_lane_scalar): Remove.
11145 * config/aarch64/aarch64-simd.md
11146 (aarch64_simd_dup): Add 'w->w' alternative.
11147 (aarch64_dup_lane<mode>): Allow for VALL.
11148 (aarch64_dup_lane_scalar<mode>): Remove.
11149 (aarch64_dup_lane_<vswap_width_name><mode>): New.
11150 (aarch64_get_lane_signed<mode>): Add w->w altenative.
11151 (aarch64_get_lane_unsigned<mode>): Likewise.
11152 (aarch64_get_lane<mode>): Likewise.
11153 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
11154 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
11155 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
11156 (VCON): Change container of V2SF.
11157 (vswap_width_name): Likewise.
11158 * config/aarch64/arm_neon.h
11159 (__aarch64_vdup_lane_any): New.
11160 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
11161 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
11162 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
11164 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
11166 PR middle-end/56382
11167 * expr.c (emit_move_complex): Do not move complex FP values as parts if
11168 the source or the destination is a single hard register.
11170 2013-09-02 Richard Biener <rguenther@suse.de>
11172 PR middle-end/57511
11173 * tree-scalar-evolution.c (instantiate_scev_name): Allow
11176 2013-09-02 Richard Biener <rguenther@suse.de>
11178 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
11179 arithmetic to sizetype.
11181 2013-09-02 Bin Cheng <bin.cheng@arm.com>
11183 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
11184 Find auto-increment use both before and after candidate.
11186 2013-09-02 Marek Polacek <polacek@redhat.com>
11188 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
11190 2013-09-01 Jan Hubicka <jh@suse.cz>
11192 * Makefile.in: Add ipa-profile.o
11193 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
11194 * cgraph.c (struct cgraph_propagate_frequency_data,
11195 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
11196 ipa-profile.c; replace cgraph_ by ipa_ prefix.
11197 * cgraph.h (cgraph_propagate_frequency): Remove.
11198 * ipa-inline-analysis.c: Include ipa-utils.h;
11199 drop duplicated cfgloop.h.
11200 (inline_update_callee_summaries): Update.
11201 * ipa-profile.c: New file.
11202 * ipa-utils.h (ipa_propagate_frequency): Declare.
11203 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
11204 data-streamer.h, value-prof.h.
11205 (symtab_remove_unreachable_nodes): Update profile.
11206 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
11207 account_time_size, cmp_counts, dump_histogram,
11208 ipa_profile_generate_summary, ipa_profile_write_summary,
11209 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
11210 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
11211 Move to ipa-profile.c.
11213 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
11215 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
11217 2013-09-01 Jan Hubicka <jh@suse.cz>
11219 * common.opt (fdevirtualize-speculatively): New function.
11220 * invoke.texi (fdevirtualize-speculatively): Document.
11221 * ipa-devirt.c: Include ipa-inline.h
11222 (likely_target_p): New function.
11223 (ipa_devirt): New function.
11224 (gate_ipa_devirt): New function.
11225 (pass_data_ipa_devirt): New static var.
11226 (pass_ipa_devirt): Likewise.
11227 (make_pass_ipa_devirt): New function.
11228 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
11229 (common_handle_option): Disable devirtualization when
11230 value range profiling is available.
11231 * passes.def (pass_ipa_devirt): Add.
11232 * timever.def (TV_IPA_DEVIRT): New timevar.
11233 * tree-pass.h (make_pass_ipa_devirt):
11235 2013-09-01 Iain Sandoe <iain@codesourcery.com>
11237 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
11238 include sanitize(undefined).
11240 2013-08-31 Diego Novillo <dnovillo@google.com>
11242 * Makefile.in (TREE_CORE_H): Define.
11244 (GTFILES): Add tree-core.h.
11245 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
11247 * tree-core.h: New file.
11248 Move all data structures, enum, typedefs, global
11249 declarations and constants from ...
11250 * tree.h: ... here.
11252 2013-08-31 Jan Hubicka <jh@suse.cz>
11254 * bulitins.c (expand_builtin): Do not early exit for gcov
11255 instrumented functions.
11257 2013-08-31 Marek Polacek <polacek@redhat.com>
11259 * ubsan.c: Include tm_p.h.
11261 2013-08-31 Jan Hubicka <jh@suse.cz>
11263 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
11266 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
11267 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
11268 * tree-cfg.c (verify_gimple_label): ... here.
11269 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
11270 (ipa_merge_profiles): New function.
11271 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
11272 (lto_input_function_body): Likewise.
11273 * ipa-utils.h (ipa_merge_profiles): Declare.
11274 * lto-streamer.h (lto_input_function_body): Update prototype.
11275 (emit_label_in_global_context_p): Remove.
11276 * lto-symtab.c: Include ipa-utils.h
11277 (lto_cgraph_replace_node): Use ipa_merge_profiles.
11279 2013-08-31 Jan Hubicka <jh@suse.cz>
11281 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
11283 2013-08-31 Jan Hubicka <jh@suse.cz>
11285 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
11287 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
11289 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
11290 "cmp" RTX before signed_comparison_operator check to account
11291 for "code" changes.
11293 2013-08-30 Jan Hubicka <jh@suse.cz>
11295 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
11296 (detect_type_change_1): Rename to ...
11297 (detect_type_change): ... this one; early return on non-polymorphic
11299 (detect_type_change_ssa): Add comp_type parameter; update
11300 use of detect_type_change.
11301 (compute_complex_assign_jump_func): Add param_type parameter;
11302 update use of detect_type_change_ssa.
11303 (compute_complex_ancestor_jump_func): Likewise.
11304 (ipa_get_callee_param_type): New function.
11305 (ipa_compute_jump_functions_for_edge): Compute parameter type;
11306 update calls to the jump function computation functions.
11308 2013-08-30 Teresa Johnson <tejohnson@google.com>
11309 Steven Bosscher <steven@gcc.gnu.org>
11311 * cfgrtl.c (fixup_new_cold_bb): New routine.
11312 (commit_edge_insertions): Invoke fixup_partitions.
11313 (find_partition_fixes): New routine.
11314 (fixup_partitions): Ditto.
11315 (verify_hot_cold_block_grouping): Update comments.
11316 (rtl_verify_edges): Invoke find_partition_fixes.
11317 (rtl_verify_bb_pointers): Update comments.
11318 (rtl_verify_bb_layout): Ditto.
11319 * basic-block.h (probably_never_executed_edge_p): Declare.
11320 (fixup_partitions): Ditto.
11321 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
11322 * bb-reorder.c (sanitize_hot_paths): New function.
11323 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
11324 sanitize_hot_paths.
11325 * predict.c (probably_never_executed_edge_p): New routine.
11326 * cfg.c (check_bb_profile): Add partition insanity warnings.
11328 2013-08-30 Meador Inge <meadori@codesourcery.com>
11330 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
11332 2013-08-30 Marek Polacek <polacek@redhat.com>
11334 * Makefile.in (ubsan.o): Add.
11335 (c-family/c-ubsan.o): Add.
11336 (builtins.o): Add ubsan.h dependency.
11337 * ubsan.h: New file.
11338 * ubsan.c: New file.
11339 * common.opt: Add -fsanitize=undefined option.
11340 (flag_sanitize): Add variable.
11341 (fsanitize=): Add option. Add Driver.
11342 (fsanitize=thread): Remove option.
11343 (fsanitize=address): Likewise.
11344 (static-libubsan): New option.
11345 * doc/invoke.texi: Document the new flag and -static-libubsan.
11346 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
11347 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
11348 * builtin-attrs.def (ATTR_COLD): Define.
11349 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
11350 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
11351 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
11352 * flag-types.h (sanitize_code): New enum.
11353 * opts.c (common_handle_option): Parse command line arguments
11354 of -fsanitize=. Add -fsanitize=unreachable option.
11355 * varasm.c (get_variable_section): Adjust.
11356 (assemble_noswitch_variable): Likewise.
11357 (assemble_variable): Likewise.
11358 (output_constant_def_contents): Likewise.
11359 (categorize_decl_for_section): Likewise.
11360 (place_block_symbol): Likewise.
11361 (output_object_block): Likewise.
11362 * builtins.def: Likewise.
11363 * toplev.c (compile_file): Likewise.
11364 (process_options): Likewise.
11365 * cppbuiltin.c: Likewise.
11366 * tsan.c (tsan_pass): Likewise.
11367 (tsan_gate): Likewise.
11368 (tsan_gate_O0): Likewise.
11369 * cfgexpand.c (partition_stack_vars): Likewise.
11370 (expand_stack_vars): Likewise.
11371 (defer_stack_allocation): Likewise.
11372 (expand_used_vars): Likewise.
11373 * cfgcleanup.c (old_insns_match_p): Likewise.
11374 * asan.c (asan_finish_file): Likewise.
11375 (asan_instrument): Likewise.
11376 (gate_asan): Likewise.
11377 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
11378 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
11379 (asan_global_struct): Use pointer_sized_int_node instead
11380 calling build_nonstandard_integer_type.
11381 (initialize_sanitizer_builtins): Likewise.
11382 (asan_finish_file): Likewise.
11383 * gcc.c: Document %{%:function(args):X}.
11384 (static_spec_functions): Add sanitize.
11385 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
11386 store funcval != NULL there.
11387 (do_spec_1): Adjust handle_spec_function caller.
11388 (handle_braces): Allow %:function(args) as condition.
11389 (sanitize_spec_function): New function.
11390 (ADD_STATIC_LIBUBSAN_LIBS): Define.
11391 (LIBUBSAN_SPEC): Likewise.
11392 (LIBUBSAN_EARLY_SPEC): Likewise.
11393 (SANITIZER_SPEC): Handle libubsan.
11394 (SANITIZER_EARLY_SPEC): Likewise.
11395 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
11396 instead of fsanitize=address.
11397 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
11398 instead of fsanitize=address*.
11399 * builtins.c: Include ubsan.h.
11400 (fold_builtin_0): Instrument __builtin_unreachable.
11401 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
11402 instead of flag_asan.
11403 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
11404 (pointer_sized_int_node): Define.
11405 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
11407 2013-08-30 Mike Stump <mikestump@comcast.net>
11409 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
11412 2013-08-29 Jan Hubicka <jh@suse.cz>
11414 * cgraph.c (cgraph_function_body_availability): Handle weakref
11416 * passes.def: Remove pass_fixup_cfg.
11417 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
11418 track when we need to remove functions.
11419 (gate_ipa_inline): Execute inlining always; add comment why.
11420 (pass_data_ipa_inline): Remove TODO_remove_functions.
11421 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
11422 do not produce summaries.
11423 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
11424 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
11425 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
11426 constant pool and vtable.
11428 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
11430 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
11431 New arm_neon.h's internal macros to specify 64-bit constants.
11432 Avoid using stdint.h's macros.
11434 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
11436 * recog.c (verify_changes): Verify that changes[i].old is non-zero
11437 before applying REG_P.
11439 2013-08-30 Jakub Jelinek <jakub@redhat.com>
11441 PR tree-optimization/58277
11442 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
11443 after seeing too many stmts with vdef in between dombb and current
11444 bb, invalidate everything.
11446 2013-08-30 Richard Biener <rguenther@suse.de>
11448 * fold-const.c (fold_single_bit_test): Fix overflow test.
11450 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
11452 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
11453 and which can live in a register, always retrieve the value on entry.
11454 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
11455 passed by invisible reference specially.
11456 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
11457 (vt_add_function_parameter): Correctly deal with a parameter passed by
11458 invisible reference.
11460 2013-08-30 Jan Hubicka <jh@suse.cz>
11462 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
11464 2013-08-30 Richard Biener <rguenther@suse.de>
11466 PR tree-optimization/58228
11467 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
11468 allow invariant loads in nested loop vectorization.
11470 2013-08-30 Richard Biener <rguenther@suse.de>
11472 PR tree-optimization/58223
11473 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
11474 (has_anti_or_output_dependence): ... this and adjust to also
11475 look for output dependences.
11476 (mark_nodes_having_upstream_mem_writes): Adjust.
11477 (rdg_flag_uses): Likewise.
11479 2013-08-30 Richard Biener <rguenther@suse.de>
11481 PR tree-optimization/58010
11482 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
11483 assert that we have a loop-closed PHI.
11485 2013-08-29 Jan Hubicka <jh@suse.cz>
11487 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
11488 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
11489 * lto-section-in.c (lto_free_function_in_decl_state): New function.
11490 (lto_free_function_in_decl_state_for_node): New function.
11492 2013-08-29 Xinliang David Li <davidxl@google.com>
11494 * loop-unroll.c (report_unroll_peel): Minor message change.
11495 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
11496 Emit alignment peeling message with default -fopt-info.
11497 (vect_loop_versioning): Emit loop version info message.
11498 * tree-vectorizer.c (vectorize_loops): Minor message change.
11499 (execute_vect_slp): Ditto.
11501 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
11503 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
11504 of the clone from the DECL_NAME of the original function.
11506 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
11508 * passes.c (register_pass): Add overload.
11509 * tree-pass.h (register_pass): Forward declare it. Add comment.
11511 2013-08-29 Jan Hubicka <jh@suse.cz>
11513 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
11514 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
11515 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
11516 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11517 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
11518 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
11519 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11520 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
11522 2013-08-29 Teresa Johnson <tejohnson@google.com>
11524 * dumpfile.c (dump_loc): Output column number.
11525 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
11526 * doc/invoke.texi: Document optall -fopt-info flag.
11527 * profile.c (read_profile_edge_counts): Use new dump framework.
11528 (compute_branch_probabilities): Ditto.
11529 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
11530 when pass not in any opt group.
11531 * pass_manager.h (pass_manager::get_pass_profile): New method.
11532 * value-prof.c (check_counter): Use new dump framework.
11533 (check_ic_target): Ditto.
11534 * coverage.c (get_coverage_counts): Ditto.
11535 (coverage_init): Setup new dump framework.
11537 2013-08-29 Richard Biener <rguenther@suse.de>
11539 PR tree-optimization/58246
11540 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
11541 handle the dominance check inside a basic-block.
11543 2013-08-29 Richard Biener <rguenther@suse.de>
11545 PR middle-end/57287
11546 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
11547 of default defs that appear in abnormal PHI nodes.
11549 2013-08-29 Richard Biener <rguenther@suse.de>
11551 PR tree-optimization/57685
11552 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
11553 single-use operands to avoid exponential complexity.
11555 2013-08-28 Dehao Chen <dehao@google.com>
11557 * ipa-inline.c (edge_badness): Fix integer underflow.
11559 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
11561 * gtm-builtins.def (_ITM_free): Declare leaf.
11563 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11566 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
11567 (*tls_local_dynamic_base_64_largepic): Likewise.
11568 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
11569 Remove predicate from call operand.
11570 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
11571 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
11573 2013-08-28 Jeff Law <law@redhat.com>
11575 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
11576 checks for the number of predecessors and successors allowed.
11577 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
11578 which require copying a joiner block if there is a request which
11579 is a subpath that requires no joiner block copying.
11581 2013-08-28 Jan Hubicka <jh@suse.cz>
11583 * lto-streamer-out.c (DFS_write_tree_body): Drop
11584 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
11585 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
11586 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
11587 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
11588 Do not read DECL_ERROR_ISSUED.
11589 (unpack_ts_decl_with_vis_value_fields): Do not read
11591 (lto_input_ts_binfo_tree_pointers): Do not read
11592 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
11593 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
11594 write DECL_ERROR_ISSUED..
11595 (pack_ts_decl_with_vis_value_fields): Do not write
11597 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
11598 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
11599 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
11600 * tree.h (tree_decl_common): Update comment.
11601 (DECL_ERROR_ISSUED): Remove.
11603 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11605 PR middle-end/58257
11606 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
11608 2013-08-28 Jan Hubicka <jh@suse.cz>
11610 * builtins.def (free): Declare leaf.
11612 2013-08-27 David Malcolm <dmalcolm@redhat.com>
11614 * gdbhooks.py: New.
11615 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
11616 * configure: Regenerate.
11618 2013-08-27 Martin Jambor <mjambor@suse.cz>
11620 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
11621 (ipa_ancestor_jf_data): Likewise.
11622 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
11623 (ipa_get_jf_pass_through_type_preserved): New function.
11624 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
11625 (ipa_get_jf_ancestor_type_preserved): New function.
11626 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
11627 (ipa_get_jf_ancestor_result): Likewise.
11628 (propagate_vals_accross_pass_through): Use
11629 ipa_get_jf_pass_through_result to do all the value mappings.
11630 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
11631 type_preserved flag.
11632 (ipa_set_jf_cst_copy): New function.
11633 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
11634 (ipa_set_jf_arith_pass_through): Likewise.
11635 (ipa_set_ancestor_jf): Likewise.
11636 (compute_complex_assign_jump_func): Set type_preserved instead of
11638 (ipa_compute_jump_functions_for_edge): Likewise.
11639 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
11640 (update_jump_functions_after_inlining): Update type_preserved.
11641 Explicitely create jump functions when combining one with pass_through.
11642 (ipa_write_jump_function): Stream the type_preserved flags.
11643 (ipa_read_jump_function): Likewise.
11645 2013-08-27 Jakub Jelinek <jakub@redhat.com>
11646 Aldy Hernandez <aldyh@redhat.com>
11648 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
11649 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
11650 * function.h (struct function): Add has_force_vect_loops and
11652 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
11653 * gimple.c (gimple_build_omp_critical): Add KIND argument and
11655 * gimple.def: Update CLAUSES comments.
11656 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
11657 (gimple_build_omp_for): Add argument to prototype.
11658 (gimple_omp_for_kind): New.
11659 (gimple_omp_for_set_kind): New.
11660 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
11661 GOVD_DATA_SHARE_CLASS.
11662 (enum omp_region_type): Add ORT_SIMD.
11663 (gimple_add_tmp_var): Handle ORT_SIMD.
11664 (gimplify_var_or_parm_decl): Same.
11665 (is_gimple_stmt): Same.
11666 (omp_firstprivatize_variable): Same.
11667 (omp_add_variable): Only use splay_tree_insert if lookup failed.
11668 (omp_notice_variable): Handle ORT_SIMD.
11669 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
11670 (omp_check_private): Handle ORT_SIMD.
11671 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11672 OMP_CLAUSE_SAFELEN.
11673 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
11674 Handle OMP_CLAUSE_LASTPRIVATE.
11675 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11676 OMP_CLAUSE_SAFELEN.
11677 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
11678 (gimplify_expr): Handle OMP_SIMD.
11679 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
11680 (expand_GOMP_SIMD_VF): New.
11681 (expand_GOMP_SIMD_LAST_LANE): New.
11682 * internal-fn.def (GOMP_SIMD_LANE): New.
11683 (GOMP_SIMD_VF): New.
11684 (GOMP_SIMD_LAST_LANE): New.
11685 * omp-low.c: Include target.h.
11686 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
11687 OMP_CLAUSE_SAFELEN.
11688 (check_omp_nesting_restrictions): Same.
11690 (lower_rec_simd_input_clauses): New.
11691 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
11693 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
11694 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
11695 (expand_omp_build_assign): New.
11696 (expand_omp_for_init_counts): New.
11697 (expand_omp_for_init_vars): New.
11698 (extract_omp_for_update_vars): New.
11699 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
11700 and rewrite accordingly.
11701 (expand_omp_simd): New.
11702 (expand_omp_for): Use expand_omp_simd.
11703 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
11704 (lower_omp_for): Do not lower the body.
11705 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
11706 in their own loops.
11707 * tree-flow.h (find_omp_clause): Remove prototype.
11708 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
11709 forcing vectorization of the loop, or if flag_tree_vectorize.
11710 (gate_tree_if_conversion): Similarly.
11711 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
11712 gimple_build_omp_for.
11713 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
11714 * tree-parloops (create_parallel_loop): Pass kind argument to
11715 gimple_build_omp_for.
11716 * tree-pretty-print.c (dump_omp_clause): Add cases for
11717 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
11718 OMP_CLAUSE__SIMDUID_.
11719 (dump_generic_node): Handle OMP_SIMD.
11720 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
11721 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
11722 unroll OMP_SIMD loops here.
11723 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
11724 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
11726 (vect_analyze_data_refs): Handle simd loops.
11727 * tree-vect-loop.c (vectorizable_live_operation): Handle
11729 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
11730 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
11731 (vectorizable_load): Same.
11732 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
11733 (struct simduid_to_vf): New.
11734 (simduid_to_vf::hash): New.
11735 (simduid_to-vf::equal): New.
11736 (struct simd_array_to_simduid): New.
11737 (simd_array_to_simduid::hash): New.
11738 (simd_array_to_simduid::equal): New.
11739 (adjust_simduid_builtins): New.
11740 (struct note_simd_array_uses_struct): New.
11741 (note_simd_array_uses_cb): New.
11742 (note_simd_array_uses): New.
11743 (vectorize_loops): Handle simd hints and adjust simd builtins
11745 * tree-vectorizer.h (struct _stmt_vec_info): Add
11746 simd_lane_access_p field.
11747 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
11748 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
11749 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
11750 (omp_clause_code_name): Same.
11751 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
11752 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
11753 * tree.def (OMP_SIMD): New entry.
11754 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
11755 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
11756 (OMP_CLAUSE_DECL): Adjust range for new clauses.
11757 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
11758 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
11759 (OMP_CLAUSE_LINEAR_STEP): New.
11760 (OMP_CLAUSE_SAFELEN_EXPR): New.
11761 (OMP_CLAUSE__SIMDUID__DECL): New.
11762 (find_omp_clause): New prototype.
11764 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
11766 * config/i386/driver-i386.c (host_detect_local_cpu): Update
11767 Haswell processor detection.
11769 2013-08-27 Christian Widmer <shadow@umbrox.de>
11772 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
11773 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
11774 AVX2 capable processors.
11776 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
11778 * config/aarch64/arm_neon.h: Replace all inline asm implementations
11779 of vget_low_* with implementations in terms of other intrinsics.
11781 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
11783 PR middle-end/57219
11784 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
11785 values to -1, 0 and 1.
11787 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
11789 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
11790 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
11791 (<optab><mode>3_insn): Remove.
11792 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
11793 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
11794 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
11795 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
11796 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
11797 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
11798 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
11799 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
11800 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
11801 (ror<mode>3_insn): Likewise.
11802 * config/aarch64/predicates.md (aarch64_simd_register): New.
11804 2013-08-27 Richard Biener <rguenther@suse.de>
11806 PR tree-optimization/57521
11807 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
11808 one edge is non-critical.
11809 (find_phi_replacement_condition): Make sure to use a non-critical
11810 edge. Cleanup and remove old bug workarounds.
11811 (bb_postdominates_preds): Remove.
11812 (if_convertible_loop_p_1): Do not compute post-dominators.
11813 (combine_blocks): Do not free post-dominators.
11814 (main_tree_if_conversion): Likewise.
11815 (pass_data_if_conversion): Add TODO_verify_ssa.
11817 2013-08-27 DJ Delorie <dj@redhat.com>
11819 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
11821 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
11823 * function.c (assign_parm_find_data_types): Set passed_mode and
11824 nominal_mode to the TYPE_MODE of nominal_type for the built
11825 pointer type in case of the struct-pass-by-reference.
11827 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11829 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
11830 (UINT16_TYPE): Change default to "unsigned int".
11832 * config/avr/avr.opt (mfract-convert-truncate): New option.
11833 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
11834 is set, round negative fractional integers according to n1169
11835 when converting to integer types.
11837 2013-08-26 Jan Hubicka <jh@suse.cz>
11839 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
11840 methods can not be called indirectly when their address is not taken.
11842 2013-08-26 Jan Hubicka <jh@suse.cz>
11844 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
11847 2013-08-26 Jan Hubicka <jh@suse.cz>
11849 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11852 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11854 * reload.c (find_valid_class): Allow classes that do not include
11855 FIRST_PSEUDO_REGISTER - 1.
11857 2013-08-26 Jan Hubicka <jh@suse.cz>
11859 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11860 fix edge count/frequency when speculation failed; fix type check
11861 for the direct call.
11863 2013-08-26 Jan Hubicka <jh@suse.cz>
11865 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11867 2013-08-26 Jan Hubicka <jh@suse.cz>
11869 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11870 to be changed by edge redirection.
11872 2013-08-26 Jan Hubicka <jh@suse.cz>
11874 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11875 formating; add sanity check.
11876 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11877 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11878 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11879 (resolve_noninline_speculation): Update callee keys, too.
11881 2013-08-26 Jan Hubicka <jh@suse.cz>
11883 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11884 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11886 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11888 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11891 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11893 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11894 (x86_64_elf_select_section): Ditto.
11895 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11897 (ix86_pass_by_reference): Ditto.
11898 (output_set_got): Ditto.
11899 (ix86_unary_operator_ok): Ditto.
11900 (ix86_expand_builtin): Ditto.
11902 2013-08-23 Jan Hubicka <jh@suse.cz>
11904 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11906 2013-08-23 Jan Hubicka <jh@suse.cz>
11908 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11909 (tree_decl_with_vis): Add FINAL field.
11911 2013-08-23 Jeff Law <law@redhat.com>
11913 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11914 the debugging dump when the expression is fully redundant.
11916 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11918 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11919 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11920 * pretty-print.c (pp_formatted_text_data): Likewise.
11921 (pp_write_text_to_stream): Likewise.
11922 (pp_write_text_as_dot_label_to_stream): Likewise.
11923 (pp_append_r): Likewise.
11924 (pp_format): Likewise.
11925 (pp_flush): Likewise.
11926 (pp_clear_output_area): Likewise.
11927 (pp_append_text): Likewise.
11928 (pp_formatted_text): Likewise.
11929 (pp_remaining_character_count_for_line): Likewise.
11930 (pp_newline): Likewise.
11931 (pp_character): Likewise.
11932 (output_buffer::~output_buffer): Define.
11933 (pretty_printer::~pretty_printer): Destruct output buffer.
11934 * pretty-print.h (output_buffer::~output_buffer): Declare.
11935 (pretty_printer::~pretty_printer): Declare virtual.
11937 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11940 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11941 HOST_WIDE_INT_M1U): New macros.
11942 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11943 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11944 unsigned -1 for lshift.
11945 * cse.c (cse_insn): Likewise.
11946 * double-int.c (rshift_double, lshift_double): Likewise.
11947 * builtins.c (fold_builtin_bitop): Likewise.
11948 * combine.c (force_to_mode): Likewise.
11949 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11950 * simplify-rtx.c (simplify_const_unary_operation,
11951 simplify_const_binary_operation): Likewise.
11952 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11953 check_va_list_escapes): Likewise.
11954 * rtlanal.c (nonzero_bits1): Likewise.
11955 * expmed.c (expand_smod_pow2): Likewise.
11956 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11958 2013-08-23 Jan Hubicka <jh@suse.cz>
11960 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11961 as having address taken.
11963 2013-08-23 Jan Hubicka <jh@suse.cz>
11965 * ipa-utils.h (method_class_type): Declare.
11966 * ipa-devirt.c (method_class_type): Export.
11968 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11969 do not walk base classes of anonymous types.
11971 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11973 PR rtl-optimization/58220
11974 PR regression/58221
11975 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11976 handle SEQUENCE insns properly.
11978 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11980 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11982 (pp_newline_and_indent): Likewise.
11983 (pp_separate_with): Likewise.
11984 * pretty-print.c (pp_newline_and_flush): Define.
11985 (pp_newline_and_indent): Likewise.
11986 (pp_separate_with): Likewise.
11988 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11991 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11992 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
11994 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
11996 * config/i386/predicates.md (ext_sse_reg_operand): New.
11997 * config/i386/i386.md (*movti_internal): Use
11998 predicate to determine if EVEX is needed.
11999 (*movsi_internal): Ditto.
12000 (*movdf_internal): Ditto.
12001 (*movsf_internal): Ditto.
12002 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
12004 2013-08-23 Jakub Jelinek <jakub@redhat.com>
12006 PR tree-optimization/58209
12007 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
12008 (find_tail_calls): Give up for pointer result types if m is non-NULL.
12009 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
12010 emit POINTER_PLUS_EXPR.
12011 (create_tailcall_accumulator): For pointer result type accumulate in
12014 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
12016 * configure.ac: Add backslashes missing from the last change.
12017 * configure: Regenerate.
12019 2013-08-22 Jan Hubicka <jh@susue.cz>
12021 * ipa.c (function_and_variable_visibility): First remember function
12022 was global and then make it local.
12024 2013-08-22 Julian Brown <julian@codesourcery.com>
12026 * configure.ac: Add aarch64 to list of arches which use "nop" in
12028 * configure: Regenerate.
12030 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12032 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
12033 gnu_libc_has_function.
12034 * config/s390/tpf.h: Likewise.
12036 2013-08-22 Jan Hubicka <jh@susue.cz>
12038 * timevar.c (validate_phases): Add cast.
12040 2013-08-22 Jan Hubicka <jh@susue.cz>
12042 * timevar.c (validate_phases): Use size_t for memory.
12043 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
12045 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
12047 * pretty-print.h (output_buffer::output_buffer): Declare.
12048 (pretty_printer::pretty_printer): Likewise.
12049 (pp_construct): Remove.
12050 * pretty-print.c (output_buffer::output_buffer): Define.
12051 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
12052 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
12053 (print_gimple_expr): Likewise.
12054 (print_gimple_seq): Likewise.
12055 (gimple_dump_bb): Likewise.
12056 * sched-vis.c (dump_value_slim): Likewise.
12057 (dump_insn_slim): Likewise.
12058 (dump_rtl_slim): Likewise.
12059 (str_pattern_slim): Likewise.
12060 * tree-mudflap.c (mf_varname_tree): Likewise.
12061 * graph.c (print_graph_cfg): Likewise.
12062 (start_graph_dump): Likewise.
12063 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
12065 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
12066 pretty printer initialization.
12067 * coretypes.h (diagnostic_context): Remove superflous type alias
12069 (pretty_printer): Likewise. Declare directly as a class.
12070 (pretty_print_info): Remove declaration as class.
12071 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
12072 and pp_clear_output_area.
12073 (asan_add_global): Likewise.
12075 2013-08-22 Jan Hubicka <jh@suse.cz>
12077 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
12078 * ipa-utils.h (update_type_inheritance_graph): Declare.
12079 (possible_polymorphic_call_target_p): Declare.
12080 (possible_polymorphic_call_target_p): New.
12081 * ipa-devirt.c: Update toplevel comments.
12082 (cached_polymorphic_call_targets): Move up.
12083 (odr_type_d): Move ID down.
12084 (polymorphic_type_binfo_p): Update comment.
12085 (odr_hasher::remove): Likewise;
12086 (get_odr_type): Set anonymous_namespace.
12087 (dump_odr_type): Dump it.
12088 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
12089 (maybe_record_node): Record node in cached_polymorphic_call_targets.
12090 (record_binfo): Add comment.
12091 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
12093 (devirt_node_removal_hook): Do not iCE when cache is freed.
12094 (possible_polymorphic_call_target_p): New predicate.
12095 (update_type_inheritance_graph): New function.
12097 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
12098 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
12099 Sergey Lega <sergey.s.lega@intel.com>
12100 Anna Tikhonova <anna.tikhonova@intel.com>
12101 Ilya Tocar <ilya.tocar@intel.com>
12102 Andrey Turetskiy <andrey.turetskiy@intel.com>
12103 Ilya Verbin <ilya.verbin@intel.com>
12104 Kirill Yukhin <kirill.yukhin@intel.com>
12105 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12107 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
12108 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
12109 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
12110 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
12111 (OPTION_MASK_ISA_AVX2_UNSET): Update.
12112 (OPTION_MASK_ISA_AVX512F_UNSET): New.
12113 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
12114 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
12115 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
12116 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
12117 OPT_mavx512pf, OPT_mavx512er cases.
12118 * config/i386/constraints.md (v): New constraint.
12119 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
12120 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
12121 (bit_AVX512CD): New.
12122 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
12123 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
12124 * config/i386/i386-c.c (ix86_target_macros_internal):
12125 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
12127 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
12128 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
12129 * config/i386/i386.c (regclass_map, dbx_register_map)
12130 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
12131 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
12132 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
12133 -mavx512pf options.
12134 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
12135 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
12136 -mavx512pf options. Fix formatting.
12137 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
12138 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
12139 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
12140 -mavx512cd, -mavx512pf options.
12141 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
12142 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
12143 (ix86_preferred_output_reload_class): Replace SSE_REGS with
12145 (ix86_hard_regno_mode_ok): Support 512-bit registers.
12146 (ix86_set_reg_reg_cost): Ditto.
12147 (x86_order_regs_for_local_alloc): Ditto.
12148 (MAX_VECT_LEN): Extend to 64-byte.
12149 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
12150 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
12151 (TARGET_AVX512ER, TARGET_AVX512CD): New.
12152 (BIGGEST_ALIGNMENT): Extend to 512-bits.
12153 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
12154 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
12155 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
12156 (SSE_REG_MODE_P): Support new modes.
12157 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
12158 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
12159 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
12160 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
12161 (REG_CLASS_CONTENTS): Add new registers.
12162 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
12163 (EXT_REX_SSE_REGNO_P): New.
12164 (HI_REGISTER_NAMES): Add new registers.
12165 * config/i386/i386.md: Define constants for new registers.
12166 (mode): Add new 512-bit modes.
12167 (prefix): Support evex prefix.
12168 (isa): Support avx512f, noavx512f, fma_avx512f.
12169 (ssemodesuffix): Add new 512-bit modes.
12171 (*movxi_internal_avx512f): Ditto.
12172 (*movdi_internal): Replace constraint "x" with the new constraint "v".
12174 (*movsi_internal): Likewise.
12175 (*movdf_internal): Likewise.
12176 (*movsf_internal): Likewise.
12177 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
12178 (<code><mode>3): Likewise.
12179 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
12181 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
12182 with the new constraint "v".
12183 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
12185 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
12186 with the new constraint "v".
12187 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
12188 (<sse2>_storedqu<avxsizesuffix>): Likewise.
12189 (*<plusminus_insn><mode>3): Likewise.
12190 (<sse>_vm<plusminus_insn><mode>3): Likewise.
12191 (*mul<mode>3): Likewise.
12192 (<sse>_vmmul<mode>3): Likewise.
12193 (<sse>_div<mode>3): Likewise.
12194 (<sse>_vmdiv<mode>3): Likewise.
12195 (<sse>_sqrt<mode>2): Likewise.
12196 (<sse>_vmsqrt<mode>2): Likewise.
12197 (*<code><mode>3_finite): Likewise.
12198 (*<code><mode>3) <smaxmin>: Likewise.
12199 (<sse>_vm<code><mode>3): Likewise.
12200 (*<code><mode>3) <any_logic>: Likewise.
12201 (*fma_fmadd_<mode>): Likewise.
12202 (*fma_fmsub_<mode>): Likewise.
12203 (*fma_fnmadd_<mode>): Likewise.
12204 (*fma_fnmsub_<mode>): Likewise.
12205 (*fma_fmaddsub_<mode>): Likewise.
12206 (*fma_fmsubadd_<mode>): Likewise.
12207 (*fmai_fmadd_<mode>): Likewise.
12208 (*fmai_fmsub_<mode>): Likewise.
12209 (*fmai_fnmadd_<mode>): Likewise.
12210 (*fmai_fnmsub_<mode>): Likewise.
12211 (sse_cvtsi2ss): Likewise.
12212 (sse_cvtsi2ssq): Likewise.
12213 (sse_cvtss2si): Likewise.
12214 (sse_cvtss2si_2): Likewise.
12215 (sse_cvtss2siq): Likewise.
12216 (sse_cvtss2siq_2): Likewise.
12217 (sse_cvttss2si): Likewise.
12218 (sse_cvtss2siq_2): Likewise.
12219 (float<sseintvecmodelower><mode>2): Likewise.
12220 (sse2_cvtsd2si_2): Likewise.
12221 (sse2_cvtsd2siq_2): Likewise.
12222 (*<plusminus_insn><mode>3): Likewise.
12223 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
12224 (*<sse4_1_avx2>_mul<mode>3): Likewise.
12225 (ashr<mode>3): Likewise.
12226 (<shift_insn><mode>3): Likewise.
12227 (avx2_<code><mode>3): Likewise.
12228 (*avx2_<code><mode>3): Likewise.
12229 (*andnot<mode>3): Likewise.
12230 (*<code><mode>3) <any_logic>: Likewise.
12231 (abs<mode>2): Likewise.
12232 (avx2_permvar<mode>): Likewise.
12233 (avx2_perm<mode>_1): Likewise.
12234 (*avx_vpermilp<mode>): Likewise.
12235 (avx_vpermilvar<mode>3): Likewise.
12236 (avx2_ashrv<mode>): Likewise.
12237 (avx2_<shift_insn>v<mode>): Likewise.
12238 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
12240 * doc/rtl.texi: Document XImode.
12242 2013-08-21 Jeff Law <law@redhat.com>
12244 * tree-flow.h (register_jump_thread): Pass vector of edges
12245 instead of each important edge.
12246 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
12247 thread path into a vector and pass that to register_jump_thread.
12248 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
12249 passed in edge vector to the current 3-edge form.
12252 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
12254 * dce.c (fini_dce): Call df_analyze again just in case
12255 delete_unmarked_insns removed anything.
12257 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
12259 * reload.h (struct reg_equivs): Rename to ..
12260 (struct reg_equivs_s): .. this.
12262 2013-08-20 Martin Liska <marxin.liska@gmail.com>
12264 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
12266 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12268 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
12270 2013-08-21 Jeff Law <law@redhat.com>
12272 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
12273 simplify assignments too. If the RHS collapses to a singleton
12274 range, then return the value for the range.
12276 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
12278 * config/i386/sse.md (V16): Rename to...
12280 (mov<mode>): Update iterator name.
12281 (*mov<mode>_internal): Ditto.
12282 (push<mode>1): Ditto.
12283 (movmisalign<mode>): Ditto.
12285 2013-08-20 Jan Hubicka <jh@suse.cz>
12288 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
12289 entry for direct edges.
12290 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
12292 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12294 Revert my last two changes, r201865 and r201864:
12297 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12299 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
12300 instances can own GC refs.
12302 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
12303 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
12304 (gcc::context::gt_pch_nx): Likewise.
12305 (gcc::context::gt_pch_nx): Likewise.
12306 * ggc.h (gt_ggc_mx <T>): New.
12307 (gt_pch_nx_with_op <T>): New.
12308 (gt_pch_nx <T>): New.
12309 * passes.c (opt_pass::gt_ggc_mx): New.
12310 (opt_pass::gt_pch_nx): New.
12311 (opt_pass::gt_pch_nx_with_op): New.
12312 (pass_manager::gt_ggc_mx): New.
12313 (pass_manager::gt_pch_nx): New.
12314 (pass_manager::gt_pch_nx_with_op): New.
12315 (pass_manager::operator new): Use
12316 ggc_internal_cleared_alloc_stat rather than xcalloc.
12317 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
12318 (pass_manager::gt_ggc_mx): New.
12319 (pass_manager::gt_pch_nx): New.
12320 (pass_manager::gt_pch_nx_with_op): New.
12321 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
12322 (opt_pass::operator new): New.
12323 (opt_pass::gt_ggc_mx): New.
12324 (opt_pass::gt_pch_nx): New.
12325 (opt_pass::gt_pch_nx_with_op): New.
12328 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12330 * Makefile.in (GTFILES): Add context.h.
12331 * context.c (gcc::context::operator new): New.
12332 (gcc::context::gt_ggc_mx): New.
12333 (gcc::context::gt_pch_nx): New.
12334 (gcc::context::gt_pch_nx): New.
12335 * context.h (gcc::context): Add GTY((user)) marking.
12336 (gcc::context::operator new): New.
12337 (gcc::context::gt_ggc_mx): New.
12338 (gcc::context::gt_pch_nx): New.
12339 (gcc::context::gt_pch_nx): New.
12340 (g): Add GTY marking.
12341 (gt_ggc_mx (gcc::context *)): New.
12342 (gt_pch_nx (gcc::context *)): New.
12343 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
12344 void *cookie)): New.
12345 * gengtype.c (open_base_files) <ifiles>: Add context.h.
12347 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
12349 * dce.c (fini_dce): Call df_analyze again just in case
12350 delete_unmarked_insns removed anything.
12352 2013-08-20 Teresa Johnson <tejohnson@google.com>
12354 PR rtl-optimizations/57451
12355 * final.c (reemit_insn_block_notes): Prevent lexical blocks
12356 from crossing split section boundaries.
12358 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12360 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
12361 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
12362 with MULTLIB_DEFAULTS.
12364 2013-08-20 Nick Clifton <nickc@redhat.com>
12366 * target.def (narrow_volatile_bitfield): Note that the default
12367 value is false, not !TARGET_STRICT_ALIGN.
12368 * doc/tm.texi: Regenerate.
12370 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
12372 Fix LIB_SPEC for systems without libpthread.
12374 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
12375 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
12377 * config/i386/linux-common.h: Likewise.
12378 * config/mips/linux-common.h: Likewise.
12380 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12382 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
12385 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12387 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
12388 instances can own GC refs.
12390 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
12391 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
12392 (gcc::context::gt_pch_nx): Likewise.
12393 (gcc::context::gt_pch_nx): Likewise.
12394 * ggc.h (gt_ggc_mx <T>): New.
12395 (gt_pch_nx_with_op <T>): New.
12396 (gt_pch_nx <T>): New.
12397 * passes.c (opt_pass::gt_ggc_mx): New.
12398 (opt_pass::gt_pch_nx): New.
12399 (opt_pass::gt_pch_nx_with_op): New.
12400 (pass_manager::gt_ggc_mx): New.
12401 (pass_manager::gt_pch_nx): New.
12402 (pass_manager::gt_pch_nx_with_op): New.
12403 (pass_manager::operator new): Use
12404 ggc_internal_cleared_alloc_stat rather than xcalloc.
12405 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
12406 (pass_manager::gt_ggc_mx): New.
12407 (pass_manager::gt_pch_nx): New.
12408 (pass_manager::gt_pch_nx_with_op): New.
12409 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
12410 (opt_pass::operator new): New.
12411 (opt_pass::gt_ggc_mx): New.
12412 (opt_pass::gt_pch_nx): New.
12413 (opt_pass::gt_pch_nx_with_op): New.
12415 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12417 * Makefile.in (GTFILES): Add context.h.
12418 * context.c (gcc::context::operator new): New.
12419 (gcc::context::gt_ggc_mx): New.
12420 (gcc::context::gt_pch_nx): New.
12421 (gcc::context::gt_pch_nx): New.
12422 * context.h (gcc::context): Add GTY((user)) marking.
12423 (gcc::context::operator new): New.
12424 (gcc::context::gt_ggc_mx): New.
12425 (gcc::context::gt_pch_nx): New.
12426 (gcc::context::gt_pch_nx): New.
12427 (g): Add GTY marking.
12428 (gt_ggc_mx (gcc::context *)): New.
12429 (gt_pch_nx (gcc::context *)): New.
12430 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
12431 void *cookie)): New.
12432 * gengtype.c (open_base_files) <ifiles>: Add context.h.
12434 2013-08-20 Alan Modra <amodra@gmail.com>
12437 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
12438 (rs6000_emit_epilogue): Likewise.
12440 2013-08-19 Dehao Chen <dehao@google.com>
12442 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
12444 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
12445 Jakub Jelinek <jakub@redhat.com>
12447 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
12448 (BUILT_IN_FABSD64): Likewise.
12449 (BUILT_IN_FABSD128): Likewise.
12450 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
12451 (fold_builtin_1): Likewise.
12452 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
12453 destination and source operands.
12454 (*abstd2_fpr): Likewise.
12455 (*nabstd2_fpr): Likewise.
12457 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
12459 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
12462 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
12464 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
12467 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
12469 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
12470 * builtins.c (default_libc_has_function): New.
12471 (gnu_libc_has_function): Ditto.
12472 (no_c99_libc_has_function): Ditto.
12473 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
12474 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
12475 (fold_builtin_sincos): Likewise.
12476 (fold_builtin_cexp): Likewise.
12477 * builtins.def (DEF_C94_BUILTIN): Likewise.
12478 (DEF_C99_BUILTIN): Likewise.
12479 (DEF_C99_C90RES_BUILTIN): Likewise.
12480 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
12481 definitions to using this define.
12482 * config/darwin-protos.h (darwin_libc_has_function): New.
12483 * config/darwin.c (darwin_libc_has_function): Ditto.
12484 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
12485 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
12486 * config/darwin.h: Ditto.
12487 * config/elfos.h: Ditto.
12488 * config/freebsd.h: Ditto.
12489 * config/i386/cygming.h: Ditto.
12490 * config/i386/djgpp.h: Ditto.
12491 * config/i386/i386-interix.h: Ditto.
12492 * config/microblaze/microblaze.h: Ditto.
12493 * config/mmix/mmix.h: Ditto.
12494 * config/gnu-user.h: Ditto.
12495 * config/ia64/hpux.h: Ditto.
12496 * config/pa/pa-hpux.h: Ditto.
12497 * config/pdp11/pdp11.h: Ditto.
12498 * config/picochip/picochip.h: Ditto.
12499 * config/linux.h: Ditto.
12500 * config/netbsd.h: Ditto.
12501 * config/openbsd.h: Ditto.
12502 * config/rs6000/aix43.h: Ditto.
12503 * config/rs6000/aix51.h: Ditto.
12504 * config/rs6000/aix52.h: Ditto.
12505 * config/rs6000/aix53.h: Ditto.
12506 * config/rs6000/aix61.h: Ditto.
12507 * config/rs6000/darwin.h: Ditto.
12508 * config/rs6000/linux.h: Ditto.
12509 * config/rs6000/linux64.h: Ditto.
12510 * config/s390/tpf.h: Ditto.
12511 * config/sol2-10.h: Ditto.
12512 * config/sol2.h: Ditto.
12513 * config/vms/vms.h: Ditto.
12514 * config/vxworks.h: Ditto.
12515 * config/linux-android.c (linux_android_libc_has_function):
12516 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
12517 * config/linux-protos.h (linux_android_libc_has_function):
12519 * config/i386/i386.c (ix86_libc_has_function): New.
12520 * config/i386/i386-protos.h
12521 (ix86_libc_has_function): New declaration.
12522 * config/i386/i386.md
12523 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
12524 ("isinf<mode>2): Likewise.
12525 * convert.c (convert_to_integer): Using new target hook
12526 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
12527 TARGET_C99_FUNCTIONS.
12528 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
12529 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
12530 * coretypes.h (function_class): New enum for different
12531 classes of functions.
12532 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
12533 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
12534 (TARGET_HAS_SINCOS): Likewise.
12535 (TARGET_LIBC_HAS_FUNCTION): New.
12536 * doc/tm.texi: Regenerated.
12537 * targhooks.h (default_libc_has_function): New declaration.
12538 (no_c99_libc_has_function): Ditto.
12539 (gnu_libc_has_function): Ditto.
12540 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
12541 and TARGET_HAS_SINCOS.
12543 2013-08-18 Jan Hubicka <jh@suse.cz>
12545 * Makeifle-in (ipa-devirt.o): New.
12546 (GTFILES): Add ipa-utils.h and ipa-devirt.c
12547 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
12548 (analyze_functions): Look into possible targets of polymorphic call.
12549 * dumpfile.c (dump_files): Add type-inheritance dump.
12550 * dumpfile.h (TDI_inheritance): New.
12551 * ipa-devirt.c: New file.
12552 * ipa-utils.h (odr_type_d): Forward declare.
12553 (odr_type): New type.
12554 (build_type_inheritance_graph): Declare.
12555 (possible_polymorphic_call_targets): Declare and introduce inline
12556 variant when only edge is pased.
12557 (dump_possible_polymorphic_call_targets): Likewise.
12558 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
12559 * tree.c (type_in_anonymous_namespace_p): Break out from ...
12560 (types_same_for_odr): ... here.
12561 * tree.h (type_in_anonymous_namespace_p): Declare.
12563 2013-08-18 Jakub Jelinek <jakub@redhat.com>
12565 PR tree-optimization/58006
12566 * tree-parloops.c (take_address_of): Don't ICE if get_name
12568 (eliminate_local_variables_stmt): Remove clobber stmts.
12570 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
12572 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
12573 error message is issued for an alias to undefined symbol.
12575 2013-08-18 Jan Hubicka <jh@suse.cz>
12577 * cgraph.c (cgraph_create_indirect_edge): Discover
12578 polymorphic calls and record basic info into indirect_info.
12579 * gimple-fold.c (gimple_fold_call): When doing BINFO based
12580 devirtualization, ignore objc function calls.
12581 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
12582 call with no parm index info.
12583 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
12584 * tree.c (virtual_method_call_p): New function.
12585 * tree.h (virtual_method_call_p): Declare.
12587 2013-08-16 Jan Hubicka <jh@suse.cz>
12589 PR middle-end/58179
12590 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
12592 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
12594 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
12597 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12599 * gengtype.c (type_for_name): Add special-case support for
12600 locating types within the "gcc::" namespace.
12601 (open_base_files): Emit a "using namespace gcc" directive.
12603 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12606 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
12607 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
12609 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
12610 array instead of each individual operand as a separate argument.
12611 (emit_fusion_gpr_load): Likewise.
12612 (expand_fusion_gpr_load): Add new function declaration.
12614 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
12615 signature to have the operands passed as an array, instead of as
12616 separate arguments. Allow ZERO_EXTEND to be in the memory
12617 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
12618 depend on the register live/dead flags when peepholes are run.
12619 (expand_fusion_gpr_load): New function to be called from the
12620 peephole2 pass, to change the register that addis sets to be the
12622 (emit_fusion_gpr_load): Change the calling signature to have the
12623 operands passed as an array, instead of as separate arguments.
12624 Allow ZERO_EXTEND to be in the memory address, and also
12625 SIGN_EXTEND if -mpower8-fusion-sign.
12627 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
12628 unspec enumeration.
12629 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
12630 adjust the register addis loads up in the peephole2 pass. Do not
12631 depend on the register live/dead state when the peephole pass is done.
12633 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12635 * gengtype.c (create_user_defined_type): Ensure that the kind
12636 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
12637 declaration is seen before the GTY((user)) marking.
12639 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
12642 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
12644 2013-08-16 Jan Hubicka <jh@suse.cz>
12646 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
12647 arugment expected_type.
12648 (gimple_fold_call): Use it.
12649 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
12650 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
12651 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
12652 (try_make_edge_direct_virtual_call): Likewise.
12653 * tree.c (obj_type_ref_class): New.
12654 * tree.h (obj_type_ref_class): Use it.
12656 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
12658 * sched-vis.c (rtl_slim_pp_initialized): Remove.
12659 (rtl_slim_pp): Likewise.
12660 (init_rtl_slim_pretty_print): Likewise.
12661 (dump_value_slim): Don't call it. Use local pretty printer.
12662 (dump_insn_slim): Likewise.
12663 (dump_rtl_slim): Likewise.
12664 (str_pattern_slim): Likewise.
12665 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
12668 2013-08-16 Jakub Jelinek <jakub@redhat.com>
12670 PR tree-optimization/58164
12671 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
12672 walk gimple_goto_dest of GIMPLE_GOTO.
12674 PR tree-optimization/58165
12675 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
12676 bi_call must be the last stmt in a bb, don't split_block, instead
12677 use fallthru edge from it and give up if there is none.
12678 Release conds vector when returning early.
12680 2013-08-14 Xinliang David Li <davidxl@google.com>
12682 * config/i386/i386.c (ix86_option_override_internal):
12683 Remove unused variable and field.
12685 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12688 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
12689 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
12690 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
12691 and Linux, correct BLKmode alignment when 128-bit alignment is
12692 required and compatibility flag is not set.
12693 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
12694 for zero-size arguments when compatibility flag is not set.
12696 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12698 PR tree-optimization/58145
12699 * tree-sra.c (build_ref_for_offset): If prev_base has
12700 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
12702 2013-08-14 Xinliang David Li <davidxl@google.com>
12704 * config/i386/i386.c (ix86_option_override_internal):
12705 Fix uninitialized variable error.
12707 2013-08-14 Xinliang David Li <davidxl@google.com>
12709 * config/i386/i386.opt: Define two new options.
12710 * config/i386/x86-tune.def: Add arch selector field in macros.
12711 * config/i386/i386.h: Adjust macro definition.
12712 * config/i386/i386.c (ix86_option_override_internal):
12714 (parse_mtune_ctrl_str): New function.
12715 (set_ix86_tune_features): New function.
12716 (ix86_function_specific_restore): Call the new helper function.
12718 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
12720 PR rtl-optimization/57662
12721 * sel-sched.c (code_motion_process_successors): When the current insn
12722 is removed after the recursive traversal, break from the loop.
12723 Add comments and debug printouts.
12725 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12726 Alexandre Oliva <aoliva@redhat.com>
12729 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
12730 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
12731 there also UNSPEC_PLTOFF.
12733 2013-08-14 Marek Polacek <polacek@redhat.com>
12735 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
12736 NUM_CONDITIONS bit positions.
12738 2013-08-13 Cary Coutant <ccoutant@google.com>
12740 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
12741 (attr_checksum): Hash vector contents instead of pointer.
12742 (attr_checksum_ordered): Likewise.
12744 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
12746 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
12747 when Pmode != word_mode. Add length_address attribute.
12748 (sse3_monitor_<mode>): Merge from sse3_monitor and
12749 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
12750 Pmode != word_mode. Update insn length attribute.
12751 * config/i386/i386.c (ix86_option_override_internal): Update
12752 ix86_gen_monitor selection for merged sse3_monitor insn.
12754 2013-08-13 Julian Brown <julian@codesourcery.com>
12756 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
12757 perform invalid legitimization on greater-than-word-size modes for
12758 TARGET_E500_DOUBLE.
12760 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
12762 * ira.c (setup_class_translate_array): Use aclass instead of cl
12763 for classes not fully covered by allocno classes.
12765 2013-08-13 Jakub Jelinek <jakub@redhat.com>
12767 PR tree-optimization/57661
12768 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
12769 * tree-inline.c (tree_function_versioning): Initialize it.
12770 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
12771 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
12772 that is not being copied.
12775 * asan.c (instrument_strlen_call): Fix typo in comment.
12776 Use char * type even for the lhs of POINTER_PLUS_EXPR.
12778 2013-08-13 Steve Ellcey <sellcey@mips.com>
12780 * config/mips/mips.md (prefetch): Use lw instead of ld on
12781 loongson in 32bit mode.
12783 2013-08-13 Nick Clifton <nickc@redhat.com>
12785 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
12787 2013-08-13 Jan Hubicka <jh@suse.cz>
12789 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
12790 introduced edge; fix typo in sanity check.
12791 (cgraph_resolve_speculation): Export; improve diagnostic.
12792 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
12793 speculation at type mismatch.
12794 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
12795 (cgraph_resolve_speculation): Declare.
12796 (symtab_can_be_discarded): New function.
12797 * value-prof.c (gimple_ic_transform): Remove actual transform code.
12798 * ipa-inline-transform.c (speculation_removed): New global var.
12799 (clone_inlined_nodes): See if speculation can be removed.
12800 (inline_call): If speculations was removed, we growths may not match.
12801 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
12802 (speculation_useful_p): New function.
12803 (resolve_noninline_speculation): New function.
12804 (inline_small_functions): Resolve useless speculations.
12805 * ipa-inline.h (speculation_useful_p): Declare
12806 * ipa.c (can_replace_by_local_alias): Simplify.
12807 (ipa_profile): Produce speculative calls in non-lto, too;
12808 add simple cost model; produce local aliases.
12810 2013-08-13 David Malcolm <dmalcolm@redhat.com>
12812 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
12815 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
12817 * config/i386/i386.c (ix86_function_versions): Use error + inform.
12819 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
12821 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
12822 iterator instead of X87MODEF.
12824 2013-08-12 Perez Read <netfirewall@gmail.com>
12827 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
12828 operand 0 for intel asm alternative.
12829 (*movabs<mode>_2): Ditto for operand 1.
12831 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
12833 * config/aarch64/arm_none.h
12834 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
12836 2013-08-12 Nick Clifton <nickc@redhat.com>
12838 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
12840 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12842 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
12843 expand for QI/HImode operand to produce more effictive code for
12844 unsigned char(short) --> float(double) conversion.
12846 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
12848 * doc/invoke.texi: Mention that -ftls-model does not force the final
12851 2013-08-12 Marek Polacek <polacek@redhat.com>
12852 Marc Glisse <marc.glisse@inria.fr>
12854 PR tree-optimization/57980
12855 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12856 when creating -1 constant.
12858 2013-08-10 Jan Hubicka <jh@suse.cz>
12860 Workaround binutils PR14342.
12861 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12862 (gimple_init_edge_profiler): Likewise.
12863 (gimple_gen_ic_func_profiler): Likewise.
12865 2013-08-09 Jan Hubicka <jh@suse.cz>
12867 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12869 2013-08-09 Xinliang David Li <davidxl@google.com>
12871 * config/i386/stringop.def: New file.
12872 * config/i386/stringop.opt: New file.
12873 * config/i386/i386-opts.h: Include stringopt.def.
12874 * config/i386/i386.opt: Include stringopt.opt.
12875 * config/i386/i386.c (ix86_option_override_internal):
12876 Override default size based stringop inline strategies with options.
12877 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12880 2013-08-09 Jan Hubicka <jh@suse.cz>
12882 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12884 2013-08-09 Jan Hubicka <jh@suse.cz>
12886 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12888 (dump_cgraph_node): Dump profile-id.
12889 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12890 and common_target_probability.
12891 * lto-cgraph.c (lto_output_edge): Stream common targets.
12892 (lto_output_node): Stream profile ids.
12893 (input_node): Stream profile ids.
12894 (input_edge): Stream common targets.
12895 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12896 * ipa.c: Include value-prof.h
12897 (ipa_profile_generate_summary): Turn indirect call statement histograms
12898 into common targets.
12899 (ipa_profile): Turn common targets into speculative edges.
12901 2013-08-09 Jan Hubicka <jh@suse.cz>
12903 * cgraph.h (cgraph_node): Add profile_id.
12904 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12905 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12906 (del_node_map): Update.
12907 (find_func_by_funcdef_no): Replace by ...
12908 (find_func_by_profile_id): ... this one.
12909 (gimple_ic_transform): Do not remove useful histograms when
12910 speculation is not done; dump info when indirect call removal
12912 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12913 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12914 (__gcov_indirect_call_profiler_v2): .. this one.
12915 * profile.h (init_node_map): Update.
12916 * coverage.c (coverage_compute_profile_id): New function.
12917 * coverage.h (coverage_compute_profile_id): Declare.
12918 * tree-profile.c (init_ic_make_global_vars): Make
12919 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12920 (gimple_init_edge_profiler): Update prototype of
12921 __gcov_indirect_call_profiler.
12922 (gimple_gen_ic_func_profiler): Simplify.
12923 (tree_profiling): Use init_node_map
12925 2013-08-09 Jan Hubicka <jh@suse.cz>
12927 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12928 non-speculative refs.
12929 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12930 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12931 (cgraph_set_call_stmt): Likewise.
12932 (cgraph_create_edge_1): Fix release checking compilatoin;
12933 clear lto_stmt_uid.
12934 (cgraph_free_edge): Free indirect info.
12935 (cgraph_turn_edge_to_speculative): New function.
12936 (cgraph_speculative_call_info): New function.
12937 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12938 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12939 (dump_cgraph_node): Dump speculation.
12940 (verify_edge_count_and_frequency): Accept speculative edges.
12941 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12942 (verify_cgraph_node): Handle speculation.
12943 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12944 (cgraph_set_call_stmt): Update prototype.
12945 (cgraph_make_edge_direct): Update prototype.
12946 (cgraph_speculative_call_info): Declare.
12947 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12948 to change; update call of ipa_find_references.
12949 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12950 lto_stmt_uid and speculative flags.
12951 (ipa_dump_references): Dump speculation.
12952 (ipa_clone_references): Clone speculative flag.
12953 (ipa_clone_referring): Likewise.
12954 (ipa_clone_ref): New function.
12955 (ipa_find_reference): Look into lto_stmt_uids
12956 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12957 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12958 (ipa_find_reference): Update declaration.
12959 (ipa_clone_ref): Declare.
12960 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12961 stream speculative flag.
12962 (lto_output_ref): Stream statements uids and speculation.
12963 (input_ref): Likewise.
12964 (input_edge): Stream speuclation.
12965 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12966 (cgraph_set_call_stmt_including_clones): Handle speculation.
12967 * ipa-inline.c (heap_edge_removal_hook): New function.
12968 (inline_small_functions): Register it.
12969 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12970 also initialize refs.
12971 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12973 (try_make_edge_direct_simple_call): Likewise.
12974 (try_make_edge_direct_simple_call): Likewise.
12975 (update_indirect_edges_after_inlining): Likewise.
12976 (remove_described_reference): Look proper lto_stmt_uid.
12977 (propagate_controlled_uses): Likewise.
12978 (propagate_controlled_uses): Liekwise.
12979 * tree-inline.c (copy_bb): Copy speculative edges.
12980 (redirect_all_calls): New function.
12981 (copy_cfg_body): Do redirection after loop info is updated.
12982 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12984 2013-08-09 Jan Hubicka <jh@suse.cz>
12986 * lto-streamer-out.c (output_function): Renumber PHIs.
12987 * lto-streamer-in.c (input_function): Likewise.
12989 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12991 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12992 (get_lane_unsigned): Likewise.
12993 (dup_lane_scalar): Likewise.
12994 (get_lane): enable for VALL.
12995 * config/aarch64/aarch64-simd.md
12996 (aarch64_dup_lane_scalar<mode>): Remove.
12997 (aarch64_get_lane_signed<mode>): Likewise.
12998 (aarch64_get_lane_unsigned<mode>): Likewise.
12999 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
13000 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
13001 (aarch64_get_lane<mode>): Enable for all vector modes.
13002 (aarch64_get_lanedi): Remove misleading constraints.
13003 * config/aarch64/arm_neon.h
13004 (__aarch64_vget_lane_any): Define.
13005 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
13006 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
13007 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
13008 * config/aarch64/iterators.md (VDQQH): New.
13009 (VDQQHS): Likewise.
13010 (vwcore): Likewise.
13012 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
13014 * configure.ac: Add GAS check for LEON instructions on SPARC.
13015 * configure: Regenerate.
13016 * config.in: Likewise.
13017 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
13019 * config/sparc/sparc.opt (LEON, LEON3): New masks.
13020 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
13022 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
13023 (AS_LEON_FLAG): New macro.
13024 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
13025 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
13026 Deal with LEON and LEON3 for the memory model.
13027 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
13028 (atomic_compare_and_swap<mode>_1): Likewise.
13029 (*atomic_compare_and_swap<mode>_1): Likewise.
13031 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13033 * config/arm/neon.md (vcond): Fix floating-point vector
13034 comparisons against 0.
13036 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
13038 * lra-constraints.c (emit_spill_move): Remove assert.
13039 (process_alt_operands): Add more debugging
13040 output. Increase reject for spilling into memory. Decrease
13041 reject for reloading scratch.
13042 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
13044 2013-08-08 Steve Ellcey <sellcey@mips.com>
13046 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
13047 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
13048 micromips incompatible. Add nan2008.
13049 (MULTILIB_DIRNAMES): Add nan2008.
13050 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
13051 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
13052 and micromips incompatible. Add nan2008.
13053 (MULTILIB_DIRNAMES): Add nan2008.
13054 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
13056 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
13058 PR rtl-optimization/58079
13059 * combine.c (combine_simplify_rtx): Avoid using SUBST if
13060 simplify_comparison has widened a comparison with an integer.
13062 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13064 * config/arm/neon.md (movmisalign<mode>): Disable when we
13065 don't allow unaligned accesses.
13066 (*movmisalign<mode>_neon_store): Likewise.
13067 (*movmisalign<mode>_neon_load): Likewise.
13068 (*movmisalign<mode>_neon_store): Likewise.
13069 (*movmisalign<mode>_neon_load): Likewise.
13071 2013-08-08 Jan Hubicka <jh@suse.cz>
13073 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
13074 (make_pass_rebuild_cgraph_edges): Also clear references.
13075 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
13076 * ipa-inline-transform.c (inline_transform): Remove all references
13078 * cgraphunit.c (expand_function): Remove all references after
13080 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
13081 (ipa_find_reference): Rewrite to iterator.
13082 (remove_stmt_references): Likewise.
13083 (ipa_clear_stmts_in_references): New function.
13084 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
13085 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
13087 * ipa-split.c (split_function): Remove references in split function.
13089 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
13092 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
13093 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
13095 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
13098 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
13099 suggested mode for the assignment isn't compatible with the
13100 registers required.
13102 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
13105 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
13107 2013-08-07 Xinliang David Li <davidxl@google.com>
13109 * config/i386/i386.opt: New option -mtune-ctrl=.
13110 * config/i386/x86-tune.def: New file.
13111 * config/i386/i386.h: include x86-tune.def.
13112 * config/i386/i386.c (ix86_option_override_internal):
13113 Parsing -mtune-ctrl= option and set tune features.
13115 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
13118 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
13119 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
13122 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
13124 * rtl.h (update_alignments): Declare.
13125 * final.c (grow_label_align): New function extracted from...
13126 (shorten_branches): ...here. Call it.
13127 (update_alignments): New function.
13128 * reorg.c (sibling_labels): New variable.
13129 (get_label_before): Add SIBLING parameter. If it is non-zero, push
13130 the new label along with it onto the sibling_labels vector.
13131 (fill_simple_delay_slots): Adjust call to get_label_before.
13132 (fill_slots_from_thread): Likewise.
13133 (relax_delay_slots): Likewise.
13134 (make_return_insns): Likewise.
13135 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
13137 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
13139 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
13140 document its semantics.
13141 (diagnostic_report_diagnostic): Adjust accordingly.
13143 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13145 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
13146 (sparc_option_override): ...and port to new C++ pass API.
13147 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
13149 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
13151 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
13153 2013-08-06 Caroline Tice <cmtice@google.com>
13155 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
13156 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
13157 * tree-pass.h: Add pass_vtable_verify.
13158 * varasm.c (assemble_variable): Add code to properly set the comdat
13159 section and name for the .vtable_map_vars section.
13160 (assemble_vtyv_preinit_initializer): New function.
13161 (default_sectin_type_flags): Make sure .vtable_map_vars section has
13163 * output.h: Add function decl for assemble_vtv_preinit_initializer.
13164 * vtable-verify.c: New file.
13165 * vtable-verify.h: New file.
13166 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
13167 initialiation levels.
13168 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
13169 * passes.def: Insert pass_vtable_verify.
13170 * aclocal.m4: Reorder includes.
13171 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
13172 -fvtv-counts options.
13173 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
13174 as appropriate, if -fvtable-verify=... is used.
13175 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
13176 -fvtable-verify=... is used.
13177 * Makefile.in (OBJS): Add vtable-verify.o to list.
13178 (vtable-verify.o): Add new build rule.
13179 (GTFILES): Add vtable-verify.c to list.
13180 * common.opt (fvtable-verify=): New flag.
13181 (vtv_priority): Values for fvtable-verify= flag.
13182 (fvtv-counts): New flag.
13183 (fvtv-debug): New flag.
13184 * tree.h (save_vtable_map_decl): New extern function decl.
13186 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13188 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
13189 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
13190 (pass_data_rl78_devirt): ...new pass_data instance and...
13191 (make_pass_rl78_devirt): ...new function.
13192 (rl78_asm_file_start): Port pass registration to new C++ API.
13194 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13196 * coretypes.h (rtl_opt_pass): Add.
13197 (gcc::context): Add.
13198 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
13199 (epiphany_init): Port to new C++ pass API.
13200 (epiphany_optimize_mode_switching): Likewise.
13201 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
13202 (pass_manager::get_pass_mode_switching): New.
13203 (pass_manager::get_pass_peephole2): New.
13204 * mode-switching.c (pass_mode_switching): Add clone method.
13205 * recog.c (pass_peephole2): Add clone method.
13206 (pass_split_all_insns): Add clone method.
13208 2013-08-06 David Malcolm <dmalcolm@redhat.com>
13210 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
13211 (mips_option_override): ...here, porting to new C++ API for passes.
13213 2013-08-06 Jan Hubicka <jh@suse.cz>
13215 * cgraph.c (cgraph_get_body): New function based on lto.c
13217 * cgraph.h (cgraph_get_body): Declare.
13218 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
13220 * cgraphunit.c (expand_function): Get body prior expanding.
13221 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
13222 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
13224 * passes.c (do_per_function_toporder): Get body.
13225 * tree-inline.c (expand_call_inline): Get body prior inlining it.
13226 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
13228 2013-08-06 Martin Jambor <mjambor@suse.cz>
13231 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
13232 is not re-finalized. Rename second parameter to no_collect.
13234 2013-08-06 Martin Jambor <mjambor@suse.cz>
13236 PR middle-end/58041
13237 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
13238 MEM_REF has proper alignment information.
13240 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
13243 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
13245 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
13246 rtx (*) (rtx, ...) with insn_gen_fn.
13247 * genoutput.c (output_insn_data): Cast gen_? function pointers to
13248 insn_gen_fn::stored_funcptr. Add initializer braces.
13250 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13252 Rewrite how instances of passes are cloned to remove assumptions
13253 about their sizes (thus allowing pass subclasses to have
13254 additional data fields, albeit non-GC-managed ones at this point).
13256 * passes.c (make_pass_instance): Now that passes have clone
13257 methods, rewrite this function to eliminate XNEW and memcpy
13258 calls that used hardcoded sizes. Since this function no longer
13259 creates pass instances, rename it to...
13260 (add_pass_instance): ...this. Document the old way that passes were
13261 numbered and flagged, and rework this function to continue using it.
13262 (next_pass_1): Add an initial_pass argument for use by
13264 (position_pass): When adding multiple instances of a pass, use
13265 the pass's clone method, rather than relying on the XNEW/memcpy
13266 within the former make_pass_instance (now add_pass_instance).
13267 (pass_manager::pass_manager): When invoking next_pass_1, also supply
13268 the initial instance of the current pass within the pass manager.
13270 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13272 This is the automated part of the conversion of passes from C
13273 structs to C++ classes.
13275 Patch autogenerated by refactor_passes.py from
13276 https://github.com/davidmalcolm/gcc-refactoring-scripts
13277 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
13279 * asan.c (pass_asan): Convert from a global struct to a subclass of
13280 gimple_opt_pass along with...
13281 (pass_data_asan): ...new pass_data instance and...
13282 (make_pass_asan): ...new function.
13283 (pass_asan_O0): Convert from a global struct to a subclass of
13284 gimple_opt_pass along with...
13285 (pass_data_asan_O0): ...new pass_data instance and...
13286 (make_pass_asan_O0): ...new function.
13287 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
13288 subclass of rtl_opt_pass along with...
13289 (pass_data_inc_dec): ...new pass_data instance and...
13290 (make_pass_inc_dec): ...new function.
13291 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
13292 a subclass of rtl_opt_pass along with...
13293 (pass_data_reorder_blocks): ...new pass_data instance and...
13294 (make_pass_reorder_blocks): ...new function.
13295 (pass_duplicate_computed_gotos): Convert from a global struct to a
13296 subclass of rtl_opt_pass along with...
13297 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
13298 (make_pass_duplicate_computed_gotos): ...new function.
13299 (pass_partition_blocks): Convert from a global struct to a subclass of
13300 rtl_opt_pass along with...
13301 (pass_data_partition_blocks): ...new pass_data instance and...
13302 (make_pass_partition_blocks): ...new function.
13303 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
13304 struct to a subclass of rtl_opt_pass along with...
13305 (pass_data_branch_target_load_optimize1): ...new pass_data instance
13307 (make_pass_branch_target_load_optimize1): ...new function.
13308 (pass_branch_target_load_optimize2): Convert from a global struct to a
13309 subclass of rtl_opt_pass along with...
13310 (pass_data_branch_target_load_optimize2): ...new pass_data instance
13312 (make_pass_branch_target_load_optimize2): ...new function.
13313 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
13314 of rtl_opt_pass along with...
13315 (pass_data_jump): ...new pass_data instance and...
13316 (make_pass_jump): ...new function.
13317 (pass_jump2): Convert from a global struct to a subclass of
13318 rtl_opt_pass along with...
13319 (pass_data_jump2): ...new pass_data instance and...
13320 (make_pass_jump2): ...new function.
13321 * cfgexpand.c (pass_expand): Convert from a global struct to a
13322 subclass of rtl_opt_pass along with...
13323 (pass_data_expand): ...new pass_data instance and...
13324 (make_pass_expand): ...new function.
13325 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
13326 of rtl_opt_pass along with...
13327 (pass_data_free_cfg): ...new pass_data instance and...
13328 (make_pass_free_cfg): ...new function.
13329 (pass_into_cfg_layout_mode): Convert from a global struct to a
13330 subclass of rtl_opt_pass along with...
13331 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
13332 (make_pass_into_cfg_layout_mode): ...new function.
13333 (pass_outof_cfg_layout_mode): Convert from a global struct to a
13334 subclass of rtl_opt_pass along with...
13335 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
13336 (make_pass_outof_cfg_layout_mode): ...new function.
13337 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
13338 struct to a subclass of gimple_opt_pass along with...
13339 (pass_data_build_cgraph_edges): ...new pass_data instance and...
13340 (make_pass_build_cgraph_edges): ...new function.
13341 (pass_rebuild_cgraph_edges): Convert from a global struct to a
13342 subclass of gimple_opt_pass along with...
13343 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
13344 (make_pass_rebuild_cgraph_edges): ...new function.
13345 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
13346 subclass of gimple_opt_pass along with...
13347 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
13349 (make_pass_remove_cgraph_callee_edges): ...new function.
13350 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
13351 struct to a subclass of rtl_opt_pass along with...
13352 (pass_data_stack_adjustments): ...new pass_data instance and...
13353 (make_pass_stack_adjustments): ...new function.
13354 * combine.c (pass_combine): Convert from a global struct to a subclass
13355 of rtl_opt_pass along with...
13356 (pass_data_combine): ...new pass_data instance and...
13357 (make_pass_combine): ...new function.
13358 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
13359 global struct to a subclass of rtl_opt_pass along with...
13360 (pass_data_compare_elim_after_reload): ...new pass_data instance
13362 (make_pass_compare_elim_after_reload): ...new function.
13363 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
13364 of rtl_opt_pass along with...
13365 (pass_data_rtl_cprop): ...new pass_data instance and...
13366 (make_pass_rtl_cprop): ...new function.
13367 * cse.c (pass_cse): Convert from a global struct to a subclass of
13368 rtl_opt_pass along with...
13369 (pass_data_cse): ...new pass_data instance and...
13370 (make_pass_cse): ...new function.
13371 (pass_cse2): Convert from a global struct to a subclass of
13372 rtl_opt_pass along with...
13373 (pass_data_cse2): ...new pass_data instance and...
13374 (make_pass_cse2): ...new function.
13375 (pass_cse_after_global_opts): Convert from a global struct to a
13376 subclass of rtl_opt_pass along with...
13377 (pass_data_cse_after_global_opts): ...new pass_data instance and...
13378 (make_pass_cse_after_global_opts): ...new function.
13379 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
13380 of rtl_opt_pass along with...
13381 (pass_data_ud_rtl_dce): ...new pass_data instance and...
13382 (make_pass_ud_rtl_dce): ...new function.
13383 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
13384 rtl_opt_pass along with...
13385 (pass_data_fast_rtl_dce): ...new pass_data instance and...
13386 (make_pass_fast_rtl_dce): ...new function.
13387 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
13388 a subclass of rtl_opt_pass along with...
13389 (pass_data_df_initialize_opt): ...new pass_data instance and...
13390 (make_pass_df_initialize_opt): ...new function.
13391 (pass_df_initialize_no_opt): Convert from a global struct to a
13392 subclass of rtl_opt_pass along with...
13393 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
13394 (make_pass_df_initialize_no_opt): ...new function.
13395 (pass_df_finish): Convert from a global struct to a subclass of
13396 rtl_opt_pass along with...
13397 (pass_data_df_finish): ...new pass_data instance and...
13398 (make_pass_df_finish): ...new function.
13399 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
13400 rtl_opt_pass along with...
13401 (pass_data_rtl_dse1): ...new pass_data instance and...
13402 (make_pass_rtl_dse1): ...new function.
13403 (pass_rtl_dse2): Convert from a global struct to a subclass of
13404 rtl_opt_pass along with...
13405 (pass_data_rtl_dse2): ...new pass_data instance and...
13406 (make_pass_rtl_dse2): ...new function.
13407 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
13408 subclass of rtl_opt_pass along with...
13409 (pass_data_dwarf2_frame): ...new pass_data instance and...
13410 (make_pass_dwarf2_frame): ...new function.
13411 * except.c (pass_set_nothrow_function_flags): Convert from a global
13412 struct to a subclass of rtl_opt_pass along with...
13413 (pass_data_set_nothrow_function_flags): ...new pass_data instance
13415 (make_pass_set_nothrow_function_flags): ...new function.
13416 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
13417 subclass of rtl_opt_pass along with...
13418 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
13420 (make_pass_convert_to_eh_region_ranges): ...new function.
13421 * final.c (pass_compute_alignments): Convert from a global struct to a
13422 subclass of rtl_opt_pass along with...
13423 (pass_data_compute_alignments): ...new pass_data instance and...
13424 (make_pass_compute_alignments): ...new function.
13425 (pass_final): Convert from a global struct to a subclass of
13426 rtl_opt_pass along with...
13427 (pass_data_final): ...new pass_data instance and...
13428 (make_pass_final): ...new function.
13429 (pass_shorten_branches): Convert from a global struct to a subclass of
13430 rtl_opt_pass along with...
13431 (pass_data_shorten_branches): ...new pass_data instance and...
13432 (make_pass_shorten_branches): ...new function.
13433 (pass_clean_state): Convert from a global struct to a subclass of
13434 rtl_opt_pass along with...
13435 (pass_data_clean_state): ...new pass_data instance and...
13436 (make_pass_clean_state): ...new function.
13437 * function.c (pass_instantiate_virtual_regs): Convert from a global
13438 struct to a subclass of rtl_opt_pass along with...
13439 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
13440 (make_pass_instantiate_virtual_regs): ...new function.
13441 (pass_leaf_regs): Convert from a global struct to a subclass of
13442 rtl_opt_pass along with...
13443 (pass_data_leaf_regs): ...new pass_data instance and...
13444 (make_pass_leaf_regs): ...new function.
13445 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
13446 subclass of rtl_opt_pass along with...
13447 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
13449 (make_pass_thread_prologue_and_epilogue): ...new function.
13450 (pass_match_asm_constraints): Convert from a global struct to a
13451 subclass of rtl_opt_pass along with...
13452 (pass_data_match_asm_constraints): ...new pass_data instance and...
13453 (make_pass_match_asm_constraints): ...new function.
13454 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
13455 subclass of rtl_opt_pass along with...
13456 (pass_data_rtl_fwprop): ...new pass_data instance and...
13457 (make_pass_rtl_fwprop): ...new function.
13458 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
13459 rtl_opt_pass along with...
13460 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
13461 (make_pass_rtl_fwprop_addr): ...new function.
13462 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
13463 rtl_opt_pass along with...
13464 (pass_data_rtl_pre): ...new pass_data instance and...
13465 (make_pass_rtl_pre): ...new function.
13466 (pass_rtl_hoist): Convert from a global struct to a subclass of
13467 rtl_opt_pass along with...
13468 (pass_data_rtl_hoist): ...new pass_data instance and...
13469 (make_pass_rtl_hoist): ...new function.
13470 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
13471 subclass of gimple_opt_pass along with...
13472 (pass_data_lower_cf): ...new pass_data instance and...
13473 (make_pass_lower_cf): ...new function.
13474 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
13475 from a global struct to a subclass of gimple_opt_pass along with...
13476 (pass_data_strength_reduction): ...new pass_data instance and...
13477 (make_pass_strength_reduction): ...new function.
13478 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
13479 of rtl_opt_pass along with...
13480 (pass_data_rtl_ifcvt): ...new pass_data instance and...
13481 (make_pass_rtl_ifcvt): ...new function.
13482 (pass_if_after_combine): Convert from a global struct to a subclass of
13483 rtl_opt_pass along with...
13484 (pass_data_if_after_combine): ...new pass_data instance and...
13485 (make_pass_if_after_combine): ...new function.
13486 (pass_if_after_reload): Convert from a global struct to a subclass of
13487 rtl_opt_pass along with...
13488 (pass_data_if_after_reload): ...new pass_data instance and...
13489 (make_pass_if_after_reload): ...new function.
13490 * init-regs.c (pass_initialize_regs): Convert from a global struct to
13491 a subclass of rtl_opt_pass along with...
13492 (pass_data_initialize_regs): ...new pass_data instance and...
13493 (make_pass_initialize_regs): ...new function.
13494 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
13495 of ipa_opt_pass_d along with...
13496 (pass_data_ipa_cp): ...new pass_data instance and...
13497 (make_pass_ipa_cp): ...new function.
13498 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
13499 global struct to a subclass of gimple_opt_pass along with...
13500 (pass_data_inline_parameters): ...new pass_data instance and...
13501 (make_pass_inline_parameters): ...new function.
13502 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
13503 subclass of gimple_opt_pass along with...
13504 (pass_data_early_inline): ...new pass_data instance and...
13505 (make_pass_early_inline): ...new function.
13506 (pass_ipa_inline): Convert from a global struct to a subclass of
13507 ipa_opt_pass_d along with...
13508 (pass_data_ipa_inline): ...new pass_data instance and...
13509 (make_pass_ipa_inline): ...new function.
13510 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
13511 struct to a subclass of gimple_opt_pass along with...
13512 (pass_data_local_pure_const): ...new pass_data instance and...
13513 (make_pass_local_pure_const): ...new function.
13514 (pass_ipa_pure_const): Convert from a global struct to a subclass of
13515 ipa_opt_pass_d along with...
13516 (pass_data_ipa_pure_const): ...new pass_data instance and...
13517 (make_pass_ipa_pure_const): ...new function.
13518 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
13519 to a subclass of ipa_opt_pass_d along with...
13520 (pass_data_ipa_reference): ...new pass_data instance and...
13521 (make_pass_ipa_reference): ...new function.
13522 * ipa-split.c (pass_split_functions): Convert from a global struct to
13523 a subclass of gimple_opt_pass along with...
13524 (pass_data_split_functions): ...new pass_data instance and...
13525 (make_pass_split_functions): ...new function.
13526 (pass_feedback_split_functions): Convert from a global struct to a
13527 subclass of gimple_opt_pass along with...
13528 (pass_data_feedback_split_functions): ...new pass_data instance and...
13529 (make_pass_feedback_split_functions): ...new function.
13530 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
13531 global struct to a subclass of simple_ipa_opt_pass along with...
13532 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
13534 (make_pass_ipa_function_and_variable_visibility): ...new function.
13535 (pass_ipa_free_inline_summary): Convert from a global struct to a
13536 subclass of simple_ipa_opt_pass along with...
13537 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
13538 (make_pass_ipa_free_inline_summary): ...new function.
13539 (pass_ipa_whole_program_visibility): Convert from a global struct to a
13540 subclass of ipa_opt_pass_d along with...
13541 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
13543 (make_pass_ipa_whole_program_visibility): ...new function.
13544 (pass_ipa_profile): Convert from a global struct to a subclass of
13545 ipa_opt_pass_d along with...
13546 (pass_data_ipa_profile): ...new pass_data instance and...
13547 (make_pass_ipa_profile): ...new function.
13548 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
13549 ipa_opt_pass_d along with...
13550 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
13551 (make_pass_ipa_cdtor_merge): ...new function.
13552 * ira.c (pass_ira): Convert from a global struct to a subclass of
13553 rtl_opt_pass along with...
13554 (pass_data_ira): ...new pass_data instance and...
13555 (make_pass_ira): ...new function.
13556 (pass_reload): Convert from a global struct to a subclass of
13557 rtl_opt_pass along with...
13558 (pass_data_reload): ...new pass_data instance and...
13559 (make_pass_reload): ...new function.
13560 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
13561 subclass of rtl_opt_pass along with...
13562 (pass_data_cleanup_barriers): ...new pass_data instance and...
13563 (make_pass_cleanup_barriers): ...new function.
13564 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
13565 of rtl_opt_pass along with...
13566 (pass_data_loop2): ...new pass_data instance and...
13567 (make_pass_loop2): ...new function.
13568 (pass_rtl_loop_init): Convert from a global struct to a subclass of
13569 rtl_opt_pass along with...
13570 (pass_data_rtl_loop_init): ...new pass_data instance and...
13571 (make_pass_rtl_loop_init): ...new function.
13572 (pass_rtl_loop_done): Convert from a global struct to a subclass of
13573 rtl_opt_pass along with...
13574 (pass_data_rtl_loop_done): ...new pass_data instance and...
13575 (make_pass_rtl_loop_done): ...new function.
13576 (pass_rtl_move_loop_invariants): Convert from a global struct to a
13577 subclass of rtl_opt_pass along with...
13578 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
13579 (make_pass_rtl_move_loop_invariants): ...new function.
13580 (pass_rtl_unswitch): Convert from a global struct to a subclass of
13581 rtl_opt_pass along with...
13582 (pass_data_rtl_unswitch): ...new pass_data instance and...
13583 (make_pass_rtl_unswitch): ...new function.
13584 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
13585 subclass of rtl_opt_pass along with...
13586 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
13588 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13589 (pass_rtl_doloop): Convert from a global struct to a subclass of
13590 rtl_opt_pass along with...
13591 (pass_data_rtl_doloop): ...new pass_data instance and...
13592 (make_pass_rtl_doloop): ...new function.
13593 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
13594 a subclass of rtl_opt_pass along with...
13595 (pass_data_lower_subreg): ...new pass_data instance and...
13596 (make_pass_lower_subreg): ...new function.
13597 (pass_lower_subreg2): Convert from a global struct to a subclass of
13598 rtl_opt_pass along with...
13599 (pass_data_lower_subreg2): ...new pass_data instance and...
13600 (make_pass_lower_subreg2): ...new function.
13601 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
13602 struct to a subclass of ipa_opt_pass_d along with...
13603 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
13604 (make_pass_ipa_lto_gimple_out): ...new function.
13605 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
13606 of ipa_opt_pass_d along with...
13607 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
13608 (make_pass_ipa_lto_finish_out): ...new function.
13609 * mode-switching.c (pass_mode_switching): Convert from a global struct
13610 to a subclass of rtl_opt_pass along with...
13611 (pass_data_mode_switching): ...new pass_data instance and...
13612 (make_pass_mode_switching): ...new function.
13613 * modulo-sched.c (pass_sms): Convert from a global struct to a
13614 subclass of rtl_opt_pass along with...
13615 (pass_data_sms): ...new pass_data instance and...
13616 (make_pass_sms): ...new function.
13617 * omp-low.c (pass_expand_omp): Convert from a global struct to a
13618 subclass of gimple_opt_pass along with...
13619 (pass_data_expand_omp): ...new pass_data instance and...
13620 (make_pass_expand_omp): ...new function.
13621 (pass_lower_omp): Convert from a global struct to a subclass of
13622 gimple_opt_pass along with...
13623 (pass_data_lower_omp): ...new pass_data instance and...
13624 (make_pass_lower_omp): ...new function.
13625 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
13626 of gimple_opt_pass along with...
13627 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
13628 (make_pass_diagnose_omp_blocks): ...new function.
13629 * passes.c (pass_early_local_passes): Convert from a global struct to
13630 a subclass of simple_ipa_opt_pass along with...
13631 (pass_data_early_local_passes): ...new pass_data instance and...
13632 (make_pass_early_local_passes): ...new function.
13633 (pass_all_early_optimizations): Convert from a global struct to a
13634 subclass of gimple_opt_pass along with...
13635 (pass_data_all_early_optimizations): ...new pass_data instance and...
13636 (make_pass_all_early_optimizations): ...new function.
13637 (pass_all_optimizations): Convert from a global struct to a subclass
13638 of gimple_opt_pass along with...
13639 (pass_data_all_optimizations): ...new pass_data instance and...
13640 (make_pass_all_optimizations): ...new function.
13641 (pass_all_optimizations_g): Convert from a global struct to a subclass
13642 of gimple_opt_pass along with...
13643 (pass_data_all_optimizations_g): ...new pass_data instance and...
13644 (make_pass_all_optimizations_g): ...new function.
13645 (pass_rest_of_compilation): Convert from a global struct to a subclass
13646 of rtl_opt_pass along with...
13647 (pass_data_rest_of_compilation): ...new pass_data instance and...
13648 (make_pass_rest_of_compilation): ...new function.
13649 (pass_postreload): Convert from a global struct to a subclass of
13650 rtl_opt_pass along with...
13651 (pass_data_postreload): ...new pass_data instance and...
13652 (make_pass_postreload): ...new function.
13653 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
13654 subclass of rtl_opt_pass along with...
13655 (pass_data_gcse2): ...new pass_data instance and...
13656 (make_pass_gcse2): ...new function.
13657 * postreload.c (pass_postreload_cse): Convert from a global struct to
13658 a subclass of rtl_opt_pass along with...
13659 (pass_data_postreload_cse): ...new pass_data instance and...
13660 (make_pass_postreload_cse): ...new function.
13661 * predict.c (pass_profile): Convert from a global struct to a subclass
13662 of gimple_opt_pass along with...
13663 (pass_data_profile): ...new pass_data instance and...
13664 (make_pass_profile): ...new function.
13665 (pass_strip_predict_hints): Convert from a global struct to a subclass
13666 of gimple_opt_pass along with...
13667 (pass_data_strip_predict_hints): ...new pass_data instance and...
13668 (make_pass_strip_predict_hints): ...new function.
13669 * recog.c (pass_peephole2): Convert from a global struct to a subclass
13670 of rtl_opt_pass along with...
13671 (pass_data_peephole2): ...new pass_data instance and...
13672 (make_pass_peephole2): ...new function.
13673 (pass_split_all_insns): Convert from a global struct to a subclass of
13674 rtl_opt_pass along with...
13675 (pass_data_split_all_insns): ...new pass_data instance and...
13676 (make_pass_split_all_insns): ...new function.
13677 (pass_split_after_reload): Convert from a global struct to a subclass
13678 of rtl_opt_pass along with...
13679 (pass_data_split_after_reload): ...new pass_data instance and...
13680 (make_pass_split_after_reload): ...new function.
13681 (pass_split_before_regstack): Convert from a global struct to a
13682 subclass of rtl_opt_pass along with...
13683 (pass_data_split_before_regstack): ...new pass_data instance and...
13684 (make_pass_split_before_regstack): ...new function.
13685 (pass_split_before_sched2): Convert from a global struct to a subclass
13686 of rtl_opt_pass along with...
13687 (pass_data_split_before_sched2): ...new pass_data instance and...
13688 (make_pass_split_before_sched2): ...new function.
13689 (pass_split_for_shorten_branches): Convert from a global struct to a
13690 subclass of rtl_opt_pass along with...
13691 (pass_data_split_for_shorten_branches): ...new pass_data instance
13693 (make_pass_split_for_shorten_branches): ...new function.
13694 * ree.c (pass_ree): Convert from a global struct to a subclass of
13695 rtl_opt_pass along with...
13696 (pass_data_ree): ...new pass_data instance and...
13697 (make_pass_ree): ...new function.
13698 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
13699 subclass of rtl_opt_pass along with...
13700 (pass_data_stack_regs): ...new pass_data instance and...
13701 (make_pass_stack_regs): ...new function.
13702 (pass_stack_regs_run): Convert from a global struct to a subclass of
13703 rtl_opt_pass along with...
13704 (pass_data_stack_regs_run): ...new pass_data instance and...
13705 (make_pass_stack_regs_run): ...new function.
13706 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
13707 subclass of rtl_opt_pass along with...
13708 (pass_data_cprop_hardreg): ...new pass_data instance and...
13709 (make_pass_cprop_hardreg): ...new function.
13710 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
13711 subclass of rtl_opt_pass along with...
13712 (pass_data_reginfo_init): ...new pass_data instance and...
13713 (make_pass_reginfo_init): ...new function.
13714 * regmove.c (pass_regmove): Convert from a global struct to a subclass
13715 of rtl_opt_pass along with...
13716 (pass_data_regmove): ...new pass_data instance and...
13717 (make_pass_regmove): ...new function.
13718 * regrename.c (pass_regrename): Convert from a global struct to a
13719 subclass of rtl_opt_pass along with...
13720 (pass_data_regrename): ...new pass_data instance and...
13721 (make_pass_regrename): ...new function.
13722 * reorg.c (pass_delay_slots): Convert from a global struct to a
13723 subclass of rtl_opt_pass along with...
13724 (pass_data_delay_slots): ...new pass_data instance and...
13725 (make_pass_delay_slots): ...new function.
13726 (pass_machine_reorg): Convert from a global struct to a subclass of
13727 rtl_opt_pass along with...
13728 (pass_data_machine_reorg): ...new pass_data instance and...
13729 (make_pass_machine_reorg): ...new function.
13730 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
13731 of rtl_opt_pass along with...
13732 (pass_data_sched): ...new pass_data instance and...
13733 (make_pass_sched): ...new function.
13734 (pass_sched2): Convert from a global struct to a subclass of
13735 rtl_opt_pass along with...
13736 (pass_data_sched2): ...new pass_data instance and...
13737 (make_pass_sched2): ...new function.
13738 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
13739 to a subclass of rtl_opt_pass along with...
13740 (pass_data_stack_ptr_mod): ...new pass_data instance and...
13741 (make_pass_stack_ptr_mod): ...new function.
13742 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
13743 to a subclass of rtl_opt_pass along with...
13744 (pass_data_rtl_store_motion): ...new pass_data instance and...
13745 (make_pass_rtl_store_motion): ...new function.
13746 * tracer.c (pass_tracer): Convert from a global struct to a subclass
13747 of gimple_opt_pass along with...
13748 (pass_data_tracer): ...new pass_data instance and...
13749 (make_pass_tracer): ...new function.
13750 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
13751 to a subclass of gimple_opt_pass along with...
13752 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
13753 (make_pass_diagnose_tm_blocks): ...new function.
13754 (pass_lower_tm): Convert from a global struct to a subclass of
13755 gimple_opt_pass along with...
13756 (pass_data_lower_tm): ...new pass_data instance and...
13757 (make_pass_lower_tm): ...new function.
13758 (pass_tm_init): Convert from a global struct to a subclass of
13759 gimple_opt_pass along with...
13760 (pass_data_tm_init): ...new pass_data instance and...
13761 (make_pass_tm_init): ...new function.
13762 (pass_tm_mark): Convert from a global struct to a subclass of
13763 gimple_opt_pass along with...
13764 (pass_data_tm_mark): ...new pass_data instance and...
13765 (make_pass_tm_mark): ...new function.
13766 (pass_tm_edges): Convert from a global struct to a subclass of
13767 gimple_opt_pass along with...
13768 (pass_data_tm_edges): ...new pass_data instance and...
13769 (make_pass_tm_edges): ...new function.
13770 (pass_tm_memopt): Convert from a global struct to a subclass of
13771 gimple_opt_pass along with...
13772 (pass_data_tm_memopt): ...new pass_data instance and...
13773 (make_pass_tm_memopt): ...new function.
13774 (pass_ipa_tm): Convert from a global struct to a subclass of
13775 simple_ipa_opt_pass along with...
13776 (pass_data_ipa_tm): ...new pass_data instance and...
13777 (make_pass_ipa_tm): ...new function.
13778 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
13779 subclass of gimple_opt_pass along with...
13780 (pass_data_call_cdce): ...new pass_data instance and...
13781 (make_pass_call_cdce): ...new function.
13782 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
13783 subclass of gimple_opt_pass along with...
13784 (pass_data_build_cfg): ...new pass_data instance and...
13785 (make_pass_build_cfg): ...new function.
13786 (pass_split_crit_edges): Convert from a global struct to a subclass of
13787 gimple_opt_pass along with...
13788 (pass_data_split_crit_edges): ...new pass_data instance and...
13789 (make_pass_split_crit_edges): ...new function.
13790 (pass_warn_function_return): Convert from a global struct to a
13791 subclass of gimple_opt_pass along with...
13792 (pass_data_warn_function_return): ...new pass_data instance and...
13793 (make_pass_warn_function_return): ...new function.
13794 (pass_warn_function_noreturn): Convert from a global struct to a
13795 subclass of gimple_opt_pass along with...
13796 (pass_data_warn_function_noreturn): ...new pass_data instance and...
13797 (make_pass_warn_function_noreturn): ...new function.
13798 (pass_warn_unused_result): Convert from a global struct to a subclass
13799 of gimple_opt_pass along with...
13800 (pass_data_warn_unused_result): ...new pass_data instance and...
13801 (make_pass_warn_unused_result): ...new function.
13802 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
13803 a subclass of gimple_opt_pass along with...
13804 (pass_data_merge_phi): ...new pass_data instance and...
13805 (make_pass_merge_phi): ...new function.
13806 * tree-complex.c (pass_lower_complex): Convert from a global struct to
13807 a subclass of gimple_opt_pass along with...
13808 (pass_data_lower_complex): ...new pass_data instance and...
13809 (make_pass_lower_complex): ...new function.
13810 (pass_lower_complex_O0): Convert from a global struct to a subclass of
13811 gimple_opt_pass along with...
13812 (pass_data_lower_complex_O0): ...new pass_data instance and...
13813 (make_pass_lower_complex_O0): ...new function.
13814 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
13815 subclass of gimple_opt_pass along with...
13816 (pass_data_lower_eh): ...new pass_data instance and...
13817 (make_pass_lower_eh): ...new function.
13818 (pass_refactor_eh): Convert from a global struct to a subclass of
13819 gimple_opt_pass along with...
13820 (pass_data_refactor_eh): ...new pass_data instance and...
13821 (make_pass_refactor_eh): ...new function.
13822 (pass_lower_resx): Convert from a global struct to a subclass of
13823 gimple_opt_pass along with...
13824 (pass_data_lower_resx): ...new pass_data instance and...
13825 (make_pass_lower_resx): ...new function.
13826 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
13827 of gimple_opt_pass along with...
13828 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
13829 (make_pass_lower_eh_dispatch): ...new function.
13830 (pass_cleanup_eh): Convert from a global struct to a subclass of
13831 gimple_opt_pass along with...
13832 (pass_data_cleanup_eh): ...new pass_data instance and...
13833 (make_pass_cleanup_eh): ...new function.
13834 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
13835 to a subclass of simple_ipa_opt_pass along with...
13836 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
13837 (make_pass_ipa_lower_emutls): ...new function.
13838 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
13839 a subclass of gimple_opt_pass along with...
13840 (pass_data_if_conversion): ...new pass_data instance and...
13841 (make_pass_if_conversion): ...new function.
13842 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
13843 subclass of gimple_opt_pass along with...
13844 (pass_data_build_ssa): ...new pass_data instance and...
13845 (make_pass_build_ssa): ...new function.
13846 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
13847 global struct to a subclass of gimple_opt_pass along with...
13848 (pass_data_loop_distribution): ...new pass_data instance and...
13849 (make_pass_loop_distribution): ...new function.
13850 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13851 subclass of gimple_opt_pass along with...
13852 (pass_data_mudflap_1): ...new pass_data instance and...
13853 (make_pass_mudflap_1): ...new function.
13854 (pass_mudflap_2): Convert from a global struct to a subclass of
13855 gimple_opt_pass along with...
13856 (pass_data_mudflap_2): ...new pass_data instance and...
13857 (make_pass_mudflap_2): ...new function.
13858 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13859 subclass of gimple_opt_pass along with...
13860 (pass_data_mudflap_1): ...new pass_data instance and...
13861 (make_pass_mudflap_1): ...new function.
13862 (pass_mudflap_2): Convert from a global struct to a subclass of
13863 gimple_opt_pass along with...
13864 (pass_data_mudflap_2): ...new pass_data instance and...
13865 (make_pass_mudflap_2): ...new function.
13866 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13867 gimple_opt_pass along with...
13868 (pass_data_nrv): ...new pass_data instance and...
13869 (make_pass_nrv): ...new function.
13870 (pass_return_slot): Convert from a global struct to a subclass of
13871 gimple_opt_pass along with...
13872 (pass_data_return_slot): ...new pass_data instance and...
13873 (make_pass_return_slot): ...new function.
13874 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13875 to a subclass of gimple_opt_pass along with...
13876 (pass_data_object_sizes): ...new pass_data instance and...
13877 (make_pass_object_sizes): ...new function.
13878 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13879 global struct to a subclass of gimple_opt_pass along with...
13880 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13882 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13883 (pass_fixup_cfg): Convert from a global struct to a subclass of
13884 gimple_opt_pass along with...
13885 (pass_data_fixup_cfg): ...new pass_data instance and...
13886 (make_pass_fixup_cfg): ...new function.
13887 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13888 (make_pass_mudflap_1): ...new function.
13889 (pass_mudflap_2): Replace declaration with that of...
13890 (make_pass_mudflap_2): ...new function.
13891 (pass_asan): Replace declaration with that of...
13892 (make_pass_asan): ...new function.
13893 (pass_asan_O0): Replace declaration with that of...
13894 (make_pass_asan_O0): ...new function.
13895 (pass_tsan): Replace declaration with that of...
13896 (make_pass_tsan): ...new function.
13897 (pass_tsan_O0): Replace declaration with that of...
13898 (make_pass_tsan_O0): ...new function.
13899 (pass_lower_cf): Replace declaration with that of...
13900 (make_pass_lower_cf): ...new function.
13901 (pass_refactor_eh): Replace declaration with that of...
13902 (make_pass_refactor_eh): ...new function.
13903 (pass_lower_eh): Replace declaration with that of...
13904 (make_pass_lower_eh): ...new function.
13905 (pass_lower_eh_dispatch): Replace declaration with that of...
13906 (make_pass_lower_eh_dispatch): ...new function.
13907 (pass_lower_resx): Replace declaration with that of...
13908 (make_pass_lower_resx): ...new function.
13909 (pass_build_cfg): Replace declaration with that of...
13910 (make_pass_build_cfg): ...new function.
13911 (pass_early_tree_profile): Replace declaration with that of...
13912 (make_pass_early_tree_profile): ...new function.
13913 (pass_cleanup_eh): Replace declaration with that of...
13914 (make_pass_cleanup_eh): ...new function.
13915 (pass_sra): Replace declaration with that of...
13916 (make_pass_sra): ...new function.
13917 (pass_sra_early): Replace declaration with that of...
13918 (make_pass_sra_early): ...new function.
13919 (pass_early_ipa_sra): Replace declaration with that of...
13920 (make_pass_early_ipa_sra): ...new function.
13921 (pass_tail_recursion): Replace declaration with that of...
13922 (make_pass_tail_recursion): ...new function.
13923 (pass_tail_calls): Replace declaration with that of...
13924 (make_pass_tail_calls): ...new function.
13925 (pass_tree_loop): Replace declaration with that of...
13926 (make_pass_tree_loop): ...new function.
13927 (pass_tree_loop_init): Replace declaration with that of...
13928 (make_pass_tree_loop_init): ...new function.
13929 (pass_lim): Replace declaration with that of...
13930 (make_pass_lim): ...new function.
13931 (pass_tree_unswitch): Replace declaration with that of...
13932 (make_pass_tree_unswitch): ...new function.
13933 (pass_predcom): Replace declaration with that of...
13934 (make_pass_predcom): ...new function.
13935 (pass_iv_canon): Replace declaration with that of...
13936 (make_pass_iv_canon): ...new function.
13937 (pass_scev_cprop): Replace declaration with that of...
13938 (make_pass_scev_cprop): ...new function.
13939 (pass_empty_loop): Replace declaration with that of...
13940 (make_pass_empty_loop): ...new function.
13941 (pass_record_bounds): Replace declaration with that of...
13942 (make_pass_record_bounds): ...new function.
13943 (pass_graphite): Replace declaration with that of...
13944 (make_pass_graphite): ...new function.
13945 (pass_graphite_transforms): Replace declaration with that of...
13946 (make_pass_graphite_transforms): ...new function.
13947 (pass_if_conversion): Replace declaration with that of...
13948 (make_pass_if_conversion): ...new function.
13949 (pass_loop_distribution): Replace declaration with that of...
13950 (make_pass_loop_distribution): ...new function.
13951 (pass_vectorize): Replace declaration with that of...
13952 (make_pass_vectorize): ...new function.
13953 (pass_slp_vectorize): Replace declaration with that of...
13954 (make_pass_slp_vectorize): ...new function.
13955 (pass_complete_unroll): Replace declaration with that of...
13956 (make_pass_complete_unroll): ...new function.
13957 (pass_complete_unrolli): Replace declaration with that of...
13958 (make_pass_complete_unrolli): ...new function.
13959 (pass_parallelize_loops): Replace declaration with that of...
13960 (make_pass_parallelize_loops): ...new function.
13961 (pass_loop_prefetch): Replace declaration with that of...
13962 (make_pass_loop_prefetch): ...new function.
13963 (pass_iv_optimize): Replace declaration with that of...
13964 (make_pass_iv_optimize): ...new function.
13965 (pass_tree_loop_done): Replace declaration with that of...
13966 (make_pass_tree_loop_done): ...new function.
13967 (pass_ch): Replace declaration with that of...
13968 (make_pass_ch): ...new function.
13969 (pass_ccp): Replace declaration with that of...
13970 (make_pass_ccp): ...new function.
13971 (pass_phi_only_cprop): Replace declaration with that of...
13972 (make_pass_phi_only_cprop): ...new function.
13973 (pass_build_ssa): Replace declaration with that of...
13974 (make_pass_build_ssa): ...new function.
13975 (pass_build_alias): Replace declaration with that of...
13976 (make_pass_build_alias): ...new function.
13977 (pass_build_ealias): Replace declaration with that of...
13978 (make_pass_build_ealias): ...new function.
13979 (pass_dominator): Replace declaration with that of...
13980 (make_pass_dominator): ...new function.
13981 (pass_dce): Replace declaration with that of...
13982 (make_pass_dce): ...new function.
13983 (pass_dce_loop): Replace declaration with that of...
13984 (make_pass_dce_loop): ...new function.
13985 (pass_cd_dce): Replace declaration with that of...
13986 (make_pass_cd_dce): ...new function.
13987 (pass_call_cdce): Replace declaration with that of...
13988 (make_pass_call_cdce): ...new function.
13989 (pass_merge_phi): Replace declaration with that of...
13990 (make_pass_merge_phi): ...new function.
13991 (pass_split_crit_edges): Replace declaration with that of...
13992 (make_pass_split_crit_edges): ...new function.
13993 (pass_pre): Replace declaration with that of...
13994 (make_pass_pre): ...new function.
13995 (pass_profile): Replace declaration with that of...
13996 (make_pass_profile): ...new function.
13997 (pass_strip_predict_hints): Replace declaration with that of...
13998 (make_pass_strip_predict_hints): ...new function.
13999 (pass_lower_complex_O0): Replace declaration with that of...
14000 (make_pass_lower_complex_O0): ...new function.
14001 (pass_lower_complex): Replace declaration with that of...
14002 (make_pass_lower_complex): ...new function.
14003 (pass_lower_vector): Replace declaration with that of...
14004 (make_pass_lower_vector): ...new function.
14005 (pass_lower_vector_ssa): Replace declaration with that of...
14006 (make_pass_lower_vector_ssa): ...new function.
14007 (pass_lower_omp): Replace declaration with that of...
14008 (make_pass_lower_omp): ...new function.
14009 (pass_diagnose_omp_blocks): Replace declaration with that of...
14010 (make_pass_diagnose_omp_blocks): ...new function.
14011 (pass_expand_omp): Replace declaration with that of...
14012 (make_pass_expand_omp): ...new function.
14013 (pass_expand_omp_ssa): Replace declaration with that of...
14014 (make_pass_expand_omp_ssa): ...new function.
14015 (pass_object_sizes): Replace declaration with that of...
14016 (make_pass_object_sizes): ...new function.
14017 (pass_strlen): Replace declaration with that of...
14018 (make_pass_strlen): ...new function.
14019 (pass_fold_builtins): Replace declaration with that of...
14020 (make_pass_fold_builtins): ...new function.
14021 (pass_stdarg): Replace declaration with that of...
14022 (make_pass_stdarg): ...new function.
14023 (pass_early_warn_uninitialized): Replace declaration with that of...
14024 (make_pass_early_warn_uninitialized): ...new function.
14025 (pass_late_warn_uninitialized): Replace declaration with that of...
14026 (make_pass_late_warn_uninitialized): ...new function.
14027 (pass_cse_reciprocals): Replace declaration with that of...
14028 (make_pass_cse_reciprocals): ...new function.
14029 (pass_cse_sincos): Replace declaration with that of...
14030 (make_pass_cse_sincos): ...new function.
14031 (pass_optimize_bswap): Replace declaration with that of...
14032 (make_pass_optimize_bswap): ...new function.
14033 (pass_optimize_widening_mul): Replace declaration with that of...
14034 (make_pass_optimize_widening_mul): ...new function.
14035 (pass_warn_function_return): Replace declaration with that of...
14036 (make_pass_warn_function_return): ...new function.
14037 (pass_warn_function_noreturn): Replace declaration with that of...
14038 (make_pass_warn_function_noreturn): ...new function.
14039 (pass_cselim): Replace declaration with that of...
14040 (make_pass_cselim): ...new function.
14041 (pass_phiopt): Replace declaration with that of...
14042 (make_pass_phiopt): ...new function.
14043 (pass_forwprop): Replace declaration with that of...
14044 (make_pass_forwprop): ...new function.
14045 (pass_phiprop): Replace declaration with that of...
14046 (make_pass_phiprop): ...new function.
14047 (pass_tree_ifcombine): Replace declaration with that of...
14048 (make_pass_tree_ifcombine): ...new function.
14049 (pass_dse): Replace declaration with that of...
14050 (make_pass_dse): ...new function.
14051 (pass_nrv): Replace declaration with that of...
14052 (make_pass_nrv): ...new function.
14053 (pass_rename_ssa_copies): Replace declaration with that of...
14054 (make_pass_rename_ssa_copies): ...new function.
14055 (pass_sink_code): Replace declaration with that of...
14056 (make_pass_sink_code): ...new function.
14057 (pass_fre): Replace declaration with that of...
14058 (make_pass_fre): ...new function.
14059 (pass_check_data_deps): Replace declaration with that of...
14060 (make_pass_check_data_deps): ...new function.
14061 (pass_copy_prop): Replace declaration with that of...
14062 (make_pass_copy_prop): ...new function.
14063 (pass_vrp): Replace declaration with that of...
14064 (make_pass_vrp): ...new function.
14065 (pass_uncprop): Replace declaration with that of...
14066 (make_pass_uncprop): ...new function.
14067 (pass_return_slot): Replace declaration with that of...
14068 (make_pass_return_slot): ...new function.
14069 (pass_reassoc): Replace declaration with that of...
14070 (make_pass_reassoc): ...new function.
14071 (pass_rebuild_cgraph_edges): Replace declaration with that of...
14072 (make_pass_rebuild_cgraph_edges): ...new function.
14073 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
14074 (make_pass_remove_cgraph_callee_edges): ...new function.
14075 (pass_build_cgraph_edges): Replace declaration with that of...
14076 (make_pass_build_cgraph_edges): ...new function.
14077 (pass_local_pure_const): Replace declaration with that of...
14078 (make_pass_local_pure_const): ...new function.
14079 (pass_tracer): Replace declaration with that of...
14080 (make_pass_tracer): ...new function.
14081 (pass_warn_unused_result): Replace declaration with that of...
14082 (make_pass_warn_unused_result): ...new function.
14083 (pass_diagnose_tm_blocks): Replace declaration with that of...
14084 (make_pass_diagnose_tm_blocks): ...new function.
14085 (pass_lower_tm): Replace declaration with that of...
14086 (make_pass_lower_tm): ...new function.
14087 (pass_tm_init): Replace declaration with that of...
14088 (make_pass_tm_init): ...new function.
14089 (pass_tm_mark): Replace declaration with that of...
14090 (make_pass_tm_mark): ...new function.
14091 (pass_tm_memopt): Replace declaration with that of...
14092 (make_pass_tm_memopt): ...new function.
14093 (pass_tm_edges): Replace declaration with that of...
14094 (make_pass_tm_edges): ...new function.
14095 (pass_split_functions): Replace declaration with that of...
14096 (make_pass_split_functions): ...new function.
14097 (pass_feedback_split_functions): Replace declaration with that of...
14098 (make_pass_feedback_split_functions): ...new function.
14099 (pass_strength_reduction): Replace declaration with that of...
14100 (make_pass_strength_reduction): ...new function.
14101 (pass_ipa_lower_emutls): Replace declaration with that of...
14102 (make_pass_ipa_lower_emutls): ...new function.
14103 (pass_ipa_function_and_variable_visibility): Replace declaration with
14105 (make_pass_ipa_function_and_variable_visibility): ...new function.
14106 (pass_ipa_tree_profile): Replace declaration with that of...
14107 (make_pass_ipa_tree_profile): ...new function.
14108 (pass_early_local_passes): Replace declaration with that of...
14109 (make_pass_early_local_passes): ...new function.
14110 (pass_ipa_whole_program_visibility): Replace declaration with that
14112 (make_pass_ipa_whole_program_visibility): ...new function.
14113 (pass_ipa_lto_gimple_out): Replace declaration with that of...
14114 (make_pass_ipa_lto_gimple_out): ...new function.
14115 (pass_ipa_increase_alignment): Replace declaration with that of...
14116 (make_pass_ipa_increase_alignment): ...new function.
14117 (pass_ipa_inline): Replace declaration with that of...
14118 (make_pass_ipa_inline): ...new function.
14119 (pass_ipa_free_lang_data): Replace declaration with that of...
14120 (make_pass_ipa_free_lang_data): ...new function.
14121 (pass_ipa_free_inline_summary): Replace declaration with that of...
14122 (make_pass_ipa_free_inline_summary): ...new function.
14123 (pass_ipa_cp): Replace declaration with that of...
14124 (make_pass_ipa_cp): ...new function.
14125 (pass_ipa_reference): Replace declaration with that of...
14126 (make_pass_ipa_reference): ...new function.
14127 (pass_ipa_pure_const): Replace declaration with that of...
14128 (make_pass_ipa_pure_const): ...new function.
14129 (pass_ipa_pta): Replace declaration with that of...
14130 (make_pass_ipa_pta): ...new function.
14131 (pass_ipa_lto_finish_out): Replace declaration with that of...
14132 (make_pass_ipa_lto_finish_out): ...new function.
14133 (pass_ipa_tm): Replace declaration with that of...
14134 (make_pass_ipa_tm): ...new function.
14135 (pass_ipa_profile): Replace declaration with that of...
14136 (make_pass_ipa_profile): ...new function.
14137 (pass_ipa_cdtor_merge): Replace declaration with that of...
14138 (make_pass_ipa_cdtor_merge): ...new function.
14139 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
14141 (make_pass_cleanup_cfg_post_optimizing): ...new function.
14142 (pass_init_datastructures): Replace declaration with that of...
14143 (make_pass_init_datastructures): ...new function.
14144 (pass_fixup_cfg): Replace declaration with that of...
14145 (make_pass_fixup_cfg): ...new function.
14146 (pass_expand): Replace declaration with that of...
14147 (make_pass_expand): ...new function.
14148 (pass_instantiate_virtual_regs): Replace declaration with that of...
14149 (make_pass_instantiate_virtual_regs): ...new function.
14150 (pass_rtl_fwprop): Replace declaration with that of...
14151 (make_pass_rtl_fwprop): ...new function.
14152 (pass_rtl_fwprop_addr): Replace declaration with that of...
14153 (make_pass_rtl_fwprop_addr): ...new function.
14154 (pass_jump): Replace declaration with that of...
14155 (make_pass_jump): ...new function.
14156 (pass_jump2): Replace declaration with that of...
14157 (make_pass_jump2): ...new function.
14158 (pass_lower_subreg): Replace declaration with that of...
14159 (make_pass_lower_subreg): ...new function.
14160 (pass_cse): Replace declaration with that of...
14161 (make_pass_cse): ...new function.
14162 (pass_fast_rtl_dce): Replace declaration with that of...
14163 (make_pass_fast_rtl_dce): ...new function.
14164 (pass_ud_rtl_dce): Replace declaration with that of...
14165 (make_pass_ud_rtl_dce): ...new function.
14166 (pass_rtl_dce): Replace declaration with that of...
14167 (make_pass_rtl_dce): ...new function.
14168 (pass_rtl_dse1): Replace declaration with that of...
14169 (make_pass_rtl_dse1): ...new function.
14170 (pass_rtl_dse2): Replace declaration with that of...
14171 (make_pass_rtl_dse2): ...new function.
14172 (pass_rtl_dse3): Replace declaration with that of...
14173 (make_pass_rtl_dse3): ...new function.
14174 (pass_rtl_cprop): Replace declaration with that of...
14175 (make_pass_rtl_cprop): ...new function.
14176 (pass_rtl_pre): Replace declaration with that of...
14177 (make_pass_rtl_pre): ...new function.
14178 (pass_rtl_hoist): Replace declaration with that of...
14179 (make_pass_rtl_hoist): ...new function.
14180 (pass_rtl_store_motion): Replace declaration with that of...
14181 (make_pass_rtl_store_motion): ...new function.
14182 (pass_cse_after_global_opts): Replace declaration with that of...
14183 (make_pass_cse_after_global_opts): ...new function.
14184 (pass_rtl_ifcvt): Replace declaration with that of...
14185 (make_pass_rtl_ifcvt): ...new function.
14186 (pass_into_cfg_layout_mode): Replace declaration with that of...
14187 (make_pass_into_cfg_layout_mode): ...new function.
14188 (pass_outof_cfg_layout_mode): Replace declaration with that of...
14189 (make_pass_outof_cfg_layout_mode): ...new function.
14190 (pass_loop2): Replace declaration with that of...
14191 (make_pass_loop2): ...new function.
14192 (pass_rtl_loop_init): Replace declaration with that of...
14193 (make_pass_rtl_loop_init): ...new function.
14194 (pass_rtl_move_loop_invariants): Replace declaration with that of...
14195 (make_pass_rtl_move_loop_invariants): ...new function.
14196 (pass_rtl_unswitch): Replace declaration with that of...
14197 (make_pass_rtl_unswitch): ...new function.
14198 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
14199 (make_pass_rtl_unroll_and_peel_loops): ...new function.
14200 (pass_rtl_doloop): Replace declaration with that of...
14201 (make_pass_rtl_doloop): ...new function.
14202 (pass_rtl_loop_done): Replace declaration with that of...
14203 (make_pass_rtl_loop_done): ...new function.
14204 (pass_web): Replace declaration with that of...
14205 (make_pass_web): ...new function.
14206 (pass_cse2): Replace declaration with that of...
14207 (make_pass_cse2): ...new function.
14208 (pass_df_initialize_opt): Replace declaration with that of...
14209 (make_pass_df_initialize_opt): ...new function.
14210 (pass_df_initialize_no_opt): Replace declaration with that of...
14211 (make_pass_df_initialize_no_opt): ...new function.
14212 (pass_reginfo_init): Replace declaration with that of...
14213 (make_pass_reginfo_init): ...new function.
14214 (pass_inc_dec): Replace declaration with that of...
14215 (make_pass_inc_dec): ...new function.
14216 (pass_stack_ptr_mod): Replace declaration with that of...
14217 (make_pass_stack_ptr_mod): ...new function.
14218 (pass_initialize_regs): Replace declaration with that of...
14219 (make_pass_initialize_regs): ...new function.
14220 (pass_combine): Replace declaration with that of...
14221 (make_pass_combine): ...new function.
14222 (pass_if_after_combine): Replace declaration with that of...
14223 (make_pass_if_after_combine): ...new function.
14224 (pass_ree): Replace declaration with that of...
14225 (make_pass_ree): ...new function.
14226 (pass_partition_blocks): Replace declaration with that of...
14227 (make_pass_partition_blocks): ...new function.
14228 (pass_match_asm_constraints): Replace declaration with that of...
14229 (make_pass_match_asm_constraints): ...new function.
14230 (pass_regmove): Replace declaration with that of...
14231 (make_pass_regmove): ...new function.
14232 (pass_split_all_insns): Replace declaration with that of...
14233 (make_pass_split_all_insns): ...new function.
14234 (pass_fast_rtl_byte_dce): Replace declaration with that of...
14235 (make_pass_fast_rtl_byte_dce): ...new function.
14236 (pass_lower_subreg2): Replace declaration with that of...
14237 (make_pass_lower_subreg2): ...new function.
14238 (pass_mode_switching): Replace declaration with that of...
14239 (make_pass_mode_switching): ...new function.
14240 (pass_sms): Replace declaration with that of...
14241 (make_pass_sms): ...new function.
14242 (pass_sched): Replace declaration with that of...
14243 (make_pass_sched): ...new function.
14244 (pass_ira): Replace declaration with that of...
14245 (make_pass_ira): ...new function.
14246 (pass_reload): Replace declaration with that of...
14247 (make_pass_reload): ...new function.
14248 (pass_clean_state): Replace declaration with that of...
14249 (make_pass_clean_state): ...new function.
14250 (pass_branch_prob): Replace declaration with that of...
14251 (make_pass_branch_prob): ...new function.
14252 (pass_value_profile_transformations): Replace declaration with that
14254 (make_pass_value_profile_transformations): ...new function.
14255 (pass_postreload_cse): Replace declaration with that of...
14256 (make_pass_postreload_cse): ...new function.
14257 (pass_gcse2): Replace declaration with that of...
14258 (make_pass_gcse2): ...new function.
14259 (pass_split_after_reload): Replace declaration with that of...
14260 (make_pass_split_after_reload): ...new function.
14261 (pass_branch_target_load_optimize1): Replace declaration with that
14263 (make_pass_branch_target_load_optimize1): ...new function.
14264 (pass_thread_prologue_and_epilogue): Replace declaration with that
14266 (make_pass_thread_prologue_and_epilogue): ...new function.
14267 (pass_stack_adjustments): Replace declaration with that of...
14268 (make_pass_stack_adjustments): ...new function.
14269 (pass_peephole2): Replace declaration with that of...
14270 (make_pass_peephole2): ...new function.
14271 (pass_if_after_reload): Replace declaration with that of...
14272 (make_pass_if_after_reload): ...new function.
14273 (pass_regrename): Replace declaration with that of...
14274 (make_pass_regrename): ...new function.
14275 (pass_cprop_hardreg): Replace declaration with that of...
14276 (make_pass_cprop_hardreg): ...new function.
14277 (pass_reorder_blocks): Replace declaration with that of...
14278 (make_pass_reorder_blocks): ...new function.
14279 (pass_branch_target_load_optimize2): Replace declaration with that
14281 (make_pass_branch_target_load_optimize2): ...new function.
14282 (pass_leaf_regs): Replace declaration with that of...
14283 (make_pass_leaf_regs): ...new function.
14284 (pass_split_before_sched2): Replace declaration with that of...
14285 (make_pass_split_before_sched2): ...new function.
14286 (pass_compare_elim_after_reload): Replace declaration with that of...
14287 (make_pass_compare_elim_after_reload): ...new function.
14288 (pass_sched2): Replace declaration with that of...
14289 (make_pass_sched2): ...new function.
14290 (pass_stack_regs): Replace declaration with that of...
14291 (make_pass_stack_regs): ...new function.
14292 (pass_stack_regs_run): Replace declaration with that of...
14293 (make_pass_stack_regs_run): ...new function.
14294 (pass_df_finish): Replace declaration with that of...
14295 (make_pass_df_finish): ...new function.
14296 (pass_compute_alignments): Replace declaration with that of...
14297 (make_pass_compute_alignments): ...new function.
14298 (pass_duplicate_computed_gotos): Replace declaration with that of...
14299 (make_pass_duplicate_computed_gotos): ...new function.
14300 (pass_variable_tracking): Replace declaration with that of...
14301 (make_pass_variable_tracking): ...new function.
14302 (pass_free_cfg): Replace declaration with that of...
14303 (make_pass_free_cfg): ...new function.
14304 (pass_machine_reorg): Replace declaration with that of...
14305 (make_pass_machine_reorg): ...new function.
14306 (pass_cleanup_barriers): Replace declaration with that of...
14307 (make_pass_cleanup_barriers): ...new function.
14308 (pass_delay_slots): Replace declaration with that of...
14309 (make_pass_delay_slots): ...new function.
14310 (pass_split_for_shorten_branches): Replace declaration with that of...
14311 (make_pass_split_for_shorten_branches): ...new function.
14312 (pass_split_before_regstack): Replace declaration with that of...
14313 (make_pass_split_before_regstack): ...new function.
14314 (pass_convert_to_eh_region_ranges): Replace declaration with that
14316 (make_pass_convert_to_eh_region_ranges): ...new function.
14317 (pass_shorten_branches): Replace declaration with that of...
14318 (make_pass_shorten_branches): ...new function.
14319 (pass_set_nothrow_function_flags): Replace declaration with that of...
14320 (make_pass_set_nothrow_function_flags): ...new function.
14321 (pass_dwarf2_frame): Replace declaration with that of...
14322 (make_pass_dwarf2_frame): ...new function.
14323 (pass_final): Replace declaration with that of...
14324 (make_pass_final): ...new function.
14325 (pass_rtl_seqabstr): Replace declaration with that of...
14326 (make_pass_rtl_seqabstr): ...new function.
14327 (pass_release_ssa_names): Replace declaration with that of...
14328 (make_pass_release_ssa_names): ...new function.
14329 (pass_early_inline): Replace declaration with that of...
14330 (make_pass_early_inline): ...new function.
14331 (pass_inline_parameters): Replace declaration with that of...
14332 (make_pass_inline_parameters): ...new function.
14333 (pass_update_address_taken): Replace declaration with that of...
14334 (make_pass_update_address_taken): ...new function.
14335 (pass_convert_switch): Replace declaration with that of...
14336 (make_pass_convert_switch): ...new function.
14337 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
14338 to a subclass of simple_ipa_opt_pass along with...
14339 (pass_data_ipa_tree_profile): ...new pass_data instance and...
14340 (make_pass_ipa_tree_profile): ...new function.
14341 * tree-sra.c (pass_sra_early): Convert from a global struct to a
14342 subclass of gimple_opt_pass along with...
14343 (pass_data_sra_early): ...new pass_data instance and...
14344 (make_pass_sra_early): ...new function.
14345 (pass_sra): Convert from a global struct to a subclass of
14346 gimple_opt_pass along with...
14347 (pass_data_sra): ...new pass_data instance and...
14348 (make_pass_sra): ...new function.
14349 (pass_early_ipa_sra): Convert from a global struct to a subclass of
14350 gimple_opt_pass along with...
14351 (pass_data_early_ipa_sra): ...new pass_data instance and...
14352 (make_pass_early_ipa_sra): ...new function.
14353 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
14354 subclass of gimple_opt_pass along with...
14355 (pass_data_ccp): ...new pass_data instance and...
14356 (make_pass_ccp): ...new function.
14357 (pass_fold_builtins): Convert from a global struct to a subclass of
14358 gimple_opt_pass along with...
14359 (pass_data_fold_builtins): ...new pass_data instance and...
14360 (make_pass_fold_builtins): ...new function.
14361 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
14362 subclass of gimple_opt_pass along with...
14363 (pass_data_copy_prop): ...new pass_data instance and...
14364 (make_pass_copy_prop): ...new function.
14365 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
14366 global struct to a subclass of gimple_opt_pass along with...
14367 (pass_data_rename_ssa_copies): ...new pass_data instance and...
14368 (make_pass_rename_ssa_copies): ...new function.
14369 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
14370 subclass of gimple_opt_pass along with...
14371 (pass_data_dce): ...new pass_data instance and...
14372 (make_pass_dce): ...new function.
14373 (pass_dce_loop): Convert from a global struct to a subclass of
14374 gimple_opt_pass along with...
14375 (pass_data_dce_loop): ...new pass_data instance and...
14376 (make_pass_dce_loop): ...new function.
14377 (pass_cd_dce): Convert from a global struct to a subclass of
14378 gimple_opt_pass along with...
14379 (pass_data_cd_dce): ...new pass_data instance and...
14380 (make_pass_cd_dce): ...new function.
14381 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
14382 subclass of gimple_opt_pass along with...
14383 (pass_data_dominator): ...new pass_data instance and...
14384 (make_pass_dominator): ...new function.
14385 (pass_phi_only_cprop): Convert from a global struct to a subclass of
14386 gimple_opt_pass along with...
14387 (pass_data_phi_only_cprop): ...new pass_data instance and...
14388 (make_pass_phi_only_cprop): ...new function.
14389 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
14390 subclass of gimple_opt_pass along with...
14391 (pass_data_dse): ...new pass_data instance and...
14392 (make_pass_dse): ...new function.
14393 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
14394 a subclass of gimple_opt_pass along with...
14395 (pass_data_forwprop): ...new pass_data instance and...
14396 (make_pass_forwprop): ...new function.
14397 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
14398 struct to a subclass of gimple_opt_pass along with...
14399 (pass_data_tree_ifcombine): ...new pass_data instance and...
14400 (make_pass_tree_ifcombine): ...new function.
14401 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
14402 subclass of gimple_opt_pass along with...
14403 (pass_data_ch): ...new pass_data instance and...
14404 (make_pass_ch): ...new function.
14405 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
14406 subclass of gimple_opt_pass along with...
14407 (pass_data_tree_loop): ...new pass_data instance and...
14408 (make_pass_tree_loop): ...new function.
14409 (pass_tree_loop_init): Convert from a global struct to a subclass of
14410 gimple_opt_pass along with...
14411 (pass_data_tree_loop_init): ...new pass_data instance and...
14412 (make_pass_tree_loop_init): ...new function.
14413 (pass_lim): Convert from a global struct to a subclass of
14414 gimple_opt_pass along with...
14415 (pass_data_lim): ...new pass_data instance and...
14416 (make_pass_lim): ...new function.
14417 (pass_tree_unswitch): Convert from a global struct to a subclass of
14418 gimple_opt_pass along with...
14419 (pass_data_tree_unswitch): ...new pass_data instance and...
14420 (make_pass_tree_unswitch): ...new function.
14421 (pass_predcom): Convert from a global struct to a subclass of
14422 gimple_opt_pass along with...
14423 (pass_data_predcom): ...new pass_data instance and...
14424 (make_pass_predcom): ...new function.
14425 (pass_vectorize): Convert from a global struct to a subclass of
14426 gimple_opt_pass along with...
14427 (pass_data_vectorize): ...new pass_data instance and...
14428 (make_pass_vectorize): ...new function.
14429 (pass_graphite): Convert from a global struct to a subclass of
14430 gimple_opt_pass along with...
14431 (pass_data_graphite): ...new pass_data instance and...
14432 (make_pass_graphite): ...new function.
14433 (pass_graphite_transforms): Convert from a global struct to a subclass
14434 of gimple_opt_pass along with...
14435 (pass_data_graphite_transforms): ...new pass_data instance and...
14436 (make_pass_graphite_transforms): ...new function.
14437 (pass_check_data_deps): Convert from a global struct to a subclass of
14438 gimple_opt_pass along with...
14439 (pass_data_check_data_deps): ...new pass_data instance and...
14440 (make_pass_check_data_deps): ...new function.
14441 (pass_iv_canon): Convert from a global struct to a subclass of
14442 gimple_opt_pass along with...
14443 (pass_data_iv_canon): ...new pass_data instance and...
14444 (make_pass_iv_canon): ...new function.
14445 (pass_scev_cprop): Convert from a global struct to a subclass of
14446 gimple_opt_pass along with...
14447 (pass_data_scev_cprop): ...new pass_data instance and...
14448 (make_pass_scev_cprop): ...new function.
14449 (pass_record_bounds): Convert from a global struct to a subclass of
14450 gimple_opt_pass along with...
14451 (pass_data_record_bounds): ...new pass_data instance and...
14452 (make_pass_record_bounds): ...new function.
14453 (pass_complete_unroll): Convert from a global struct to a subclass of
14454 gimple_opt_pass along with...
14455 (pass_data_complete_unroll): ...new pass_data instance and...
14456 (make_pass_complete_unroll): ...new function.
14457 (pass_complete_unrolli): Convert from a global struct to a subclass of
14458 gimple_opt_pass along with...
14459 (pass_data_complete_unrolli): ...new pass_data instance and...
14460 (make_pass_complete_unrolli): ...new function.
14461 (pass_parallelize_loops): Convert from a global struct to a subclass
14462 of gimple_opt_pass along with...
14463 (pass_data_parallelize_loops): ...new pass_data instance and...
14464 (make_pass_parallelize_loops): ...new function.
14465 (pass_loop_prefetch): Convert from a global struct to a subclass of
14466 gimple_opt_pass along with...
14467 (pass_data_loop_prefetch): ...new pass_data instance and...
14468 (make_pass_loop_prefetch): ...new function.
14469 (pass_iv_optimize): Convert from a global struct to a subclass of
14470 gimple_opt_pass along with...
14471 (pass_data_iv_optimize): ...new pass_data instance and...
14472 (make_pass_iv_optimize): ...new function.
14473 (pass_tree_loop_done): Convert from a global struct to a subclass of
14474 gimple_opt_pass along with...
14475 (pass_data_tree_loop_done): ...new pass_data instance and...
14476 (make_pass_tree_loop_done): ...new function.
14477 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
14478 struct to a subclass of gimple_opt_pass along with...
14479 (pass_data_cse_reciprocals): ...new pass_data instance and...
14480 (make_pass_cse_reciprocals): ...new function.
14481 (pass_cse_sincos): Convert from a global struct to a subclass of
14482 gimple_opt_pass along with...
14483 (pass_data_cse_sincos): ...new pass_data instance and...
14484 (make_pass_cse_sincos): ...new function.
14485 (pass_optimize_bswap): Convert from a global struct to a subclass of
14486 gimple_opt_pass along with...
14487 (pass_data_optimize_bswap): ...new pass_data instance and...
14488 (make_pass_optimize_bswap): ...new function.
14489 (pass_optimize_widening_mul): Convert from a global struct to a
14490 subclass of gimple_opt_pass along with...
14491 (pass_data_optimize_widening_mul): ...new pass_data instance and...
14492 (make_pass_optimize_widening_mul): ...new function.
14493 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
14494 subclass of gimple_opt_pass along with...
14495 (pass_data_phiopt): ...new pass_data instance and...
14496 (make_pass_phiopt): ...new function.
14497 (pass_cselim): Convert from a global struct to a subclass of
14498 gimple_opt_pass along with...
14499 (pass_data_cselim): ...new pass_data instance and...
14500 (make_pass_cselim): ...new function.
14501 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
14502 subclass of gimple_opt_pass along with...
14503 (pass_data_phiprop): ...new pass_data instance and...
14504 (make_pass_phiprop): ...new function.
14505 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
14506 subclass of gimple_opt_pass along with...
14507 (pass_data_pre): ...new pass_data instance and...
14508 (make_pass_pre): ...new function.
14509 (pass_fre): Convert from a global struct to a subclass of
14510 gimple_opt_pass along with...
14511 (pass_data_fre): ...new pass_data instance and...
14512 (make_pass_fre): ...new function.
14513 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
14514 subclass of gimple_opt_pass along with...
14515 (pass_data_reassoc): ...new pass_data instance and...
14516 (make_pass_reassoc): ...new function.
14517 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
14518 subclass of gimple_opt_pass along with...
14519 (pass_data_sink_code): ...new pass_data instance and...
14520 (make_pass_sink_code): ...new function.
14521 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
14522 subclass of gimple_opt_pass along with...
14523 (pass_data_strlen): ...new pass_data instance and...
14524 (make_pass_strlen): ...new function.
14525 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
14526 struct to a subclass of gimple_opt_pass along with...
14527 (pass_data_build_alias): ...new pass_data instance and...
14528 (make_pass_build_alias): ...new function.
14529 (pass_build_ealias): Convert from a global struct to a subclass of
14530 gimple_opt_pass along with...
14531 (pass_data_build_ealias): ...new pass_data instance and...
14532 (make_pass_build_ealias): ...new function.
14533 (pass_ipa_pta): Convert from a global struct to a subclass of
14534 simple_ipa_opt_pass along with...
14535 (pass_data_ipa_pta): ...new pass_data instance and...
14536 (make_pass_ipa_pta): ...new function.
14537 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
14538 subclass of gimple_opt_pass along with...
14539 (pass_data_uncprop): ...new pass_data instance and...
14540 (make_pass_uncprop): ...new function.
14541 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
14542 global struct to a subclass of gimple_opt_pass along with...
14543 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
14544 (make_pass_late_warn_uninitialized): ...new function.
14545 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
14546 to a subclass of gimple_opt_pass along with...
14547 (pass_data_init_datastructures): ...new pass_data instance and...
14548 (make_pass_init_datastructures): ...new function.
14549 (pass_early_warn_uninitialized): Convert from a global struct to a
14550 subclass of gimple_opt_pass along with...
14551 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
14552 (make_pass_early_warn_uninitialized): ...new function.
14553 (pass_update_address_taken): Convert from a global struct to a
14554 subclass of gimple_opt_pass along with...
14555 (pass_data_update_address_taken): ...new pass_data instance and...
14556 (make_pass_update_address_taken): ...new function.
14557 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
14558 struct to a subclass of gimple_opt_pass along with...
14559 (pass_data_release_ssa_names): ...new pass_data instance and...
14560 (make_pass_release_ssa_names): ...new function.
14561 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
14562 subclass of gimple_opt_pass along with...
14563 (pass_data_stdarg): ...new pass_data instance and...
14564 (make_pass_stdarg): ...new function.
14565 * tree-switch-conversion.c (pass_convert_switch): Convert from a
14566 global struct to a subclass of gimple_opt_pass along with...
14567 (pass_data_convert_switch): ...new pass_data instance and...
14568 (make_pass_convert_switch): ...new function.
14569 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
14570 to a subclass of gimple_opt_pass along with...
14571 (pass_data_tail_recursion): ...new pass_data instance and...
14572 (make_pass_tail_recursion): ...new function.
14573 (pass_tail_calls): Convert from a global struct to a subclass of
14574 gimple_opt_pass along with...
14575 (pass_data_tail_calls): ...new pass_data instance and...
14576 (make_pass_tail_calls): ...new function.
14577 * tree-vect-generic.c (pass_lower_vector): Convert from a global
14578 struct to a subclass of gimple_opt_pass along with...
14579 (pass_data_lower_vector): ...new pass_data instance and...
14580 (make_pass_lower_vector): ...new function.
14581 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
14582 gimple_opt_pass along with...
14583 (pass_data_lower_vector_ssa): ...new pass_data instance and...
14584 (make_pass_lower_vector_ssa): ...new function.
14585 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
14586 to a subclass of gimple_opt_pass along with...
14587 (pass_data_slp_vectorize): ...new pass_data instance and...
14588 (make_pass_slp_vectorize): ...new function.
14589 (pass_ipa_increase_alignment): Convert from a global struct to a
14590 subclass of simple_ipa_opt_pass along with...
14591 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
14592 (make_pass_ipa_increase_alignment): ...new function.
14593 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
14594 gimple_opt_pass along with...
14595 (pass_data_vrp): ...new pass_data instance and...
14596 (make_pass_vrp): ...new function.
14597 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
14598 subclass of simple_ipa_opt_pass along with...
14599 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
14600 (make_pass_ipa_free_lang_data): ...new function.
14601 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
14602 gimple_opt_pass along with...
14603 (pass_data_tsan): ...new pass_data instance and...
14604 (make_pass_tsan): ...new function.
14605 (pass_tsan_O0): Convert from a global struct to a subclass of
14606 gimple_opt_pass along with...
14607 (pass_data_tsan_O0): ...new pass_data instance and...
14608 (make_pass_tsan_O0): ...new function.
14609 * var-tracking.c (pass_variable_tracking): Convert from a global
14610 struct to a subclass of rtl_opt_pass along with...
14611 (pass_data_variable_tracking): ...new pass_data instance and...
14612 (make_pass_variable_tracking): ...new function.
14613 * web.c (pass_web): Convert from a global struct to a subclass of
14614 rtl_opt_pass along with...
14615 (pass_data_web): ...new pass_data instance and...
14616 (make_pass_web): ...new function.
14617 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
14618 declaration with that of...
14619 (make_pass_mode_switch_use): ...new function.
14620 (pass_resolve_sw_modes): Replace declaration with that of...
14621 (make_pass_resolve_sw_modes): ...new function.
14622 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
14623 from a global struct to a subclass of rtl_opt_pass along with...
14624 (pass_data_mode_switch_use): ...new pass_data instance and...
14625 (make_pass_mode_switch_use): ...new function.
14626 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
14627 from a global struct to a subclass of rtl_opt_pass along with...
14628 (pass_data_resolve_sw_modes): ...new pass_data instance and...
14629 (make_pass_resolve_sw_modes): ...new function.
14630 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
14631 struct to a subclass of rtl_opt_pass along with...
14632 (pass_data_insert_vzeroupper): ...new pass_data instance and...
14633 (make_pass_insert_vzeroupper): ...new function.
14634 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
14635 global struct to a subclass of rtl_opt_pass along with...
14636 (pass_data_work_around_errata): ...new pass_data instance and...
14637 (make_pass_work_around_errata): ...new function.
14638 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
14639 struct to a subclass of rtl_opt_pass along with...
14640 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
14641 (make_pass_mips_machine_reorg2): ...new function.
14643 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14645 * passes.c (pass_manager::operator new): New.
14647 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14649 Handwritten part of conversion of passes to C++ classes.
14651 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
14652 (toplev.o): Add dep on PASS_MANAGER_H.
14653 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
14654 of early local pases to reflect this moving from a global to a
14655 member of gcc::pass_manager.
14656 (cgraph_add_new_function): Likewise.
14657 * lto-cgraph.c (lto_output_node): Update for conversion of
14658 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
14659 * passes.c (opt_pass::clone): New.
14660 (opt_pass::gate): New.
14661 (opt_pass::execute): New.
14662 (opt_pass::opt_pass): New.
14663 (pass_manager::execute_early_local_passes): New.
14664 (pass_manager::execute_pass_mode_switching): new.
14665 (finish_optimization_passes): Convert to...
14666 (pass_manager::finish_optimization_passes): ...this.
14667 (finish_optimization_passes): Update for conversion of passes to
14669 (register_dump_files_1): Use has_gate since we cannot portably
14670 check a vtable entry against NULL.
14671 (dump_one_pass): Likewise.
14672 (ipa_write_summaries_2): Likewise.
14673 (ipa_write_optimization_summaries_1): Likewise.
14674 (ipa_read_summaries_1): Likewise.
14675 (ipa_read_optimization_summaries_1): Likewise.
14676 (execute_ipa_stmt_fixups): Likewise.
14677 (pass_manager::pass_manager): Rewrite pass-creation, invoking
14678 pass-creation functions rather than wiring up globals, and
14679 storing the results in fields of pass_manager generated using
14680 pass-instances.def.
14681 (pass_manager::dump_profile_report): Update for conversion of
14682 passes to C++ classes.
14683 (pass_manager::execute_ipa_summary_passes): Likewise.
14684 (execute_one_ipa_transform_pass): Likewise.
14685 (execute_one_pass): Use has_gate and has_execute since we cannot
14686 portably check a vtable entry against NULL.
14687 * pass_manager.h (pass_manager::finish_optimization_passes): New.
14688 (pass_manager): Use pass-instances.def to add fields for the
14689 various pass instances.
14690 * toplev.c (finalize): Update for move of
14691 finish_optimization_passes to a method of gcc::pass_manager.
14692 * toplev.h (finish_optimization_passes): Move to method of class
14694 * tree-pass.h (struct pass_data): New.
14695 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
14696 (opt_pass::gate): Convert to virtual function.
14697 (opt_pass::~opt_pass): New.
14698 (opt_pass::clone): New.
14699 (opt_pass::execute): Convert to virtual function.
14700 (opt_pass::opt_pass): New.
14701 (opt_pass::ctxt_): new.
14702 (gimple_opt_pass): Convert to subclass of opt_pass.
14703 (gimple_opt_pass::gimple_opt_pass): New.
14704 (rtl_opt_pass): Convert to subclass of opt_pass.
14705 (rtl_opt_pass::rtl_opt_pass): New.
14706 (ipa_opt_pass_d): Convert to subclass of opt_pass.
14707 (ipa_opt_pass_d::ipa_opt_pass_d): New.
14708 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
14709 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
14710 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
14711 invocation of pass_mode_switching to reflect this moving from a
14712 global to a member of gcc::pass_manager.
14713 (ix86_option_override): Rework how pass_insert_vzeroupper is
14714 added to the pass_manager to reflect autogenerated changes.
14715 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
14717 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
14719 PR rtl-optimization/57708
14720 * recog.c (peep2_find_free_register): Validate all regs in a
14723 2013-08-05 Jan Hubicka <jh@suse.cz>
14726 * cgraph.c (verify_cgraph_node): Accept local flags from other
14728 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
14729 (function_and_variable_visibility): Likewise.
14730 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
14732 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14734 * graph.c (init_graph_slim_pretty_print): Remove.
14735 (print_graph_cfg): Do not call it. Use local pretty printer.
14736 (start_graph_dump): Likewise.
14738 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14740 * gimple-pretty-print.c (buffer): Remove.
14741 (initialized): Likewise.
14742 (maybe_init_pretty_print): Likewise.
14743 (print_gimple_stmt): Do not call it. Use non-static local
14744 pretty_printer variable.
14745 (print_gimple_expr): Likewise.
14746 (print_gimple_seq): Likewise.
14747 (gimple_dump_bb): Likewise.
14749 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14751 * asan.c (asan_pp): Remove.
14752 (asan_pp_initialized): Likewise.
14753 (asan_pp_initialize): Likewise.
14754 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
14755 (asan_emit_stack_protection): Tidy. Use local pretty printer.
14756 (asan_add_global): Likewise.
14758 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
14760 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
14761 * diagnostic.h (diagnostic_flush_buffer): Adjust.
14762 * pretty-print.c (pp_formatted_text_data): Likewise.
14763 (pp_indent): Rename from pp_base_indent.
14764 (pp_format): Rename from pp_base_format.
14765 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
14766 (pp_format_verbatim): Rename from pp_base_format_verbatim.
14767 (pp_flush): Rename from pp_base_flush.
14768 (pp_set_line_maximum_length): Rename from
14769 pp_base_set_line_maximum_length.
14770 (pp_clear_output_area): Rename from pp_base_clear_output_area.
14771 (pp_set_prefix): Rename from pp_base_set_prefix.
14772 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
14773 (pp_emit_prefix): Rename from pp_base_emit_prefix.
14774 (pp_append_text): Rename from pp_base_append_text.
14775 (pp_formatted_text): Rename from pp_base_formatted_text.
14776 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
14777 (pp_remaining_character_count_for_line): Rename from
14778 pp_base_remaining_character_count_for_line.
14779 (pp_newline): Rename from pp_base_newline.
14780 (pp_character): Rename from pp_base_character.
14781 (pp_string): Rename from pp_base_string.
14782 (pp_maybe_space): Rename from pp_base_maybe_space.
14783 * asan.c (asan_pp_string): Adjust.
14784 (asan_emit_stack_protection): Likewise.
14785 (asan_add_global): Likewise.
14786 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
14787 * tree-mudflap.c (mf_varname_tree): Likewise.
14788 * tree-pretty-print.c (pp_tree_identifier): Rename from
14789 pp_base_tree_identifier.
14790 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
14791 Declare as function.
14793 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14795 * pretty-print.h (pp_bar_bar): New.
14796 (pp_ampersand_ampersand): Likewise.
14797 (pp_less_equal): Likewise.
14798 (pp_greater_equal): Likewise.
14799 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
14800 printer functions instead of pp_string or operators and punctuators.
14801 (dump_gimple_call): Likewise.
14802 (dump_gimple_omp_for): Likewise.
14803 (dump_gimple_transaction): Likewise.
14804 (dump_gimple_phi): Likewise.
14805 (pp_gimple_stmt_1): Likewise.
14806 * sched-vis.c (print_insn): Likewise.
14807 * tree-mudflap.c (mf_varname_tree): Likewise.
14808 * tree-pretty-print.c (dump_block_node): Likewise.
14809 (dump_generic_node): Likewise.
14811 2013-08-02 Jan Hubicka <jh@suse.cz>
14813 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
14815 * lto-streamer-out.c (tree_is_indexable): Results decls and
14816 parm decls are not indexable.
14817 (DFS_write_tree_body): Do not follow args and results.
14818 (hash_tree): Likewise.
14819 (output_functions): Rearrange so struct function is needed
14820 only when real body is output; be able to also ouptut abstract
14821 functions; output DECL_ARGUMENTS and DECL_RESULT.
14822 (lto_output): When not in WPA, ale store abstract functions.
14823 (write_symbol): Do not care about RESULT_DECL.
14824 (output_symbol_p): Handle correctly sbtract decls.
14825 * lto-streamer-in.c (input_function): Rearrange so struct
14826 function can be NULL at entry; allow streaming of
14827 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
14828 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
14829 sanity check during LTO.
14830 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
14831 RESULT_DECl and DECL_ARGUMENTS.
14832 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
14835 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14837 * pretty-print.h (pp_underscore): New.
14839 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
14840 printer functions instead of pp_character.
14841 (dump_binary_rhs): Likewise.
14842 (dump_ternary_rhs): Likewise.
14843 (dump_gimple_call_args): Likewise.
14844 (pp_points_to_solution): Likewise.
14845 (dump_gimple_call): Likewise.
14846 (dump_gimple_switch): Likewise.
14847 (dump_gimple_cond): Likewise.
14848 (dump_gimple_bind): Likewise.
14849 (dump_gimple_try): Likewise.
14850 (dump_gimple_omp_for): Likewise.
14851 (dump_gimple_omp_continue): Likewise.
14852 (dump_gimple_omp_single): Likewise.
14853 (dump_gimple_omp_sections): Likewise.
14854 (dump_gimple_omp_block): Likewise.
14855 (dump_gimple_omp_critical): Likewise.
14856 (dump_gimple_transaction): Likewise.
14857 (dump_gimple_asm): Likewise.
14858 (dump_gimple_phi): Likewise.
14859 (dump_gimple_omp_parallel): Likewise.
14860 (dump_gimple_omp_task): Likewise.
14861 (dump_gimple_omp_atomic_load): Likewise.
14862 (dump_gimple_omp_atomic_store): Likewise.
14863 (dump_gimple_mem_ops): Likewise.
14864 (pp_gimple_stmt_1): Likewise.
14865 (pp_cfg_jump): Likewise.
14866 (dump_implicit_edges): Likewise.
14867 (gimple_dump_bb_for_graph): Likewise.
14868 * graph.c (draw_cfg_node): Likewise.
14869 * langhooks.c (lhd_print_error_function): Likewise.
14870 * sched-vis.c (print_exp): Likewise.
14871 (print_value): Likewise.
14872 (print_pattern): Likewise.
14873 (print_insn): Likewise.
14874 (rtl_dump_bb_for_graph): Likewise.
14875 * tree-pretty-print.c (dump_function_declaration): Likewise.
14876 (dump_array_domain): Likewise.
14877 (dump_omp_clause): Likewise.
14878 (dump_location): Likewise.
14879 (dump_generic_node): Likewise.
14880 (print_struct_decl): Likewise.
14881 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14883 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14885 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14886 candidate table when replacing a candidate statement.
14887 (replace_rhs_if_not_dup): Likewise.
14888 (replace_one_candidate): Likewise.
14890 2013-08-02 Jan Hubicka <jh@suse.cz>
14891 Martin Liska <marxin.liska@gmail.com>
14893 * cgraphunit.c (add_new_function): Fix logic when adding from
14895 (assemble_thunk): Rename to ...
14896 (expand_thunk); .. this one; export; get it working with
14897 general functions; make produced gimple valid.
14898 * cgraph.h (expand_thunk): Declare.
14900 2013-08-02 Jan Hubicka <jh@suse.cz>
14902 * ipa-cp.c (gather_context_independent_values): Use
14903 ipa_get_param_move_cost.
14904 (get_replacement_map): Remove PARAM; move parameter folding
14906 (create_specialized_node): Update.
14907 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14908 assert that we have gimple body; update move_cost.
14909 (count_formal_params): Assert that we have gimple body.
14910 (ipa_dump_param): New function.
14911 (ipa_alloc_node_params): Break out from ...
14912 (ipa_initialize_node_params): ... here.
14913 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14914 (ipa_write_node_info): Stream move costs.
14915 (ipa_read_node_info): Read move costs.
14916 (ipa_update_after_lto_read): Do not recompute node params.
14917 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14918 (ipa_get_param): Check we are not in WPA.
14919 (ipa_get_param_move_cost): New.
14920 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14921 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14922 parm numbers to be present.
14924 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14926 PR rtl-optimization/58048
14927 * lra-constraints.c (process_alt_operands): Don't check asm
14928 operand on register.
14930 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14932 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14933 the implied StoreLoad barrier for atomic operations if before.
14935 2013-08-02 Jan Hubicka <jh@suse.cz>
14936 Martin Liska <marxin.liska@gmail.com>
14938 * cgraph.c (cgraph_function_body_availability): Do not check
14940 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14941 symtab_node_availability): Declare.
14942 * ipa.c (can_replace_by_local_alias): New.
14943 (function_and_variable_visibility): Use it.
14944 * symtab.c (symtab_for_node_and_aliases,
14945 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14947 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14949 PR rtl-optimization/57963
14950 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14951 (lra_constraints): Use them.
14953 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14955 * config/arm/types.md (define_attr "type"): Add "load_acq"
14957 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14959 (cortex_a53_store1): Likewise.
14961 2013-08-01 Jan Hubicka <jh@suse.cz>
14963 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14964 partitions are not needed.
14966 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14968 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14969 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14970 MAYBE_NON_Q_CLASS_P where appropriate.
14972 2013-08-01 Jan Hubicka <jh@suse.cz>
14974 * cgraph.h (release_function_body): Declare.
14975 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14976 of unused delcarations.
14978 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14980 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14981 RTL form when subtracting a constant.
14983 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14985 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14986 Generate canonical plus rtx with negated immediate instead of minus
14988 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14990 2013-08-01 Jan Hubicka <jh@suse.cz>
14992 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14993 (cgraph_release_function_body): Likewise.
14994 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
14995 * cgraph.h (cgrpah_node): Rename abstract_and_needed
14996 to used_as_abstract_origin.
14997 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
14998 symbols used as abstract origins.
14999 * cgraphunit.c (analyze_functions): Update.
15000 * ipa.c (symtab_remove_unreachable_nodes): Recompute
15001 used_as_abstract_origin.
15002 * tree-inline.c (tree_function_versioning): Update
15003 used_as_abstract_origin; be ready for DECL_RESULT and
15004 DECL_ARGUMENTS to be NULL.
15006 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
15007 for abstract functions.
15008 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
15011 2013-08-01 Jan Hubicka <jh@suse.cz>
15013 * profile.c (compute_value_histograms): Fix thinko.
15015 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
15017 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
15018 aarch-common-protos.h to extra_headers.
15019 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
15020 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
15021 * config/aarch64/t-aarch64 (aarch-common.o): Define.
15023 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
15025 * config/aarch64/aarch64.md (define_attr "type"): Delete.
15026 Include "../arm/types.md". Define "type" attribute for all patterns.
15027 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
15030 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
15032 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
15033 to support power8 load fusion.
15034 (fusion_gpr_mem_load): Likewise.
15036 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
15038 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
15039 declarations for power8 load fusion.
15040 (emit_fusion_gpr_load): Likewise.
15042 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
15043 tuning for power8, turn on fusion mode by default. Turn on sign
15044 extending fusion mode if normal fusion mode is on, and we are at
15046 (fusion_gpr_load_p): New function, return true if we can fuse an
15047 addis instruction with a dependent load to a GPR.
15048 (emit_fusion_gpr_load): Emit the instructions for power8 load
15051 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
15052 (VSX load fusion peepholes): New peepholes to fuse together an
15053 addi instruction with a VSX load instruction.
15055 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
15056 peepholes to fuse an addis instruction with a load to a GPR base
15057 register. If we are supporting sign extending fusions, convert
15058 sign extending loads to zero extending loads and add an explicit
15061 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
15063 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
15064 aarch-common-protos.h to extra_headers.
15065 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
15066 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
15067 (arm_early_store_addr_dep): Likewise.
15068 (arm_no_early_alu_shift_dep): Likewise.
15069 (arm_no_early_alu_shift_value_dep): Likewise.
15070 (arm_no_early_mul_dep): Likewise.
15071 (arm_no_early_store_addr_dep): Likewise.
15072 (arm_mac_accumulator_is_mul_result): Likewise.
15073 (arm_mac_accumulator_is_result): Likewise.
15074 * config/arm/aarch-common.c: ... here. New file.
15075 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
15077 (arm_early_store_addr_dep): Likewise.
15078 (arm_no_early_alu_shift_dep): Likewise.
15079 (arm_no_early_alu_shift_value_dep): Likewise.
15080 (arm_no_early_mul_dep): Likewise.
15081 (arm_no_early_store_addr_dep): Likewise.
15082 (arm_mac_accumulator_is_mul_result): Likewise.
15083 (arm_mac_accumulator_is_result): Likewise.
15084 * config/arm/aarch-common-protos.h: ... here. New file.
15085 * config/arm/t-arm (aarch-common.o): Define.
15087 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
15089 * config/arm/arm.md: Include new file "types.md".
15090 (define_attr "type"): Move from here to ...
15091 (define_attr "mul32"): Likewise.
15092 (define_attr "mul64"): Likewise.
15093 * config/arm/types.md: ... here. New file.
15095 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
15097 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
15098 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
15100 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15102 * gen-pass-instances.awk: Fix offset of substr().
15104 2013-07-31 David Malcolm <dmalcolm@redhat.com>
15106 * Makefile.in (pass-instances.def): New.
15107 (passes.o): Replace dependency on passes.def with one on
15110 * gen-pass-instances.awk: New.
15112 * passes.c (pass_manager::pass_manager): Use pass-instances.def
15113 rather than passes.def, updating local definition of NEXT_PASS
15114 macro to add an extra NUM parameter (currently unused).
15116 2013-07-30 David Malcolm <dmalcolm@redhat.com>
15118 * Makefile.in (PASS_MANAGER_H): New.
15119 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
15120 (passes.o): Likewise.
15121 (statistics.o): Likewise.
15122 (cgraphunit.o): Likewise.
15123 (context.o): Depend on PASS_MANAGER_H.
15125 * pass_manager.h: New.
15127 * cgraphunit.c (cgraph_add_new_function): Update for moves
15128 of globals to fields of pass_manager.
15129 (analyze_function): Likewise.
15130 (expand_function): Likewise.
15131 (ipa_passes): Likewise.
15132 (compile): Likewise.
15134 * context.c (context::context): New.
15135 * context.h (context::context): New.
15136 (context::get_passes): New.
15137 (context::passes_): New.
15139 * lto-cgraph.c (input_node): Update for moves of globals to
15140 fields of pass_manager.
15142 * passes.c (all_passes): Remove, in favor of a field of the
15143 same name within the new class pass_manager.
15144 (all_small_ipa_passes): Likewise.
15145 (all_lowering_passes): Likewise.
15146 (all_regular_ipa_passes): Likewise.
15147 (all_late_ipa_passes): Likewise.
15148 (all_lto_gen_passes): Likewise.
15149 (passes_by_id): Likewise.
15150 (passes_by_id_size): Likewise.
15151 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
15152 the new class pass_manager.
15153 (set_pass_for_id): Convert to...
15154 (pass_manager::set_pass_for_id): ...method.
15155 (get_pass_for_id): Convert to...
15156 (pass_manager::get_pass_for_id): ...method.
15157 (register_one_dump_file): Move body of implementation into...
15158 (pass_manager::register_one_dump_file): ...here.
15159 (register_dump_files_1): Convert to...
15160 (pass_manager::register_dump_files_1): ...method.
15161 (register_dump_files): Convert to...
15162 (pass_manager::register_dump_files): ...method.
15163 (create_pass_tab): Update for moves of globals to fields of
15165 (dump_passes): Move body of implementation into...
15166 (pass_manager::dump_passes): ...here.
15167 (register_pass): Move body of implementation into...
15168 (pass_manager::register_pass): ...here.
15169 (init_optimization_passes): Convert into...
15170 (pass_manager::pass_manager): ...constructor for new
15171 pass_manager class, and initialize the pass_lists array.
15172 (check_profile_consistency): Update for moves of globals to
15173 fields of pass_manager.
15174 (dump_profile_report): Move body of implementation into...
15175 (pass_manager::dump_profile_report): ...here.
15176 (ipa_write_summaries_1): Update for moves of pass lists from
15177 being globals to fields of pass_manager.
15178 (ipa_write_optimization_summaries): Likewise.
15179 (ipa_read_summaries): Likewise.
15180 (ipa_read_optimization_summaries): Likewise.
15181 (execute_all_ipa_stmt_fixups): Likewise.
15183 * statistics.c (statistics_fini): Update for moves of globals to
15184 fields of pass_manager.
15186 * toplev.c (general_init): Replace call to
15187 init_optimization_passes with construction of the pass_manager
15190 * tree-pass.h (all_passes): Remove, in favor of a field of the
15191 same name within the new class pass_manager.
15192 (all_small_ipa_passes): Likewise.
15193 (all_lowering_passes): Likewise.
15194 (all_regular_ipa_passes): Likewise.
15195 (all_lto_gen_passes): Likewise.
15196 (all_late_ipa_passes): Likewise.
15197 (passes_by_id): Likewise.
15198 (passes_by_id_size): Likewise.
15199 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
15200 the new class pass_manager.
15201 (get_pass_for_id): Remove.
15203 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
15205 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
15208 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
15210 * arm.md (mulhi3): New expand pattern.
15212 2013-07-30 Jan Hubicka <jh@suse.cz>
15213 Martin Liska <marxin.liska@gmail.com>
15215 * profile.c (compute_value_histograms): Do not ICE when
15216 there is mismatch only on some counters.
15218 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15220 PR rtl-optimization/57637
15221 * function.c (move_insn_for_shrink_wrap): Also check the
15222 GEN set of the LIVE problem for the liveness analysis
15223 if it exists, otherwise give up.
15225 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15227 PR tree-optimization/57993
15228 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
15229 replaced statement in the candidate table.
15230 (phi_add_costs): Return infinite cost when the hidden basis does
15231 not dominate all phis on which the candidate is dependent.
15232 (replace_one_candidate): Record replaced statement in the
15235 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
15237 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
15238 (ashlv2si3): New expander.
15239 (*ashlv2si3_i): New define_insn_and_split.
15240 * predicates.md (float_operation): Allow patterns with three
15241 basic sub-patterns.
15243 PR rtl-optimization/58021
15244 * mode-switching.c (create_pre_exit): Always split off preceding
15245 insns if we are not at the basic block head.
15247 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
15249 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
15250 (UCLIBC_DYNAMIC_LINKER): New macro.
15251 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
15253 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
15254 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
15256 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
15257 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
15258 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
15259 for SF and DF modes. Use ieee_quad_format for TF mode.
15260 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
15261 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
15262 (mips_option_override): Handle `-mnan=legacy'.
15263 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
15264 `-mabs=2008' and `-mnan=2008'.
15265 (OPTION_DEFAULT_SPECS): Add "nan" default.
15266 (ASM_SPEC): Handle `-mnan='.
15267 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
15268 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
15269 comment accordingly.
15270 (neg<mode>2): Likewise.
15271 * config/mips/mips.opt (mabs, mnan): New options.
15272 * doc/install.texi (Configuration): Document `--with-nan=' option.
15273 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
15275 (MIPS Options): Document them.
15276 * config.gcc <mips*-*-*>: Handle `--with-nan='.
15277 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
15278 * configure: Regenerate.
15279 * config.in: Regenerate.
15281 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
15283 * config/i386/i386.md (float post-reload splitters): Do not check
15284 for subregs of SSE registers.
15286 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
15287 H.J. Lu <hongjiu.lu@intel.com>
15291 * config/i386/i386.md (post-reload splitter
15292 to avoid partial SSE reg dependency stalls): New pattern.
15294 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
15296 * config/s390/s390.md ("movcc"): Swap load and store instructions.
15298 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
15300 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
15301 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
15303 2013-07-26 Cary Coutant <ccoutant@google.com>
15305 * dwarf2out.c (die_checksum_ordered): Don't include template
15306 instantiations in signature.
15307 (is_template_parameter): New function.
15308 (is_template_instantiation): New function.
15309 (generate_skeleton_bottom_up): Don't include template instantiations
15311 (generate_skeleton): Likewise.
15312 (break_out_comdat_types): Move recursive call to break out nested
15314 (prune_unused_types_mark_generic_parms_dies): Call
15315 is_template_parameter.
15317 2013-07-26 Ian Bolton <ian.bolton@arm.com>
15319 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
15320 uses vector registers.
15321 * config/aarch64/iterators.md: Add attributes rtn and vas.
15323 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15324 Richard Earnshaw <richard.earnshaw@arm.com>
15326 * combine.c (simplify_comparison): Re-canonicalize operands
15328 * config/arm/arm.md (movcond_addsi): New splitter.
15330 2013-07-25 Sterling Augustine <saugustine@google.com>
15332 * dwarf2out.c (size_of_pubnames): Move code to...
15333 (include_pubname_in_output): ...here. New.
15334 (want_pubnames): Rearrange.
15335 (output_pubnames): Call include_pubname_in_output. Move assertion.
15337 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
15339 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
15341 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
15344 * doc/extend.texi: Remove obsolete builtins. Fix
15345 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
15347 2013-07-25 Jan Hubicka <jh@suse.cz>
15349 * cgraph.c (release_function_body): Break out from ...
15350 (cgraph_release_function_body): ... this one; also release DECL_RESULT
15351 and DECL_ARGUMENTS.
15352 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
15353 old_tree in the map.
15354 (create_specialized_node): Update.
15355 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
15357 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
15358 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
15359 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
15360 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
15363 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15365 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
15366 addsi3_carryin_alt2_<optab>): Correct output template.
15368 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15370 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
15371 Adjust for arm_restrict_it.
15372 Remove trailing whitespace.
15374 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
15376 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
15377 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
15379 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
15381 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
15383 PR rtl-optimization/57960
15384 * lra-constraints.c (process_alt_operands): Use the right mode
15385 when checking strict_low.
15387 2013-07-25 Jan Hubicka <jh@suse.cz>
15389 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
15390 * cgraph.c (cgraph_remove_node): Do not release function body
15391 when in cgraph streaming.
15392 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
15393 in other partitions are not considered reachable; fix handling of
15396 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15398 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
15400 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15405 * config/arm/arm.md ("*sibcall_insn): Replace use of
15406 Ss with US. Adjust output for v5 and v4t.
15407 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
15409 * config/arm/constraints.md ("Ss"): Rename to US.
15411 2013-07-25 Terry Guo <terry.guo@arm.com>
15413 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
15414 shift_add/shift_sub0/shift_sub1 RTXs.
15416 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
15417 Anton Blanchard <anton@au1.ibm.com>
15419 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
15420 (altivec_vpks<VI_char>ss): Likewise.
15421 (altivec_vpks<VI_char>us): Likewise.
15422 (altivec_vpku<VI_char>us): Likewise.
15423 (altivec_vpku<VI_char>um): Likewise.
15425 2013-07-24 David Malcolm <dmalcolm@redhat.com>
15427 Introduce context class.
15429 * Makefile.in (CONTEXT_H): New.
15430 (OBJS): Add context.o.
15431 (toplev.o): Add CONTEXT_H to dependencies.
15434 * toplev.c (general_init): Create the singleton gcc::context instance.
15440 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
15442 PR rtl-optimization/57968
15443 * mode-switching.c (create_pre_exit): Allow instructions that
15444 don't set a return register to need a non-exit mode.
15446 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15447 Anton Blanchard <anton@au1.ibm.com>
15449 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
15450 operands to vperm for little endian.
15451 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
15452 of lvsl to create the control mask for a vperm for little endian.
15454 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15455 Anton Blanchard <anton@au1.ibm.com>
15457 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
15458 two operands for little-endian.
15460 2013-07-23 Steve Ellcey <sellcey@mips.com>
15462 * config/mips/mips.c (mips_case_values_threshold): New.
15463 (TARGET_CASE_VALUES_THRESHOLD): Define.
15465 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15466 Anton Blanchard <anton@au1.ibm.com>
15468 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
15469 selection of field for vector splat in little endian mode.
15471 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
15473 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
15474 expanders to rs6000.md.
15475 (ior<mode>3): Likewise.
15476 (and<mode>3): Likewise.
15477 (one_cmpl<mode>2): Likewise.
15478 (nor<mode>3): Likewise.
15479 (andc<mode>3): Likewise.
15480 (eqv<mode>3): Likewise.
15481 (nand<mode>3): Likewise.
15482 (orc<mode>3): Likewise.
15484 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
15487 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
15488 to split multi-word logical operations.
15489 (rs6000_split_logical_di): Likewise.
15490 (rs6000_split_logical): Likewise.
15492 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
15493 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
15494 and allow TImode operations in 32-bit.
15495 (vsx_and<mode>3_64bit): Likewise.
15496 (vsx_ior<mode>3_32bit): Likewise.
15497 (vsx_ior<mode>3_64bit): Likewise.
15498 (vsx_xor<mode>3_32bit): Likewise.
15499 (vsx_xor<mode>3_64bit): Likewise.
15500 (vsx_one_cmpl<mode>2_32bit): Likewise.
15501 (vsx_one_cmpl<mode>2_64bit): Likewise.
15502 (vsx_nor<mode>3_32bit): Likewise.
15503 (vsx_nor<mode>3_64bit): Likewise.
15504 (vsx_andc<mode>3_32bit): Likewise.
15505 (vsx_andc<mode>3_64bit): Likewise.
15506 (vsx_eqv<mode>3_32bit): Likewise.
15507 (vsx_eqv<mode>3_64bit): Likewise.
15508 (vsx_nand<mode>3_32bit): Likewise.
15509 (vsx_nand<mode>3_64bit): Likewise.
15510 (vsx_orc<mode>3_32bit): Likewise.
15511 (vsx_orc<mode>3_64bit): Likewise.
15513 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
15514 logical types in GPRs.
15516 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
15517 logical insns to rs6000.md, and allow TImode operations in
15519 (altivec_ior<mode>3): Likewise.
15520 (altivec_xor<mode>3): Likewise.
15521 (altivec_one_cmpl<mode>2): Likewise.
15522 (altivec_nor<mode>3): Likewise.
15523 (altivec_andc<mode>3): Likewise.
15525 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
15526 attributes for moving the 128-bit logical operations into
15528 (BOOL_REGS_OUTPUT): Likewise.
15529 (BOOL_REGS_OP1): Likewise.
15530 (BOOL_REGS_OP2): Likewise.
15531 (BOOL_REGS_UNARY): Likewise.
15532 (BOOL_REGS_AND_CR0): Likewise.
15533 (one_cmpl<mode>2): Add support for DI logical operations on
15534 32-bit, splitting the operations to 32-bit.
15535 (anddi3): Likewise.
15536 (iordi3): Likewise.
15537 (xordi3): Likewise.
15538 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
15539 changes to combine the 32/64-bit code, allow logical operations on
15540 TI mode in 32-bit, and to use similar match_operator patterns like
15541 scalar mode uses. Combine the Altivec and VSX code for logical
15542 operations, and move it here.
15543 (ior<mode>3, 128-bit types): Likewise.
15544 (xor<mode>3, 128-bit types): Likewise.
15545 (one_cmpl<mode>3, 128-bit types): Likewise.
15546 (nor<mode>3, 128-bit types): Likewise.
15547 (andc<mode>3, 128-bit types): Likewise.
15548 (eqv<mode>3, 128-bit types): Likewise.
15549 (nand<mode>3, 128-bit types): Likewise.
15550 (orc<mode>3, 128-bit types): Likewise.
15551 (and<mode>3_internal): Likewise.
15552 (bool<mode>3_internal): Likewise.
15553 (boolc<mode>3_internal1): Likewise.
15554 (boolc<mode>3_internal2): Likewise.
15555 (boolcc<mode>3_internal1): Likewise.
15556 (boolcc<mode>3_internal2): Likewise.
15557 (eqv<mode>3_internal1): Likewise.
15558 (eqv<mode>3_internal2): Likewise.
15559 (one_cmpl1<mode>3_internal): Likewise.
15561 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15563 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15564 Rename flag_stack_usage to flag_stack_usage_info.
15566 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15568 * config/microblaze/sync.md: New file.
15569 * config/microblaze/microblaze.md: Include sync.md
15570 * config/microblaze/microblaze.c: Add print_operand 'y'.
15571 * config/microblaze/constraints.md: Add memory_contraint
15572 'Q' which is a single register.
15574 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
15576 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
15578 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15580 * reload.c (find_reloads): Exit loop once we find this operand
15581 cannot be reloaded somehow for this alternative.
15583 * reload.c (find_reloads): Exit loop once we find a hard register.
15585 * rtlanal.c (computed_jump_p): Exit loop once we find label
15588 * i386.c (ix86_pad_returns): Exit loop after setting replace.
15590 * cfgloopmanip.c (remove_path): Exit loop after setting
15593 * gensupport.c (subst_dup): Avoid loop if code is not
15594 MATCH_DUP nor MATCH_OP_DUP.
15596 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
15598 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
15600 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15602 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
15603 true for SP_REGNUM if mode == ptr_mode.
15604 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
15605 with value R0_REGNUM + 31.
15607 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15609 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
15610 pad pointer-typed argument downward.
15612 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15614 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
15615 and __ILP32__ when the ILP32 model is in use.
15617 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15619 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
15620 (aarch64_load_symref_appropriately): In the case of
15621 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
15622 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
15623 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
15624 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
15625 if 'mode' doesn't equal to 'ptr_mode'.
15626 (aarch64_output_mi_thunk): Add an assertion on the alignment of
15627 'vcall_offset'; change to call aarch64_emit_move differently depending
15628 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
15629 to calculate the upper bound of 'vcall_offset'.
15630 (aarch64_cannot_force_const_mem): Change to also return true if
15632 (aarch64_legitimize_reload_address): In the case of large
15633 displacements, add new local variable 'xmode' and an assertion
15634 based on it; change to use 'xmode' to generate the new rtx and
15636 (aarch64_asm_trampoline_template): Change to generate the template
15637 differently depending on TARGET_ILP32 or not; change to use
15638 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
15639 (aarch64_trampoline_size): Removed.
15640 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
15641 and replace immediate literals with it. Change to use 'ptr_mode'
15642 instead of 'DImode' and call convert_memory_address if the mode
15643 of 'fnaddr' doesn't equal to 'ptr_mode'.
15644 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
15646 (aarch64_elf_asm_destructor): Likewise.
15647 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
15648 on TARGET_ILP32 instead of aarch64_trampoline_size.
15649 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
15650 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
15651 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
15652 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
15653 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
15654 (storewb_pair<GPI:mode>_<P:mode>): ... this.
15655 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
15656 depending on the value of 'mode'.
15657 (add_losym_<mode>): New.
15658 (ldr_got_small_<mode>): New, based on ldr_got_small.
15659 (ldr_got_small): Remove.
15660 (ldr_got_small_sidi): New.
15661 * config/aarch64/iterators.md (P): New.
15662 (PTR): Change to 'ptr_mode' in the condition.
15664 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15666 * config.gcc (aarch64*-*-*): Support --with-abi.
15667 (aarch64*-*-elf): Support --with-multilib-list.
15668 (aarch64*-*-linux*): Likewise.
15669 (supported_defaults): Add abi to aarch64*-*-*.
15670 * configure.ac: Mention AArch64 for --with-multilib-list.
15671 * configure: Re-generated.
15672 * config/aarch64/biarchilp32.h: New file.
15673 * config/aarch64/biarchlp64.h: New file.
15674 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
15676 (MULTILIB_DEFAULTS): Ditto.
15677 (DRIVER_SELF_SPECS): Ditto.
15678 (ASM_SPEC): Update to also substitute -mabi.
15679 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
15680 file whose name depends on -mabi= and -mbig-endian.
15681 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
15683 (POINTER_SIZE): New define.
15684 (POINTERS_EXTEND_UNSIGNED): Ditto.
15685 (enum aarch64_abi_type): New enumeration tag.
15686 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
15687 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
15688 (TARGET_ILP32): New define.
15689 * config/aarch64/aarch64.opt (mabi): New.
15690 (aarch64_abi): New.
15691 (ilp32, lp64): New values for -mabi.
15692 * config/aarch64/t-aarch64 (comma): New define.
15693 (MULTILIB_OPTIONS): Ditto.
15694 (MULTILIB_DIRNAMES): Ditto.
15695 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
15696 * doc/invoke.texi: Document -mabi for AArch64.
15698 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
15700 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
15702 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15703 Anton Blanchard <anton@au1.ibm.com>
15705 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
15706 endianness when selecting field to splat.
15708 2013-07-22 Eric Christopher <echristo@gmail.com>
15710 * dwarf2out.c (die_odr_checksum): New function to use
15711 CHECKSUM_ macros and ULEB128 for DIE tag.
15712 (generate_type_signature): Use.
15714 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
15716 * config.gcc (sparc*-*-*): Accept leon3 processor.
15717 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
15718 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
15719 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
15720 * config/sparc/sparc.opt (enum processor_type): Add leon3.
15721 (mfix-ut699): Adjust comment.
15722 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
15723 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
15724 (CPP_CPU_SPEC): Likewise.
15725 (ASM_CPU_SPEC): Likewise.
15726 * config/sparc/sparc.c (leon3_cost): New constant.
15727 (sparc_option_override): Add leon3 support.
15728 (mem_ref): New function.
15729 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
15730 (sparc_do_work_around_errata): Look into the instruction in the delay
15731 slot and adjust accordingly. Add fix for the data cache nullify issues
15732 of the UT699. Change insertion position for the NOP.
15733 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
15734 (leon3_load): New reservation.
15735 (leon_store): Bump latency to 2.
15736 (grfpu): New automaton.
15737 (grfpu_alu): New unit.
15738 (grfpu_ds): Likewise.
15739 (leon_fp_alu): Adjust.
15740 (leon_fp_mult): Delete.
15741 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
15742 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
15743 * config/sparc/sparc.md (cpu): Add leon3.
15744 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
15745 (swapsi): Likewise.
15746 (atomic_test_and_set): Likewise.
15747 (ldstub): Likewise.
15749 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
15751 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
15752 default for R5900 targets.
15753 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
15754 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
15755 * config/mips/mips.c (mips_option_override): Report an error for
15756 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
15757 for -march=r5900 -mhard-float.
15759 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15761 * df-problems.c (can_move_insns_across): Exit loop once we
15762 find a non-fixed, non-global register.
15764 * ipa-pure-const.c (propagate_nothrow): Exit loop after
15767 * omega.c (omega_eliminate_red): Break after setting red_found.
15768 (omega_problem_has_red_equations): Similarly after setting found.
15769 (omega_query_variable): Similarly after setting coupled.
15771 2013-07-22 Marek Polacek <polacek@redhat.com>
15773 * gimplify.c: Don't include gimple.h twice.
15775 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15777 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
15778 instead of TARGET_THUMB1.
15779 (Pz): New constraint.
15780 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
15782 (compare_negsi_si): Likewise.
15783 (compare_addsi2_op0): Likewise.
15784 (compare_addsi2_op1): Likewise.
15785 (addsi3_carryin_<optab>): Likewise.
15786 (addsi3_carryin_alt2_<optab>): Likewise.
15787 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
15788 for arm_restrict_it.
15789 (subsi3_carryin): Likewise.
15790 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
15791 (minmax_arithsi): Disable for arm_restrict_it.
15792 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
15793 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
15794 (satsi_<SAT:code>_shift): Likewise.
15795 (arm_shiftsi3): Add alternative for 16-bit encoding.
15796 (arm32_movhf): Disable for arm_restrict_it.
15797 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
15798 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
15800 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
15802 * config/arm/arm.md (attribute "insn"): Delete.
15803 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
15804 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
15805 (not_shiftsi): Update for attribute change.
15806 (not_shiftsi_compare0): Likewise.
15807 (not_shiftsi_compare0_scratch): Likewise.
15808 (arm_one_cmplsi2): Likewise.
15809 (thumb1_one_cmplsi2): Likewise.
15810 (notsi_compare0): Likewise.
15811 (notsi_compare0_scratch): Likewise.
15812 (thumb1_movdi_insn): Likewise.
15813 (arm_movsi_insn): Likewise.
15814 (movhi_insn_arch4): Likewise.
15815 (movhi_bytes): Likewise.
15816 (arm_movqi_insn): Likewise.
15817 (thumb1_movqi_insn): Likewise.
15818 (arm32_movhf): Likewise.
15819 (thumb1_movhf): Likewise.
15820 (arm_movsf_soft_insn): Likewise.
15821 (thumb1_movsf_insn): Likewise.
15822 (thumb_movdf_insn): Likewise.
15823 (movsicc_insn): Likewise.
15824 (movsfcc_soft_insn): Likewise.
15825 (and_scc): Likewise.
15826 (cond_move): Likewise.
15827 (if_move_not): Likewise.
15828 (if_not_move): Likewise.
15829 (if_shift_move): Likewise.
15830 (if_move_shift): Likewise.
15831 (if_shift_shift): Likewise.
15832 (if_not_arith): Likewise.
15833 (if_arith_not): Likewise.
15834 (cond_move_not): Likewise.
15835 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
15836 (neon_mov<mode>): Likewise.
15837 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
15838 (thumb2_movsi_vfp): Likewise.
15839 (movsf_vfp): Likewise.
15840 (thumb2_movsf_vfp): Likewise.
15841 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15843 (cortexa7_older_only): Likewise.
15844 (cortexa7_younger): Likewise.
15845 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
15846 (1020alu_shift_op): Likewise.
15847 (1020alu_shift_reg_op): Likewise.
15848 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15849 (alu_shift_op): Likewise.
15850 (alu_shift_reg_op): Likewise.
15851 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15852 (11_alu_shift_op): Likewise.
15853 (11_alu_shift_reg_op): Likewise.
15854 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15855 (9_alu_shift_reg_op): Likewise.
15856 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15858 (cortex_a15_alu_shift): Likewise.
15859 (cortex_a15_alu_shift_reg): Likewise.
15860 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15861 (cortex_a5_alu_shift): Likewise.
15862 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15864 (cortex_a53_alu_shift): Likewise.
15865 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15867 (cortex_a7_alu_reg): Likewise.
15868 (cortex_a7_alu_shift): Likewise.
15869 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15870 (cortex_a8_alu_shift): Likewise.
15871 (cortex_a8_alu_shift_reg): Likewise.
15872 (cortex_a8_mov): Likewise.
15873 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15874 (cortex_a9_dp_shift): Likewise.
15875 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15876 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15877 (cortex_r4_mov): Likewise.
15878 (cortex_r4_alu_shift): Likewise.
15879 (cortex_r4_alu_shift_reg): Likewise.
15880 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15881 (526_alu_shift_op): Likewise.
15882 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15883 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15884 (626te_alu_shift_op): Likewise.
15885 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15886 (726te_alu_op): Likewise.
15887 (726te_alu_shift_op): Likewise.
15888 (726te_alu_shift_reg_op): Likewise.
15889 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15890 (mp626_alu_shift_op): Likewise.
15891 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15892 (pj4_alu_e1_conds): Likewise.
15893 (pj4_alu): Likewise.
15894 (pj4_alu_conds): Likewise.
15895 (pj4_shift): Likewise.
15896 (pj4_shift_conds): Likewise.
15897 (pj4_alu_shift): Likewise.
15898 (pj4_alu_shift_conds): Likewise.
15900 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15902 * config/arm/predicates.md (shiftable_operator_strict_it):
15904 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15905 Disable cond_exec version for arm_restrict_it.
15906 (thumb2_smaxsi3): Convert to generate cond_exec.
15907 (thumb2_sminsi3): Likewise.
15908 (thumb32_umaxsi3): Likewise.
15909 (thumb2_uminsi3): Likewise.
15910 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15911 (thumb2_neg_abssi2): Likewise.
15912 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15913 (thumb2_movsicc_insn): Adjust alternatives.
15914 (thumb2_mov_negscc): Disable for arm_restrict_it.
15915 (thumb2_mov_negscc_strict_it): New pattern.
15916 (thumb2_mov_notscc_strict_it): New pattern.
15917 (thumb2_mov_notscc): Disable for arm_restrict_it.
15918 (thumb2_ior_scc): Likewise.
15919 (thumb2_ior_scc_strict_it): New pattern.
15920 (thumb2_cond_move): Adjust for arm_restrict_it.
15921 (thumb2_cond_arith): Disable for arm_restrict_it.
15922 (thumb2_cond_arith_strict_it): New pattern.
15923 (thumb2_cond_sub): Adjust for arm_restrict_it.
15924 (thumb2_movcond): Likewise.
15925 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15926 (thumb2_zero_extendhisi2_v6): Likewise.
15927 (thumb2_zero_extendqisi2_v6): Likewise.
15928 (orsi_notsi_si): Likewise.
15929 (orsi_not_shiftsi_si): Likewise.
15931 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15933 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15934 instruction sequence is 1 byte shorter.
15936 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15938 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15939 it is not needed after split.
15941 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15944 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15945 second label for nonlocal goto receivers. Don't output pic base labels
15946 unless we're producing PIC; mark that action unreachable().
15947 (ix86_save_reg): If the function contains a nonlocal label, save the
15949 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15950 * config/darwin.c (emitted_pic_label_num): New GTY.
15951 (update_pic_label_number_if_needed): New.
15952 (machopic_output_function_base_name): Adjust for nonlocal receiver
15954 (machopic_should_output_picbase_label): New.
15955 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15956 (nonlocal_goto_receiver): New insn and split.
15958 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15960 * config/aarch64/aarch64-builtins.c
15961 (aarch64_fold_builtin): Fold abs in all modes.
15962 * config/aarch64/aarch64-simd-builtins.def
15963 (abs): Enable for all modes.
15964 * config/aarch64/arm_neon.h
15965 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15966 (vabs_f64): Add missing intrinsic.
15968 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15970 * config/aarch64/arm_neon.h (vabs_s64): New function
15972 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15975 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15976 * config/avr/avr.md (adjust_len): Add `round'.
15977 * config/avr/avr-protos.h (avr_out_round): New prototype.
15978 (avr_out_plus): Add `out_label' argument.
15979 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15980 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15981 Handle the case where `insn' is just a pattern.
15982 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15983 (avr_out_round): New function.
15984 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15986 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15988 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15989 Add check for flag_stack_usage to handle -fstack-usage support
15991 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15993 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
15994 interaction for new Power8 flags and VSX.
15996 2013-07-18 Sriraman Tallam <tmsriram@google.com>
15998 PR middle-end/57698
15999 * tree-inline.c (expand_call_inline): Emit errors during
16000 early_inlining only if optimization is not turned on.
16002 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16006 * passes.c (init_optimization_passes): Move the construction of
16007 the pass hierarchy into a new passes.def file.
16009 * Makefile.in (passes.o): Add dependency on passes.def.
16011 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16013 * passes.c (init_optimization_passes): Introduce macros for
16014 constructing the tree of passes (INSERT_PASSES_AFTER,
16015 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
16016 TERMINATE_PASS_LIST).
16018 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
16019 Wei Mi <wmi@google.com>
16021 PR rtl-optimization/57878
16022 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
16024 (reload_pseudo_compare_func): Check nregs first for reload
16027 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16029 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
16031 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
16033 * read-rtl.c (validate_const_int): Once an invalid character is
16034 seen, quit the loop.
16036 * gengtype.c (write_roots): Similarly once we find the "deletable"
16037 or "if_marked" option.
16039 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
16041 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
16042 "xtab" and "sat". Move value "clz" from here to ...
16043 (attriubte "type"): ... here.
16044 (satsi_<SAT:code>): Delete "insn" attribute.
16045 (satsi_<SAT:code>_shift): Likewise.
16046 (arm_zero_extendqisi2addsi): Likewise.
16047 (arm_extendqisi2addsi): Likewise.
16048 (clzsi2): Update for attribute changes.
16049 (rbitsi2): Likewise.
16050 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
16052 (arm_usatsihi): Likewise.
16053 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
16055 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
16057 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
16058 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
16059 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
16060 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
16061 in alphabetical order.
16062 (attribute "core_cycles"): Update for attribute changes.
16063 (arm_addsi3): Likewise.
16064 (addsi3_compare0): Likewise.
16065 (addsi3_compare0_scratch): Likewise.
16066 (addsi3_compare_op1): Likewise.
16067 (addsi3_compare_op2): Likewise.
16068 (compare_addsi2_op0): Likewise.
16069 (compare_addsi2_op1): Likewise.
16070 (addsi3_carryin_shift_<optab>): Likewise.
16071 (subsi3_carryin_shift): Likewise.
16072 (rsbsi3_carryin_shift): Likewise.
16073 (arm_subsi3_insn): Likewise.
16074 (subsi3_compare0): Likewise.
16075 (subsi3_compare): Likewise.
16076 (arm_andsi3_insn): Likewise.
16077 (thumb1_andsi3_insn): Likewise.
16078 (andsi3_compare0): Likewise.
16079 (andsi3_compare0_scratch): Likewise.
16080 (zeroextractsi_compare0_scratch
16081 (andsi_not_shiftsi_si): Likewise.
16082 (iorsi3_insn): Likewise.
16083 (iorsi3_compare0): Likewise.
16084 (iorsi3_compare0_scratch): Likewise.
16085 (arm_xorsi3): Likewise.
16086 (thumb1_xorsi3_insn): Likewise.
16087 (xorsi3_compare0): Likewise.
16088 (xorsi3_compare0_scratch): Likewise.
16089 (satsi_<SAT:code>_shift): Likewise.
16091 (arm_shiftsi3): Likewise.
16092 (shiftsi3_compare0): Likewise.
16093 (not_shiftsi): Likewise.
16094 (not_shiftsi_compare0): Likewise.
16095 (not_shiftsi_compare0_scratch): Likewise.
16096 (arm_one_cmplsi2): Likewise.
16097 (thumb_one_complsi2): Likewise.
16098 (notsi_compare0): Likewise.
16099 (notsi_compare0_scratch): Likewise.
16100 (thumb1_zero_extendhisi2): Likewise.
16101 (arm_zero_extendhisi2): Likewise.
16102 (arm_zero_extendhisi2_v6): Likewise.
16103 (arm_zero_extendhisi2addsi): Likewise.
16104 (thumb1_zero_extendqisi2): Likewise.
16105 (thumb1_zero_extendqisi2_v6): Likewise.
16106 (arm_zero_extendqisi2): Likewise.
16107 (arm_zero_extendqisi2_v6): Likewise.
16108 (arm_zero_extendqisi2addsi): Likewise.
16109 (thumb1_extendhisi2): Likewise.
16110 (arm_extendhisi2): Likewise.
16111 (arm_extendhisi2_v6): Likewise.
16112 (arm_extendqisi): Likewise.
16113 (arm_extendqisi_v6): Likewise.
16114 (arm_extendqisi2addsi): Likewise.
16115 (thumb1_extendqisi2): Likewise.
16116 (thumb1_movdi_insn): Likewise.
16117 (arm_movsi_insn): Likewise.
16118 (movsi_compare0): Likewise.
16119 (movhi_insn_arch4): Likewise.
16120 (movhi_bytes): Likewise.
16121 (arm_movqi_insn): Likewise.
16122 (thumb1_movqi_insn): Likewise.
16123 (arm32_movhf): Likewise.
16124 (thumb1_movhf): Likewise.
16125 (arm_movsf_soft_insn): Likewise.
16126 (thumb1_movsf_insn): Likewise.
16127 (movdf_soft_insn): Likewise.
16128 (thumb_movdf_insn): Likewise.
16129 (arm_cmpsi_insn): Likewise.
16130 (cmpsi_shiftsi): Likewise.
16131 (cmpsi_shiftsi_swp): Likewise.
16132 (arm_cmpsi_negshiftsi_si): Likewise.
16133 (movsicc_insn): Likewise.
16134 (movsfcc_soft_insn): Likewise.
16135 (arith_shiftsi): Likewise.
16136 (arith_shiftsi_compare0
16137 (arith_shiftsi_compare0_scratch
16138 (sub_shiftsi): Likewise.
16139 (sub_shiftsi_compare0
16140 (sub_shiftsi_compare0_scratch
16141 (and_scc): Likewise.
16142 (cond_move): Likewise.
16143 (if_plus_move): Likewise.
16144 (if_move_plus): Likewise.
16145 (if_move_not): Likewise.
16146 (if_not_move): Likewise.
16147 (if_shift_move): Likewise.
16148 (if_move_shift): Likewise.
16149 (if_shift_shift): Likewise.
16150 (if_not_arith): Likewise.
16151 (if_arith_not): Likewise.
16152 (cond_move_not): Likewise.
16153 (thumb1_ashlsi3): Set type attribute.
16154 (thumb1_ashrsi3): Likewise.
16155 (thumb1_lshrsi3): Likewise.
16156 (thumb1_rotrsi3): Likewise.
16157 (shiftsi3_compare0_scratch): Likewise.
16158 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
16159 (neon_mov<mode>): Likewise.
16160 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
16162 (thumb2_movsi_insn): Likewise.
16163 (thumb2_cmpsi_neg_shiftsi): Likewise.
16164 (thumb2_extendqisi_v6): Likewise.
16165 (thumb2_zero_extendhisi2_v6): Likewise.
16166 (thumb2_zero_extendqisi2_v6): Likewise.
16167 (thumb2_shiftsi3_short): Likewise.
16168 (thumb2_addsi3_compare0_scratch): Likewise.
16169 (orsi_not_shiftsi_si): Likewise.
16170 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
16171 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
16173 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
16174 (1020alu_shift_op): Likewise.
16175 (1020alu_shift_reg_op): Likewise.
16176 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
16177 (alu_shift_op): Likewise.
16178 (alu_shift_reg_op): Likewise.
16179 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
16180 (11_alu_shift_op): Likewise.
16181 (11_alu_shift_reg_op): Likewise.
16182 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
16183 (9_alu_shift_reg_op): Likewise.
16184 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
16186 (cortex_a15_alu_shift): Likewise.
16187 (cortex_a15_alu_shift_reg): Likewise.
16188 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
16190 (cortex_a5_alu_shift): Likewise.
16191 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
16193 (cortex_a53_alu_shift): Likewise.
16194 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
16196 (cortex_a7_alu_reg): Likewise.
16197 (cortex_a7_alu_shift): Likewise.
16198 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
16200 (cortex_a8_alu_shift): Likewise.
16201 (cortex_a8_alu_shift_reg): Likewise.
16202 (cortex_a8_mov): Likewise.
16203 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
16204 (cortex_a9_dp_shift): Likewise.
16205 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
16207 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
16209 (cortex_r4_mov): Likewise.
16210 (cortex_r4_alu_shift): Likewise.
16211 (cortex_r4_alu_shift_reg): Likewise.
16212 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
16213 (526_alu_shift_op): Likewise.
16214 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
16215 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
16216 (626te_alu_shift_op): Likewise.
16217 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
16218 (726te_alu_op): Likewise.
16219 (726te_alu_shift_op): Likewise.
16220 (726te_alu_shift_reg_op): Likewise.
16221 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
16222 (mp626_alu_shift_op): Likewise.
16223 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
16224 (pj4_alu_e1_conds): Likewise.
16225 (pj4_alu): Likewise.
16226 (pj4_alu_conds): Likewise.
16227 (pj4_shift): Likewise.
16228 (pj4_shift_conds): Likewise.
16229 (pj4_alu_shift): Likewise.
16230 (pj4_alu_shift_conds): Likewise.
16231 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
16233 (cortexa7_older_only): Likewise.
16234 (cortexa7_younger): Likewise.
16236 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16238 * ipa-pure-const.c (generate_summary): Rename to...
16239 (pure_const_generate_summary): ... this.
16241 2013-07-17 Iain Sandoe <iain@codesourcery.com>
16243 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
16245 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
16248 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
16251 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16253 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
16254 enabled without -march=zEC12.
16255 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
16258 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16260 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
16261 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
16262 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
16263 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
16264 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
16265 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
16266 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
16268 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
16269 and ISA_HAS_NMADD3_NMSUB3.
16270 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
16271 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
16272 (nmsub4<mode>, nmsub3<mode>): Likewise.
16273 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
16275 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16277 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
16278 TARGET_MIPS5400 checking.
16280 2013-07-16 Jakub Jelinek <jakub@redhat.com>
16281 Peter Bergner <bergner@vnet.ibm.com>
16283 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
16284 registers in the comment.
16285 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
16286 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
16287 rather than FIRST_PSEUDO_REGISTERS.
16289 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
16291 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
16292 enable extra ISA flags with TARGET_HTM.
16294 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16296 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
16299 2013-07-15 Cong Hou <congh@google.com>
16301 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
16302 in compare function for sorting.
16304 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
16306 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
16307 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
16308 * config/rs6000/rs6000.opt: Add -mhtm option.
16309 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
16310 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
16311 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16312 __HTM__ if the HTM instructions are available.
16313 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
16314 htm_spr_reg_operand): New define_predicates.
16315 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
16316 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
16318 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
16319 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
16320 HTM builtin functions.
16321 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
16322 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
16323 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
16324 (rs6000_builtin_mask_calculate): Likewise.
16325 (rs6000_option_override_internal): Likewise.
16326 (bdesc_htm): Add new HTM builtin support.
16327 (htm_spr_num): New function.
16328 (htm_spr_regno): Likewise.
16329 (rs6000_htm_spr_icode): Likewise.
16330 (htm_expand_builtin): Likewise.
16331 (htm_init_builtins): Likewise.
16332 (rs6000_expand_builtin): Add support for HTM builtin functions.
16333 (rs6000_init_builtins): Likewise.
16334 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
16336 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
16337 (TARGET_HTM, MASK_HTM): Define macros.
16338 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
16339 (FIXED_REGISTERS): Likewise.
16340 (CALL_USED_REGISTERS): Likewise.
16341 (CALL_REALLY_USED_REGISTERS): Likewise.
16342 (REG_ALLOC_ORDER): Likewise.
16343 (enum reg_class): Likewise.
16344 (REG_CLASS_NAMES): Likewise.
16345 (REG_CLASS_CONTENTS): Likewise.
16346 (REGISTER_NAMES): Likewise.
16347 (ADDITIONAL_REGISTER_NAMES): Likewise.
16348 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
16349 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
16350 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
16351 * config/rs6000/htm.md: New file.
16352 * config/rs6000/htmintrin.h: New file.
16353 * config/rs6000/htmxlintrin.h: New file.
16355 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
16357 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16358 Define SYMBOL_TINY_GOT, update comment.
16359 * config/aarch64/aarch64.c
16360 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
16361 (aarch64_expand_mov_immediate): Likewise.
16362 (aarch64_print_operand): Likewise.
16363 (aarch64_classify_symbol): Likewise.
16364 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
16365 (ldr_got_tiny): Define.
16367 2013-07-13 Tobias Grosser <tobias@grosser.es>
16369 PR tree-optimization/54094
16370 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
16371 scheduling dimension for the parallelism check from the polyhedral
16372 information in the AST.
16373 * graphite-dependences.c (carries_deps): Do not assume the schedule is
16376 2013-07-13 Jason Merrill <jason@redhat.com>
16378 * print-tree.c (debug_vec_tree): Use debug_raw.
16379 (debug_raw (vec<tree, va_gc> &)): New.
16380 (debug_raw (vec<tree, va_gc> *)): New.
16381 * tree.h: Declare them.
16383 2013-07-13 Bin Cheng <bin.cheng@arm.com>
16385 * ifcvt.c (ifcvt_after_combine): New static variable.
16386 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
16388 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
16389 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
16390 rest_of_handle_if_after_reload): Pass new argument for if_convert.
16392 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
16394 * config/mips/mips.c (mips_expand_call): Remove empty statement.
16396 2013-07-12 Michael Matz <matz@suse.de>
16398 PR middle-end/55771
16399 * convert.c (convert_to_real): Reject non-float inner types.
16401 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
16403 * config/aarch64/aarch64-protos.h
16404 (aarch64_simd_immediate_valid_for_move): Remove.
16405 * config/aarch64/aarch64.c (simd_immediate_info): New member.
16406 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
16408 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
16410 2013-07-11 Steve Ellcey <sellcey@mips.com>
16412 * config/mips/mips.c (mips_conditional_register_usage): Do not
16413 use t[0-7] registers in MIPS16 mode when optimizing for size.
16415 2013-07-11 Sriraman Tallam <tmsriram@google.com>
16417 * config/i386/i386.c (dispatch_function_versions): Fix array
16418 indexing of function_version_info to match actual_versions.
16420 2013-07-11 Teresa Johnson <tejohnson@google.com>
16422 * vec.h (struct va_gc): Move release out-of-line.
16423 (va_gc::release): Call ggc_free on released vec.
16425 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
16427 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
16428 Require GOT register as additional operand in UNSPEC.
16429 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
16430 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
16431 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
16432 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
16433 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
16434 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
16435 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
16437 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
16440 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
16441 name seen by assembler/linker rather if available.
16443 2013-07-11 Andreas Schwab <schwab@suse.de>
16445 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
16447 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
16449 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
16451 2013-07-10 Joseph Myers <joseph@codesourcery.com>
16453 * doc/tm.texi.in: Move hook documentation to ....
16454 * target.def: ... here.
16456 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
16457 text on @hook line.
16458 * doc/tm.texi: Regenerate.
16460 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
16463 * doc/invoke.texi: Document -Wconditionally-supported.
16465 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16468 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
16471 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16474 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
16475 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
16476 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
16477 Remove duplicate devices.
16478 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
16479 * config/avr/t-multilib: Regenerate.
16480 * config/avr/avr-tables.opt: Regenerate.
16481 * doc/avr-mmcu.texi: Regenerate.
16483 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16486 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
16488 2013-07-10 Graham Stott <graham.stott@btinternet.com>
16490 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
16491 the cost of MULT when optimizing for size.
16493 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16495 * config/cr16/cr16-protos.h: Don't include target.h.
16497 2013-07-09 Joseph Myers <joseph@codesourcery.com>
16499 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
16500 adjust register size for TDmode and TFmode for VSX registers.
16502 2013-07-08 Kai Tietz <ktietz@redhat.com>
16505 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
16506 hook_bool_const_tree_true.
16508 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16510 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
16511 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
16512 * config/s390/s390.md: Define FPR*_REGNUM constants.
16513 Fix FPR2_REGNUM constant (18 -> 17).
16514 ("*trunc<BFP:mode><DFP_ALL:mode>2")
16515 ("*trunc<DFP_ALL:mode><BFP:mode>2")
16516 ("trunc<BFP:mode><DFP_ALL:mode>2")
16517 ("trunc<DFP_ALL:mode><BFP:mode>2")
16518 ("*extend<BFP:mode><DFP_ALL:mode>2")
16519 ("*extend<DFP_ALL:mode><BFP:mode>2")
16520 ("extend<BFP:mode><DFP_ALL:mode>2")
16521 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
16524 2013-07-08 Graham Stott <graham.stott@btinternet.com>
16526 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
16528 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16530 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
16531 and cfun_fpr_bit_p to cfun_fpr_save_p.
16532 (s390_frame_area, s390_register_info, s390_frame_info)
16533 (s390_emit_prologue, s390_emit_epilogue)
16534 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
16536 * config/s390/s390.h: Define *_REGNUM macros for floating point
16539 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
16541 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
16543 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
16545 PR rtl-optimization/57786
16546 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
16547 and break out of the loop when it is set to false.
16549 2013-07-08 Jakub Jelinek <jakub@redhat.com>
16552 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
16553 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
16554 (const_int 63)) 0)).
16555 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
16556 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
16557 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
16559 PR rtl-optimization/57829
16560 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
16561 mask bits outside of mode are just sign-extension from mode to HWI.
16563 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
16565 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
16566 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
16567 adjust_address instead of change_address to keep info about alignment.
16568 (emit_strmov): Remove.
16569 (emit_memmov): New function.
16570 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
16571 (expand_movmem_epilogue): Likewise and return updated rtx for
16573 (expand_constant_movmem_prologue): Likewise and return updated rtx for
16574 destination and source.
16575 (decide_alignment): Refactor, handle vector_loop.
16576 (ix86_expand_movmem): Likewise.
16577 (ix86_expand_setmem): Likewise.
16578 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
16580 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
16582 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
16583 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
16585 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16587 * config/i386/sse.md (sse_movlhps): Change alternative 3
16588 of operand 2 to "m".
16590 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16593 * config/i386/sse.md (iptr): New mode attribute.
16594 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
16595 (<sse>_vm<plusminus_insn><mode>3): Ditto.
16596 (<sse>_vmmul<mode>3): Ditto.
16597 (<sse>_vmdiv<mode>3): Ditto.
16598 (sse_vmrcpv4sf2): Ditto.
16599 (<sse>_vmsqrt<mode>2): Ditto.
16600 (sse_vmrsqrtv4sf2): Ditto.
16601 (<sse>_vm<code><mode>3): Ditto.
16602 (avx_vmcmp<mode>3): Ditto.
16603 (<sse>_vmmaskcmp<mode>3): Ditto.
16604 (<sse>_comi): Ditto.
16605 (<sse>_ucomi): Ditto.
16606 (*xop_vmfrcz_<mode>): Ditto.
16607 (*fmai_fmadd_<mode>): Ditto.
16608 (*fmai_fmsub_<mode>): Ditto.
16609 (*fmai_fnmadd_<mode>): Ditto.
16610 (*fmai_fnmsub_<mode>): Ditto.
16611 (*fma4i_vmfmadd_<mode>): Ditto.
16612 (*fma4i_vmfmsub_<mode>): Ditto.
16613 (*fma4i_vmfnmadd_<mode>): Ditto.
16614 (*fma4i_vmfnmsub_<mode>): Ditto.
16615 (*xop_vmfrcz_<mode>): Ditto.
16616 (sse_cvtps2pi): Ditto.
16617 (sse_cvttps2pi): Ditto.
16618 (sse_cvtss2si): Ditto.
16619 (sse_cvtss2si_2): Ditto.
16620 (sse_cvtss2siq_2): Ditto.
16621 (sse_cvttss2si): Ditto.
16622 (sse_cvttss2siq): Ditto.
16623 (sse_cvtsd2si): Ditto.
16624 (sse_cvtsd2si_2): Ditto.
16625 (sse_cvtsd2siq_2): Ditto.
16626 (sse_cvttsd2si): Ditto.
16627 (sse_cvttsd2siq): Ditto.
16628 (sse_cvtsd2ss): Ditto.
16629 (sse_cvtss2sd): Ditto.
16630 (avx2_pbroadcast<mode>): Ditto.
16631 (avx2_pbroadcast<mode>_1): Ditto.
16632 (*avx_vperm_broadcast_v4sf): Ditto.
16634 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
16635 (sse_movlhps): Ditto.
16636 (sse_storehps): Ditto.
16637 (sse_loadhps): Ditto.
16638 (sse_storelps): Ditto.
16639 (sse_loadlps): Ditto.
16640 (*vec_concatv4sf): Ditto.
16641 (*vec_interleave_highv2df): Ditto.
16642 (*vec_interleave_lowv2df): Ditto.
16643 (*vec_extractv2df_1_sse): Ditto.
16644 (*vec_extractv2df_0_sse): Ditto.
16645 (sse2_storelpd): Ditto.
16646 (sse2_loadlpd): Ditto.
16647 (sse2_movsd): Ditto.
16648 (*vec_concatv4si): Ditto.
16649 (vec_concatv2di): Ditto.
16651 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
16652 for Intel asm dialect.
16653 (mmx_punpcklwd): Ditto.
16654 (mmx_punpckldq): Ditto.
16656 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
16657 for intel assembler dialect.
16659 2013-07-06 Jakub Jelinek <jakub@redhat.com>
16662 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
16663 for BUILT_IN_C{LZ,LRSB}*.
16664 * tree.h (CASE_INT_FN): Add FN##IMAX case.
16665 * tree-vrp.c (extract_range_basic): Handle
16666 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
16667 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
16668 fall thru to code calling set_value*.
16669 * builtins.c (expand_builtin): Remove *IMAX cases.
16670 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
16671 if width is bigger than 2*HWI.
16673 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
16675 PR rtl-optimization/55342
16676 * lra-int.h (lra_subreg_reload_pseudos): New.
16677 * lra.c: Add undoing optional reloads to the block diagram.
16678 (lra_subreg_reload_pseudos): New.
16679 (lra_optional_reload_pseudos): Change comments.
16680 (lra): Init and clear lra_subreg_reload_pseudos. Clear
16681 lra_optional_reload_pseudos after undo transformations.
16682 * lra-assigns.c (pseudo_prefix_title): New.
16683 (lra_setup_reg_renumber): Use it.
16684 (spill_for): Ditto. Check subreg reload pseudos too.
16685 (assign_by_spills): Consider subreg reload pseudos too.
16686 * lra-constraints.c (simplify_operand_subreg): Use
16687 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
16688 (curr_insn_transform): Recognize and do optional reloads.
16689 (undo_optional_reloads): New.
16690 (lra_undo_inheritance): Call undo_optional_reloads.
16692 2013-07-05 Thomas Quinot <quinot@adacore.com>
16694 * tree-complex.c (expand_complex_operations_1): Fix typo.
16696 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
16698 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
16699 (tune_params): New member 'const vec_costs'.
16700 * config/aarch64/aarch64.c (generic_vector_cost): New.
16701 (generic_tunings): New member 'generic_vector_cost'.
16702 (aarch64_builtin_vectorization_cost): New.
16703 (aarch64_add_stmt_cost): New.
16704 (TARGET_VECTORIZE_ADD_STMT_COST): New.
16705 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
16707 2013-07-03 Jakub Jelinek <jakub@redhat.com>
16710 * config/i386/predicates.md (vsib_address_operand): Disallow
16711 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
16713 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
16715 PR middle-end/55030
16716 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
16717 expand_builtin_setjmp_receiver.
16718 (expand_label): Adjust, call expand_builtin_setjmp_receiver
16719 with NULL for the label parameter.
16720 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
16721 the frame-pointer. Adjust comments.
16722 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
16723 only if LABEL is non-NULL.
16725 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
16727 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
16728 (ARM_ABI_AAPCS64): Ditto.
16730 (ARM_DEFAULT_ABI): Ditto.
16732 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
16734 * config/aarch64/aarch64-builtins.c
16735 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
16736 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
16738 * config/aarch64/aarch64-simd.md
16739 (aarch64_ld1<VALL:mode>): New.
16740 (aarch64_st1<VALL:mode>): Likewise.
16741 * config/aarch64/arm_neon.h
16742 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
16744 2013-07-02 Sriraman Tallam <tmsriram@google.com>
16746 * config/i386/i386.c (gate_insert_vzeroupper): Check if
16748 (ix86_option_override_internal):Turn on all -mavx target flags by
16749 default as they are dependent on AVX anyway.
16751 2013-07-02 Cary Coutant <ccoutant@google.com>
16753 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
16754 deterministic hash.
16755 (loc_checksum_ordered): Likewise.
16756 (hash_loc_operands): Remove inline keyword.
16758 2013-07-02 Jakub Jelinek <jakub@redhat.com>
16760 PR tree-optimization/57741
16761 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
16762 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
16763 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
16764 Allow REAL_CST step_exprs if flag_associative_math.
16765 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
16767 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16769 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
16771 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16773 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
16775 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16777 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
16779 (iorsi3_insn): Likewise.
16780 (arm_xorsi3): Likewise.
16782 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
16784 * arm.md (attribute "wtype"): Delete. Move attribute values from here
16786 (attribute "type"): ... here, and prefix with "wmmx_".
16787 (attribute "core_cycles"): Update for attribute changes.
16788 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
16789 (tbcstv4hi): Likewise.
16790 (tbcstv2si): Likewise.
16791 (iwmmxt_iordi3): Likewise.
16792 (iwmmxt_xordi3): Likewise.
16793 (iwmmxt_anddi3): Likewise.
16794 (iwmmxt_nanddi3): Likewise.
16795 (iwmmxt_arm_movdi): Likewise.
16796 (iwmmxt_movsi_insn): Likewise.
16797 (mov<mode>_internal): Likewise.
16798 (and<mode>3_iwmmxt): Likewise.
16799 (ior<mode>3_iwmmxt): Likewise.
16800 (xor<mode>3_iwmmxt): Likewise.
16801 (add<mode>3_iwmmxt): Likewise.
16802 (ssaddv8qi3): Likewise.
16803 (ssaddv4hi3): Likewise.
16804 (ssaddv2si3): Likewise.
16805 (usaddv8qi3): Likewise.
16806 (usaddv4hi3): Likewise.
16807 (usaddv2si3): Likewise.
16808 (sub<mode>3_iwmmxt): Likewise.
16809 (sssubv8qi3): Likewise.
16810 (sssubv4hi3): Likewise.
16811 (sssubv2si3): Likewise.
16812 (ussubv8qi3): Likewise.
16813 (ussubv4hi3): Likewise.
16814 (ussubv2si3): Likewise.
16815 (mulv4hi3_iwmmxt): Likewise.
16816 (smulv4hi3_highpart): Likewise.
16817 (umulv4hi3_highpart): Likewise.
16818 (iwmmxt_wmacs): Likewise.
16819 (iwmmxt_wmacsz): Likewise.
16820 (iwmmxt_wmacu): Likewise.
16821 (iwmmxt_wmacuz): Likewise.
16822 (iwmmxt_clrdi): Likewise.
16823 (iwmmxt_clrv8qi): Likewise.
16824 (iwmmxt_clr4hi): Likewise.
16825 (iwmmxt_clr2si): Likewise.
16826 (iwmmxt_uavgrndv8qi3): Likewise.
16827 (iwmmxt_uavgrndv4hi3): Likewise.
16828 (iwmmxt_uavgv8qi3): Likewise.
16829 (iwmmxt_uavgv4hi3): Likewise.
16830 (iwmmxt_tinsrb): Likewise.
16831 (iwmmxt_tinsrh): Likewise.
16832 (iwmmxt_tinsrw): Likewise.
16833 (iwmmxt_textrmub): Likewise.
16834 (iwmmxt_textrmsb): Likewise.
16835 (iwmmxt_textrmuh): Likewise.
16836 (iwmmxt_textrmsh): Likewise.
16837 (iwmmxt_textrmw): Likewise.
16838 (iwmxxt_wshufh): Likewise.
16839 (eqv8qi3): Likewise.
16840 (eqv4hi3): Likewise.
16841 (eqv2si3): Likewise.
16842 (gtuv8qi3): Likewise.
16843 (gtuv4hi3): Likewise.
16844 (gtuv2si3): Likewise.
16845 (gtv8qi3): Likewise.
16846 (gtv4hi3): Likewise.
16847 (gtv2si3): Likewise.
16848 (smax<mode>3_iwmmxt): Likewise.
16849 (umax<mode>3_iwmmxt): Likewise.
16850 (smin<mode>3_iwmmxt): Likewise.
16851 (umin<mode>3_iwmmxt): Likewise.
16852 (iwmmxt_wpackhss): Likewise.
16853 (iwmmxt_wpackwss): Likewise.
16854 (iwmmxt_wpackdss): Likewise.
16855 (iwmmxt_wpackhus): Likewise.
16856 (iwmmxt_wpackwus): Likewise.
16857 (iwmmxt_wpackdus): Likewise.
16858 (iwmmxt_wunpckihb): Likewise.
16859 (iwmmxt_wunpckihh): Likewise.
16860 (iwmmxt_wunpckihw): Likewise.
16861 (iwmmxt_wunpckilb): Likewise.
16862 (iwmmxt_wunpckilh): Likewise.
16863 (iwmmxt_wunpckilw): Likewise.
16864 (iwmmxt_wunpckehub): Likewise.
16865 (iwmmxt_wunpckehuh): Likewise.
16866 (iwmmxt_wunpckehuw): Likewise.
16867 (iwmmxt_wunpckehsb): Likewise.
16868 (iwmmxt_wunpckehsh): Likewise.
16869 (iwmmxt_wunpckehsw): Likewise.
16870 (iwmmxt_wunpckelub): Likewise.
16871 (iwmmxt_wunpckeluh): Likewise.
16872 (iwmmxt_wunpckeluw): Likewise.
16873 (iwmmxt_wunpckelsb): Likewise.
16874 (iwmmxt_wunpckelsh): Likewise.
16875 (iwmmxt_wunpckelsw): Likewise.
16876 (ror<mode>3): Likewise.
16877 (ashr<mode>3_iwmmxt): Likewise.
16878 (lshr<mode>3_iwmmxt): Likewise.
16879 (ashl<mode>3_iwmmxt): Likewise.
16880 (ror<mode>3_di): Likewise.
16881 (ashr<mode>3_di): Likewise.
16882 (lshr<mode>3_di): Likewise.
16883 (ashl<mode>3_di): Likewise.
16884 (iwmmxt_wmadds): Likewise.
16885 (iwmmxt_wmaddu): Likewise.
16886 (iwmmxt_tmia): Likewise.
16887 (iwmmxt_tmiaph): Likewise.
16888 (iwmmxt_tmiabb): Likewise.
16889 (iwmmxt_tmiatb): Likewise.
16890 (iwmmxt_tmiabt): Likewise.
16891 (iwmmxt_tmiatt): Likewise.
16892 (iwmmxt_tmovmskb): Likewise.
16893 (iwmmxt_tmovmskh): Likewise.
16894 (iwmmxt_tmovmskw): Likewise.
16895 (iwmmxt_waccb): Likewise.
16896 (iwmmxt_wacch): Likewise.
16897 (iwmmxt_waccw): Likewise.
16898 (iwmmxt_waligni): Likewise.
16899 (iwmmxt_walignr): Likewise.
16900 (iwmmxt_walignr0): Likewise.
16901 (iwmmxt_walignr1): Likewise.
16902 (iwmmxt_walignr2): Likewise.
16903 (iwmmxt_walignr3): Likewise.
16904 (iwmmxt_wsadb): Likewise.
16905 (iwmmxt_wsadh): Likewise.
16906 (iwmmxt_wsadbz): Likewise.
16907 (iwmmxt_wsadhz): Likewise.
16908 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16909 (iwmmxt_wabsdiffb): Likewise.
16910 (iwmmxt_wabsdiffh): Likewise.
16911 (iwmmxt_wabsdiffw): Likewise.
16912 (iwmmxt_waddsubhx): Likewise
16913 (iwmmxt_wsubaddhx): Likewise.
16914 (addc<mode>3): Likewise.
16915 (iwmmxt_avg4): Likewise.
16916 (iwmmxt_avg4r): Likewise.
16917 (iwmmxt_wmaddsx): Likewise.
16918 (iwmmxt_wmaddux): Likewise.
16919 (iwmmxt_wmaddsn): Likewise.
16920 (iwmmxt_wmaddun): Likewise.
16921 (iwmmxt_wmulwsm): Likewise.
16922 (iwmmxt_wmulwum): Likewise.
16923 (iwmmxt_wmulsmr): Likewise.
16924 (iwmmxt_wmulumr): Likewise.
16925 (iwmmxt_wmulwsmr): Likewise.
16926 (iwmmxt_wmulwumr): Likewise.
16927 (iwmmxt_wmulwl): Likewise.
16928 (iwmmxt_wqmulm): Likewise.
16929 (iwmmxt_wqmulwm): Likewise.
16930 (iwmmxt_wqmulmr): Likewise.
16931 (iwmmxt_wqmulwmr): Likewise.
16932 (iwmmxt_waddbhusm): Likewise.
16933 (iwmmxt_waddbhusl): Likewise.
16934 (iwmmxt_wqmiabb): Likewise.
16935 (iwmmxt_wqmiabt): Likewise.
16936 (iwmmxt_wqmiatb): Likewise.
16937 (iwmmxt_wqmiatt): Likewise.
16938 (iwmmxt_wqmiabbn): Likewise.
16939 (iwmmxt_wqmiabtn): Likewise.
16940 (iwmmxt_wqmiatbn): Likewise.
16941 (iwmmxt_wqmiattn): Likewise.
16942 (iwmmxt_wmiabb): Likewise.
16943 (iwmmxt_wmiabt): Likewise.
16944 (iwmmxt_wmiatb): Likewise.
16945 (iwmmxt_wmiatt): Likewise.
16946 (iwmmxt_wmiabbn): Likewise.
16947 (iwmmxt_wmiabtn): Likewise.
16948 (iwmmxt_wmiatbn): Likewise.
16949 (iwmmxt_wmiattn): Likewise.
16950 (iwmmxt_wmiawbb): Likewise.
16951 (iwmmxt_wmiawbt): Likewise.
16952 (iwmmxt_wmiawtb): Likewise.
16953 (iwmmxt_wmiawtt): Likewise.
16954 (iwmmxt_wmiawbbn): Likewise.
16955 (iwmmxt_wmiawbtn): Likewise.
16956 (iwmmxt_wmiawtbn): Likewise.
16957 (iwmmxt_wmiawttn): Likewise.
16958 (iwmmxt_wmerge): Likewise.
16959 (iwmmxt_tandc<mode>3): Likewise.
16960 (iwmmxt_torc<mode>3): Likewise.
16961 (iwmmxt_torvsc<mode>3): Likewise.
16962 (iwmmxt_textrc<mode>3): Likewise.
16963 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16964 (wmmxt_pack): Likewise.
16965 (wmmxt_mult_c1): Likewise.
16966 (wmmxt_mult_c2): Likewise.
16967 (wmmxt_alu_c1): Likewise.
16968 (wmmxt_alu_c2): Likewise.
16969 (wmmxt_alu_c3): Likewise.
16970 (wmmxt_transfer_c1): Likewise.
16971 (wmmxt_transfer_c2): Likewise.
16972 (wmmxt_transfer_c3): Likewise.
16973 (marvell_f_iwmmxt_wstr): Likewise.
16974 (marvell_f_iwmmxt_wldr): Likewise.
16976 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16978 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16980 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16983 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16984 * lra-constraints.c (need_for_split_p): Check call used hard regs
16985 living through calls.
16987 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16988 call used regs for call insn.
16990 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16993 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
16994 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
16996 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16998 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
17000 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
17002 * lra-constraints.c (need_for_split_p): Check call used hard regs
17003 living through calls.
17005 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
17008 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
17009 to tie with any other modes. Eliminate Altivec vector mode tests,
17010 since these are a subset of ALTIVEC or VSX vector modes. Simplify
17011 code, to return 0 if testing MODE2 for a condition, if we've
17012 already tested MODE1 for the same condition.
17014 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
17016 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
17019 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
17021 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
17022 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
17024 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
17026 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
17028 (aarch64_symbolic_constant_p): Remove.
17029 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
17030 static. Fix line length and white space.
17031 (aarch64_symbolic_constant_p): Remove.
17032 * config/aarch64/predicates.md (aarch64_valid_symref):
17033 Use aarch64_classify_symbol_expression.
17035 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17037 * config/arm/constraints.md (Ts): New constraint.
17038 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
17040 (compare_scc): Use "Ts" constraint for operand 0.
17041 (ior_scc_scc): Likewise.
17042 (and_scc_scc): Likewise.
17043 (and_scc_scc_nodom): Likewise.
17044 (ior_scc_scc_cmp): Likewise for operand 7.
17045 (and_scc_scc_cmp): Likewise.
17046 * config/arm/thumb2.md (thumb2_movsi_insn):
17047 Add alternatives for 16-bit encodings.
17048 (thumb2_movhi_insn): Likewise.
17049 (thumb2_movsicc_insn): Likewise.
17050 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
17051 (thumb2_negscc): Use "Ts" constraint.
17052 Move mvn instruction outside cond_exec block.
17053 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
17054 for 16-bit encodings.
17056 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17058 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
17060 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
17061 (mulsi3subsi): Likewise.
17062 (mulsidi3adddi): Likewise.
17063 (mulsidi3_v6): Likewise.
17064 (umulsidi3_v6): Likewise.
17065 (umulsidi3adddi_v6): Likewise.
17066 (smulsi3_highpart_v6): Likewise.
17067 (umulsi3_highpart_v6): Likewise.
17068 (mulhisi3tb): Likewise.
17069 (mulhisi3bt): Likewise.
17070 (mulhisi3tt): Likewise.
17071 (maddhisi4): Likewise.
17072 (maddhisi4tb): Likewise.
17073 (maddhisi4tt): Likewise.
17074 (maddhidi4): Likewise.
17075 (maddhidi4tb): Likewise.
17076 (maddhidi4tt): Likewise.
17077 (zeroextractsi_compare0_scratch): Likewise.
17078 (insv_zero): Likewise.
17079 (insv_t2): Likewise.
17080 (anddi_notzesidi_di): Likewise.
17081 (anddi_notsesidi_di): Likewise.
17082 (andsi_notsi_si): Likewise.
17083 (iordi_zesidi_di): Likewise.
17084 (xordi_zesidi_di): Likewise.
17085 (andsi_iorsi3_notsi): Likewise.
17086 (smax_0): Likewise.
17087 (smax_m1): Likewise.
17088 (smin_0): Likewise.
17089 (not_shiftsi): Likewise.
17090 (unaligned_loadsi): Likewise.
17091 (unaligned_loadhis): Likewise.
17092 (unaligned_loadhiu): Likewise.
17093 (unaligned_storesi): Likewise.
17094 (unaligned_storehi): Likewise.
17095 (extv_reg): Likewise.
17096 (extzv_t2): Likewise.
17097 (divsi3): Likewise.
17098 (udivsi3): Likewise.
17099 (arm_zero_extendhisi2addsi): Likewise.
17100 (arm_zero_extendqisi2addsi): Likewise.
17101 (compareqi_eq0): Likewise.
17102 (arm_extendhisi2_v6): Likewise.
17103 (arm_extendqisi2addsi): Likewise.
17104 (arm_movt): Likewise.
17105 (thumb2_ldrd): Likewise.
17106 (thumb2_ldrd_base): Likewise.
17107 (thumb2_ldrd_base_neg): Likewise.
17108 (thumb2_strd): Likewise.
17109 (thumb2_strd_base): Likewise.
17110 (thumb2_strd_base_neg): Likewise.
17111 (arm_negsi2): Add alternative for 16-bit encoding.
17112 (arm_one_cmplsi2): Likewise.
17114 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17116 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
17117 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
17118 (movdfcc): Likewise.
17119 * config/arm/vfp.md (*thumb2_movsf_vfp):
17120 Disable predication for arm_restrict_it.
17121 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
17122 (*thumb2_movdfcc_vfp): Likewise.
17123 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
17124 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
17125 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
17126 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
17127 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
17128 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
17129 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
17130 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
17131 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
17132 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
17133 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
17134 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
17135 Disable predication for arm_restrict_it.
17137 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
17139 * config/i386/bmiintrin.h (_bextr_u32): New.
17140 (_bextr_u64): Ditto.
17142 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
17144 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
17145 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
17146 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
17147 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
17148 * config/mips/n32-elf.h: ...this new file.
17150 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
17153 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
17154 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
17156 2013-06-27 Catherine Moore <clm@codesourcery.com>
17158 * config/mips/mips-tables.opt: Regenerate.
17159 * config/mips/mips-cpus.def: Add m14ke and m14kec.
17160 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
17161 * doc/invoke.texi: Add -m14kc.
17163 2013-06-27 Jakub Jelinek <jakub@redhat.com>
17166 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
17167 constraints of operand 1 and 2.
17170 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
17171 to match RTL canonicalization. Swap predicates and
17172 constraints of operand 1 and 2.
17174 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
17176 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
17177 Process OP_INOUT regs for splitting too.
17179 2013-06-27 Jakub Jelinek <jakub@redhat.com>
17181 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
17182 decl before the loop, initialize to NULL.
17183 (vectorizable_load): Initialize ptr_incr to NULL.
17185 2013-06-27 Martin Jambor <mjambor@suse.cz>
17188 * ipa-ref.h (ipa_maybe_record_reference): Declare.
17189 * ipa-ref.c (ipa_maybe_record_reference): New function.
17190 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
17191 * ipa-cp.c (create_specialized_node): Record potential references from
17193 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
17195 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
17197 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
17198 parameter 'mode' of type 'enum machine_mode mode'; change to pass
17199 'mode' to force_reg.
17200 (aarch64_add_offset): Update calls to aarch64_force_temporary.
17201 (aarch64_expand_mov_immediate): Likewise.
17203 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
17205 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
17206 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
17208 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17210 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
17211 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
17212 (struct machine_function): Add tbegin_p.
17213 (s390_canonicalize_comparison): Fold CC mode compares to
17214 conditional jump if possible.
17215 (s390_emit_jump): Return the emitted jump.
17216 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
17217 Handle CCRAWmode compares.
17218 (s390_option_override): Default to -mhtm if available.
17219 (s390_reg_clobbered_rtx): Handle floating point regs as well.
17220 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
17221 FPRs instead of df_regs_ever_live_p.
17222 (s390_optimize_nonescaping_tx): New function.
17223 (s390_init_frame_layout): Extend clobbered_regs array to cover
17225 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
17226 (s390_expand_tbegin): New function.
17227 (enum s390_builtin): New enum definition.
17228 (code_for_builtin): New array definition.
17229 (s390_init_builtins): New function.
17230 (s390_expand_builtin): New function.
17231 (TARGET_INIT_BUILTINS): Define.
17232 (TARGET_EXPAND_BUILTIN): Define.
17233 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
17234 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
17235 (s390_alc_comparison): Likewise.
17236 * config/s390/s390-modes.def: Add CCRAWmode.
17237 * config/s390/s390.h (processor_flags): Add PF_TX.
17238 (TARGET_CPU_HTM): Define macro.
17239 (TARGET_HTM): Define macro.
17240 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
17241 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
17242 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
17243 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
17244 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
17245 (TBEGIN_MASK, TBEGINC_MASK): New constants.
17246 ("*cc_to_int"): Move up.
17247 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
17248 constants other than 0.
17249 ("*ccraw_to_int"): New insn and splitter definition.
17250 ("tbegin", "tbegin_nofloat", "tbegin_retry")
17251 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
17252 ("tx_assist"): New expander.
17253 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
17254 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
17255 * config/s390/s390.opt: Add -mhtm option.
17256 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
17257 * config/s390/htmxlintrin.h: New file.
17258 * config/s390/htmintrin.h: New file.
17259 * config/s390/s390intrin.h: New file.
17260 * doc/extend.texi: Document htm builtins.
17261 * config.gcc: Add the new header files to extra_headers.
17263 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
17265 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
17266 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
17268 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
17269 Pat Haugen <pthaugen@us.ibm.com>
17270 Peter Bergner <bergner@vnet.ibm.com>
17272 * config/rs6000/power8.md: New.
17273 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
17274 setting for power8 entry.
17275 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
17276 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
17277 test for Power4/Power5 only.
17278 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
17280 (force_new_group): Adjust comment.
17281 * config/rs6000/rs6000.md: Include power8.md.
17283 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
17285 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
17286 * config/arm/arm-protos.h (arm_max_conditional_execute): New
17288 (tune_params): Update comment.
17289 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
17290 (arm_max_conditional_execute): New function.
17291 (thumb2_final_prescan_insn): Use max_insn_skipped and
17292 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
17294 2013-06-25 Jakub Jelinek <jakub@redhat.com>
17296 PR tree-optimization/57705
17297 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
17298 SSA_NAME step, provided that it is not defined inside the loop.
17299 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
17300 (get_initial_def_for_induction): Handle SSA_NAME IV step.
17302 2013-06-25 Martin Jambor <mjambor@suse.cz>
17304 PR middle-end/57670
17305 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
17306 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
17308 (ipa_note_param_call): Initialize member_ptr flag.
17309 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
17310 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
17311 (ipa_write_indirect_edge_info): Stream member_ptr flag.
17312 (ipa_read_indirect_edge_info): Likewise.
17314 2013-06-25 Richard Biener <rguenther@suse.de>
17316 PR middle-end/56977
17317 * passes.c (init_optimization_passes): Move pass_fold_builtins
17318 and pass_dce earlier with -Og.
17320 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
17322 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
17323 <BIT_FIELD_REF>: Remove trailing TAB.
17324 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
17327 2013-06-24 Martin Jambor <mjambor@suse.cz>
17329 PR tree-optimization/57358
17330 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
17331 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
17332 (ipa_analyze_params_uses): Generate pessimistic info when true.
17334 2013-06-24 Martin Jambor <mjambor@suse.cz>
17336 PR tree-optimization/57539
17337 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
17338 global.inlined_to of the new node to it. All callers changed.
17339 * ipa-inline-transform.c (clone_inlined_nodes): New variable
17340 inlining_into, pass it to cgraph_clone_node.
17341 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
17342 ipa_free_edge_args_substructures.
17343 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
17344 rdesc linked list. Do not assert rdesc edges have inlined caller.
17345 Assert we have found an rdesc in the rdesc list.
17347 2013-06-24 Richard Biener <rguenther@suse.de>
17349 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
17350 (pointer_set_lookup): Declare.
17351 (class pointer_map): New template class implementing a
17352 generic pointer to T map.
17353 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
17354 pointer_map<T>::contains, pointer_map<T>::insert,
17355 pointer_map<T>::traverse): New functions.
17356 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
17357 (pointer_set_lookup): New function.
17358 (pointer_set_contains): Use pointer_set_lookup.
17359 (pointer_set_insert): Likewise.
17360 (insert_aux): Remove.
17361 (struct pointer_map_t): Embed a pointer_set_t.
17362 (pointer_map_create): Adjust.
17363 (pointer_map_destroy): Likewise.
17364 (pointer_map_contains): Likewise.
17365 (pointer_map_insert): Likewise.
17366 (pointer_map_traverse): Likewise.
17367 * tree-streamer.h (struct streamer_tree_cache_d): Use a
17368 pointer_map<unsigned> instead of a pointer_map_t.
17369 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
17370 (streamer_tree_cache_lookup): Likewise.
17371 (streamer_tree_cache_create): Likewise.
17372 (streamer_tree_cache_delete): Likewise.
17373 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
17374 pointer_map<unsigned> instead of a pointer_map_t.
17375 (lto_init_tree_ref_encoder): Adjust.
17376 (lto_destroy_tree_ref_encoder): Likewise.
17377 * lto-section-out.c (lto_output_decl_index): Likewise.
17378 (lto_record_function_out_decl_state): Likewise.
17379 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
17381 2013-06-24 Richard Biener <rguenther@suse.de>
17383 PR tree-optimization/57488
17384 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
17386 2013-06-24 Alan Modra <amodra@gmail.com>
17388 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
17389 (gen_easy_altivec_constant): Likewise.
17390 * config/rs6000/predicates.md (easy_vector_constant_add_self,
17391 easy_vector_constant_msb): Likewise.
17393 2013-06-23 Jakub Jelinek <jakub@redhat.com>
17396 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
17397 add missing return true.
17399 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
17402 * config/sh/predicates.md (general_extend_operand): Invoke
17403 general_movsrc_operand for memory operands.
17404 (general_movsrc_operand): Allow reg+reg addressing, do not use
17405 general_operand for memory operands.
17407 2013-06-23 Sriraman Tallam <tmsriram@google.com>
17409 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
17410 when current target options does not apply.
17411 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
17412 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
17413 * config/i386/bmiintrin.h: Pass appropriate target
17414 attributes to header.
17415 * config/i386/mmintrin.h: Ditto.
17416 * config/i386/nmmintrin.h: Ditto.
17417 * config/i386/avx2intrin.h: Ditto.
17418 * config/i386/fxsrintrin.h: Ditto.
17419 * config/i386/tbmintrin.h: Ditto.
17420 * config/i386/xsaveintrin.h: Ditto.
17421 * config/i386/f16cintrin.h: Ditto.
17422 * config/i386/xtestintrin.h: Ditto.
17423 * config/i386/xsaveoptintrin.h: Ditto.
17424 * config/i386/bmi2intrin.h: Ditto.
17425 * config/i386/lzcntintrin.h: Ditto.
17426 * config/i386/smmintrin.h: Ditto.
17427 * config/i386/wmmintrin.h: Ditto.
17428 * config/i386/x86intrin.h: Remove all header include guards.
17429 * config/i386/prfchwintrin.h: Ditto.
17430 * config/i386/pmmintrin.h: Ditto.
17431 * config/i386/tmmintrin.h: Ditto.
17432 * config/i386/xmmintrin.h: Ditto.
17433 * config/i386/popcntintrin.h: Ditto.
17434 * config/i386/rdseedintrin.h: Ditto.
17435 * config/i386/ammintrin.h: Ditto.
17436 * config/i386/emmintrin.h: Ditto.
17437 * config/i386/immintrin.h: Remove all header include guards.
17438 * config/i386/fma4intrin.h: Ditto.
17439 * config/i386/lwpintrin.h: Ditto.
17440 * config/i386/xopintrin.h: Ditto.
17441 * config/i386/ia32intrin.h: Ditto.
17442 * config/i386/avxintrin.h: Ditto.
17443 * config/i386/rtmintrin.h: Ditto.
17444 * config/i386/fmaintrin.h: Ditto.
17445 * config/i386/mm3dnow.h: Ditto.
17447 2013-06-22 Sriraman Tallam <tmsriram@google.com>
17449 * common/config/i386/i386-common.c: Handle LZCNT.
17451 2013-06-22 Andi Kleen <ak@linux.intel.com>
17453 * doc/extend.texi: Use __atomic_store_n instead of
17454 __atomic_store in HLE example.
17456 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
17458 * config/sh/sh.c: Remove <cstdlib> workaround.
17460 2013-06-21 Andi Kleen <ak@linux.intel.com>
17462 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
17464 2013-06-21 Andi Kleen <ak@linux.intel.com>
17466 * doc/extend.texi: Document that __atomic_clear and
17467 __atomic_test_and_set should only be used with bool.
17469 2013-06-20 Jan Hubicka <jh@suse.cz>
17471 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
17472 types_same_for_odr.
17473 * tree.c (decls_same_for_odr): New function.
17474 (same_for_edr): New function.
17475 (types_same_for_odr): New function.
17476 (get_binfo_at_offset): Use it.
17477 * tree.h (types_same_for_odr): Declare.
17479 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
17480 Jason Merrill <jason@redhat.com>
17482 * system.h: Include <cstdlib> as well as <stdlib.h>.
17484 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
17487 * config/i386/i386.c (construct_container): Report error if
17488 long double is used with disabled x87 float returns.
17490 2013-06-20 Jan Hubicka <jh@suse.cz>
17492 * lto-cgraph.c (input_symtab): Do not set cgraph state.
17494 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
17496 PR rtl-optimization/57425
17497 PR rtl-optimization/57569
17498 * alias.c (write_dependence_p): Remove parameters mem_mode and
17499 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
17500 Changed all callers.
17501 (canon_anti_dependence): Get comments and semantics in sync.
17502 Add parameter mem_canonicalized. Changed all callers.
17503 * rtl.h (canon_anti_dependence): Update prototype.
17505 2013-06-20 Richard Biener <rguenther@suse.de>
17507 * data-streamer-in.c (streamer_read_uhwi): Optimize single
17508 byte case, inline streamer_read_uchar and defer section
17511 2013-06-20 Richard Biener <rguenther@suse.de>
17513 PR tree-optimization/57584
17514 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
17515 SSA names into the expanded expression that take part in
17516 abnormal coalescing.
17518 2013-06-19 Sharad Singhai <singhai@google.com>
17520 * gcov.c (print_usage): Handle new option.
17521 (process_args): Ditto.
17522 (get_gcov_intermediate_filename): New function.
17523 (output_intermediate_file): New function.
17524 (output_gcov_file): New function
17525 (generate_results): Handle new option.
17526 (release_function): Relase demangled name.
17527 (read_graph_file): Handle demangled name.
17528 (output_lines): Ditto.
17529 * doc/gcov.texi: Document gcov intermediate format.
17531 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
17534 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
17535 (lra_emit_add): Use the functions. Add comment about Y as an
17538 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
17541 * collect2.c (collect_atexit): New.
17542 (collect_exit): Delete.
17543 (main): Register collect_atexit with atexit.
17544 (collect_wait): Change collect_exit to exit.
17546 * collect2.h (collect_exit): Delete.
17547 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
17549 2013-06-19 Wei Mi <wmi@google.com>
17551 PR rtl-optimization/57518
17552 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
17553 if regno is used in paradoxical subreg.
17554 (update_equiv_regs): Check pdx_subregs[regno] before
17555 set a reg to be equivalent with a mem.
17557 2013-06-19 Matthias Klose <doko@ubuntu.com>
17560 * file-find.h (find_a_file): Add a mode parameter.
17561 * file-find.c (find_a_file): Likewise.
17562 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
17563 with X_OK for the executables.
17564 * collect2.c (main): Call find_a_file with X_OK.
17566 2013-06-19 Steve Ellcey <sellcey@mips.com>
17569 * config/mips/mips.md (casesi_internal_mips16_<mode>):
17570 Use NEXT_INSN instead of next_real_insn.
17572 2013-06-19 Jan Hubicka <jh@suse.cz>
17574 * cgraph.h (const_value_known_p): Replace by ...
17575 (ctor_for_folding): .. this one.
17576 * cgraphunit.c (process_function_and_variable_attributes): Use it.
17577 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
17578 * expr.c (expand_expr_real_1): Likewise.
17579 (string_constant): Likewise.
17580 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
17581 * ipa.c (process_references): Likewise.
17582 (symtab_remove_unreachable_nodes): Likewise.
17583 * ipa-inline-analysis.c (param_change_prob): Likewise.
17584 * gimple-fold.c (canonicalize_constructor_val): Likewise.
17585 (get_base_constructor): Likwise.
17586 * varpool.c (varpool_remove_node): Likewise.
17587 (varpool_remove_initializer): LIkewise.
17588 (dump_varpool_node): LIkwise.
17589 (const_value_known_p): Rewrite to ...
17590 (ctor_for_folding): ... this one.
17592 2013-06-19 Jakub Jelinek <jakub@redhat.com>
17595 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
17596 PERSONALITY in $PATH derived prefixes.
17598 2013-06-19 Jeff Law <law@redhat.com>
17600 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
17603 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
17604 (simplify_bitwise_binary): Use it to simpify certain binary ops on
17607 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
17609 * config/arm/vfp.md: Move VFP instruction classification documentation
17611 * config/arm/arm.md: ... here. Update instruction classification
17614 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
17616 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
17617 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
17618 pattern. Use more efficient sequences on ARMv5 and Thumb2.
17620 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
17623 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17624 with NEXT_INSN. Use tablejump_p to check for jump table data
17627 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
17630 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
17631 that now in C++ the value is correct per the C++ standards.
17633 2013-06-19 Richard Biener <rguenther@suse.de>
17635 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
17636 for global context.
17638 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17641 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17644 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17645 with next_active_insn.
17647 2013-06-18 Sriraman Tallam <tmsriram@google.com>
17649 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
17650 functions are inlined during failures to flag an error.
17651 * tree-inline.c (expand_call_inline): Allow the error to be flagged
17652 in early inline pass.
17654 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
17656 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
17659 2013-06-18 Julian Brown <julian@codesourcery.com>
17661 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
17662 Permit virtual register pre-reload if !strict.
17663 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
17665 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
17667 * config/arm/neon.md (movmisalign<mode>): Use
17668 neon_perm_struct_or_reg_operand instead of
17669 neon_struct_or_register_operand.
17670 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
17671 neon_permissive_struct_operand instead of neon_struct_operand.
17672 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
17673 neon_vector_mem_operand.
17674 * config/arm/predicates.md (neon_struct_operand): Adjust call to
17675 neon_vector_mem_operand.
17676 (neon_permissive_struct_operand): New.
17677 (neon_struct_or_register_operand): Rename to...
17678 (neon_perm_struct_or_reg_operand): This. Adjust call to
17679 neon_vector_mem_operand.
17681 2013-06-18 Richard Biener <rguenther@suse.de>
17683 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
17684 * lto-streamer.h: Include pointer-set.h.
17685 (struct lto_decl_slot): Remove.
17686 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
17687 Remove next_index entry.
17688 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17689 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17690 (lto_init_tree_ref_encoder): Adjust.
17691 (lto_destroy_tree_ref_encoder): Likewise.
17692 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17693 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17694 (lto_output_decl_index): Adjust.
17695 (lto_new_out_decl_state): Likewise.
17696 (lto_record_function_out_decl_state): Likewise.
17697 * lto-streamer-out.c (copy_function): Likewise.
17699 2013-06-18 Richard Biener <rguenther@suse.de>
17701 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
17702 * cgraphunit.c: Include cfgloop.h.
17703 (init_lowered_empty_function): Initialize the loop tree.
17704 (assemble_thunk): Insert new BBs into loops.
17706 2013-06-18 Richard Biener <rguenther@suse.de>
17708 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
17709 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
17710 the map from cache entry to cache index optional.
17711 (streamer_tree_cache_replace_tree): Adjust accordingly.
17712 (streamer_tree_cache_append): Likewise.
17713 (streamer_tree_cache_delete): Likewise.
17714 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
17715 streamer cache map from cache entry to cache index.
17716 * lto-streamer-out.c (create_output_block): Adjust.
17718 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
17720 * config/arm/arm.md (attribute "insn"): Move multiplication and
17721 division attributes to...
17722 (attribute "type"): ... here. Remove mult.
17723 (attribute "mul32"): New attribute.
17724 (attribute "mul64"): Add umaal.
17725 (*arm_mulsi3): Update attributes.
17726 (*arm_mulsi3_v6): Likewise.
17727 (*thumb_mulsi3): Likewise.
17728 (*thumb_mulsi3_v6): Likewise.
17729 (*mulsi3_compare0): Likewise.
17730 (*mulsi3_compare0_v6): Likewise.
17731 (*mulsi_compare0_scratch): Likewise.
17732 (*mulsi_compare0_scratch_v6): Likewise.
17733 (*mulsi3addsi): Likewise.
17734 (*mulsi3addsi_v6): Likewise.
17735 (*mulsi3addsi_compare0): Likewise.
17736 (*mulsi3addsi_compare0_v6): Likewise.
17737 (*mulsi3addsi_compare0_scratch): Likewise.
17738 (*mulsi3addsi_compare0_scratch_v6): Likewise.
17739 (*mulsi3subsi): Likewise.
17740 (*mulsidi3adddi): Likewise.
17741 (*mulsi3addsi_v6): Likewise.
17742 (*mulsidi3adddi_v6): Likewise.
17743 (*mulsidi3_nov6): Likewise.
17744 (*mulsidi3_v6): Likewise.
17745 (*umulsidi3_nov6): Likewise.
17746 (*umulsidi3_v6): Likewise.
17747 (*umulsidi3adddi): Likewise.
17748 (*umulsidi3adddi_v6): Likewise.
17749 (*smulsi3_highpart_nov6): Likewise.
17750 (*smulsi3_highpart_v6): Likewise.
17751 (*umulsi3_highpart_nov6): Likewise.
17752 (*umulsi3_highpart_v6): Likewise.
17753 (mulhisi3): Likewise.
17754 (*mulhisi3tb): Likewise.
17755 (*mulhisi3bt): Likewise.
17756 (*mulhisi3tt): Likewise.
17757 (maddhisi4): Likewise.
17758 (*maddhisi4tb): Likewise.
17759 (*maddhisi4tt): Likewise.
17760 (maddhidi4): Likewise.
17761 (*maddhidi4tb): Likewise.
17762 (*maddhidi4tt): Likewise.
17763 (divsi3): Likewise.
17764 (udivsi3): Likewise.
17765 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
17766 (thumb2_mulsi_short_compare0): Likewise.
17767 (thumb2_mulsi_short_compare0_scratch): Likewise.
17768 * config/arm/arm1020e.md (1020mult1): Update attribute change.
17769 (1020mult2): Likewise.
17770 (1020mult3): Likewise.
17771 (1020mult4): Likewise.
17772 (1020mult5): Likewise.
17773 (1020mult6): Likewise.
17774 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
17776 (cortex_a15_mult64): Likewise.
17777 (cortex_a15_sdiv): Likewise.
17778 (cortex_a15_udiv): Likewise.
17779 * config/arm/arm1026ejs.md (mult1): Update attribute change.
17785 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
17786 (pj4_ir_div): Likewise.
17787 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
17788 (11_mult2): Likewise.
17789 (11_mult3): Likewise.
17790 (11_mult4): Likewise.
17791 (11_mult5): Likewise.
17792 (11_mult6): Likewise.
17793 (11_mult7): Likewise.
17794 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
17795 (cortex_a8_mla): Likewise.
17796 (cortex_a8_mull): Likewise.
17797 (cortex_a8_smulwy): Likewise.
17798 (cortex_a8_smlald): Likewise.
17799 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
17800 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
17801 (cortex_r4_mul_3): Likewise.
17802 (cortex_r4_mla_4): Likewise.
17803 (cortex_r4_mla_3): Likewise.
17804 (cortex_r4_smlald): Likewise.
17805 (cortex_r4_mull): Likewise.
17806 (cortex_r4_sdiv): Likewise.
17807 (cortex_r4_udiv): Likewise.
17808 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
17809 (cortex_a7_idiv): Likewise.
17810 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
17811 (9_mult2): Likewise.
17812 (9_mult3): Likewise.
17813 (9_mult4): Likewise.
17814 (9_mult5): Likewise.
17815 (9_mult6): Likewise.
17816 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
17817 (cortex_a53_sdiv): Likewise.
17818 (cortex_a53_udiv): Likewise.
17819 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
17820 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
17821 (mp626_mult2): Likewise.
17822 (mp626_mult3): Likewise.
17823 (mp626_mult4): Likewise.
17824 * config/arm/fa526.md (526_mult1): Update attribute change.
17825 (526_mult2): Likewise.
17826 * config/arm/arm-generic.md (mult): Update attribute change.
17827 (mult_ldsched_strongarm): Likewise.
17828 (mult_ldsched): Likewise.
17829 (multi_cycle): Likewise.
17830 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
17831 * config/arm/fa606te.md (606te_mult1): Update attribute change.
17832 (606te_mult2): Likewise.
17833 (606te_mult3): Likewise.
17834 (606te_mult4): Likewise.
17835 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
17836 (cortex_a9_mac16): Likewise.
17837 (cortex_a9_multiply): Likewise.
17838 (cortex_a9_mac): Likewise.
17839 (cortex_a9_multiply_long): Likewise.
17840 * config/arm/fa626te.md (626te_mult1): Update attribute change.
17841 (626te_mult2): Likewise.
17842 (626te_mult3): Likewise.
17843 (626te_mult4): Likewise.
17845 2013-06-18 Richard Biener <rguenther@suse.de>
17848 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17850 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17853 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17854 with next_active_insn.
17856 2013-06-18 Alan Modra <amodra@gmail.com>
17858 * config/rs6000/rs6000.h (enum data_align): New.
17859 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17860 (DATA_ABI_ALIGNMENT): Define.
17861 (CONSTANT_ALIGNMENT): Correct comment.
17862 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17863 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17865 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17867 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17868 ATTRIBUTE_UNUSED marking.
17870 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17872 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17873 alternative and update.
17874 (aarch64_dup_lanedi): Delete.
17875 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17876 * config/aarch64/aarch64-simd-builtins.def: Update.
17878 2013-06-17 Richard Biener <rguenther@suse.de>
17880 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17881 (lto_input_scc): Declare.
17882 (lto_input_tree_1): Likewise.
17883 (struct lto_stats_d): Add num_tree_bodies_output and
17884 num_pickle_refs_output.
17885 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17886 (lto_read_tree_1): Split out from ...
17887 (lto_read_tree): ... this.
17888 (lto_input_scc): New function.
17889 (lto_input_tree_1): Split out from ...
17890 (lto_input_tree): ... this. Handle LTO_tree_scc.
17891 (lto_data_in_create): Create the streamer cache without hashes.
17892 * lto-streamer-out.c (create_output_block): Create the streamer
17893 cache with hashes when not doing WPA.
17894 (lto_write_tree_1): Split out from ...
17895 (lto_write_tree): ... this.
17896 (get_symbol_initial_value): New function.
17897 (lto_output_tree_1): Split out from ...
17898 (lto_output_tree): ... this. Write trees as series of SCCs
17899 using a DFS walk via DFS_write_tree.
17900 (struct sccs, struct scc_entry): New types.
17901 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17902 (DFS_write_tree_body): New function.
17903 (DFS_write_tree): Likewise.
17904 (hash_tree): Likewise.
17905 (scc_entry_compare): Likewise.
17906 (hash_scc): Likewise.
17907 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17908 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17909 TREE_CHAIN as regular reference.
17910 (streamer_read_integer_cst): Remove.
17911 (streamer_get_pickled_tree): Adjust.
17912 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17913 of DECL_EXTERNALs in BLOCK_VARS for now.
17914 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17916 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17917 Add hash value argument and record that if hashes are recorded
17919 (streamer_tree_cache_insert_1): Adjust.
17920 (streamer_tree_cache_insert): Likewise.
17921 (streamer_tree_cache_insert_at): Rename to ...
17922 (streamer_tree_cache_replace_tree): ... this and adjust.
17923 (streamer_tree_cache_append): Adjust.
17924 (record_common_node): Likewise.
17925 (streamer_tree_cache_create): Add argument whether to
17926 record hash values together with trees.
17927 (streamer_tree_cache_delete): Adjust.
17928 * tree-streamer.h (struct streamer_tree_cache_d): Add
17930 (streamer_read_integer_cst): Remove.
17931 (streamer_tree_cache_insert): Adjust.
17932 (streamer_tree_cache_append): Likewise.
17933 (streamer_tree_cache_insert_at): Rename to ...
17934 (streamer_tree_cache_replace_tree): ... this and adjust.
17935 (streamer_tree_cache_create): Add argument whether to record hashes.
17936 (streamer_tree_cache_get): Rename to ...
17937 (streamer_tree_cache_get_tree): ... this.
17938 (streamer_tree_cache_get_hash): New function.
17939 * tree.c (cache_integer_cst): New function.
17940 * tree.h (cache_integer_cst): Declare.
17941 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17942 * lto-symtab.c (lto_varpool_replace_node): Only release
17943 DECL_INITIAL of non-prevailing decls.
17944 * varpool.c (varpool_remove_initializer): Do not release
17945 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17947 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17949 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17950 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17951 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17952 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17953 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17954 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17955 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17956 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17957 instead of TARGET_64BIT.
17958 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17959 Require ISA_HAS_<D>DIV.
17961 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17963 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17964 (mips*-*-linux*): Move default with_llsc setting to where other
17966 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17968 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17969 Likewise. Remove default with_tune setting. Move default float
17970 setting to its own block. Handle with_llsc in the same block as above.
17972 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17974 PR rtl-optimization/57425
17975 PR rtl-optimization/57569
17976 * alias.c (write_dependence_p): Add new parameters mem_mode,
17977 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17978 Changed all callers.
17979 (canon_anti_dependence): New function.
17980 * cse.c (check_dependence): Use canon_anti_dependence.
17981 * cselib.c (cselib_invalidate_mem): Likewise.
17982 * rtl.h (canon_anti_dependence): Declare.
17984 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17986 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17987 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17988 ".set mips3" for 64-bit targets.
17990 2013-06-15 Dehao Chen <dehao@google.com>
17992 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17993 * gimple-low.c (gimple_check_call_matching_types): Likewise.
17994 (gimple_check_call_args): Likewise.
17995 * value-prof.c (check_ic_target): Likewise.
17996 * ipa-inline.c (early_inliner): Likewise.
17997 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
17998 * cgraph.c (cgraph_create_edge_1): Likewise.
17999 (cgraph_make_edge_direct): Likewise.
18001 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
18004 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
18005 rs6000_output_move_128bit to handle emitting quad memory
18006 operations. Set attribute length to 8 bytes.
18008 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
18010 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
18012 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
18013 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
18014 (aarch64_<su>mlsl<mode>): Likewise.
18016 2013-06-14 Mike Stump <mikestump@comcast.net>
18018 * Makefile.in (TARGET_H): Add insn-codes.h.
18020 2013-06-14 Alan Modra <amodra@gmail.com>
18022 PR middle-end/57134
18023 PR middle-end/57586
18024 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
18025 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
18026 bitfield expansion when EXPAND_MEMORY.
18027 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
18029 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
18031 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
18032 test for clearing quad memory on 32-bit later.
18034 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
18036 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
18037 (fold_negate_expr): Likewise.
18038 (fold_real_zero_addition_p): Handle vectors.
18039 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
18041 2013-06-14 Alan Modra <amodra@gmail.com>
18043 * varasm.c (force_const_mem): Revert 2013-06-07 change.
18045 2013-06-13 Jan Hubicka <jh@suse.cz>
18047 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
18048 Local comdats are not externally visible.
18049 * symtab.c (dump_symtab_base): Dump externally visible.
18050 (verify_symtab_base): Verify back links in the symtab hash.
18052 2013-06-13 Bin Cheng <bin.cheng@arm.com>
18054 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
18055 CONVERT_EXPR as equal nodes.
18057 2013-06-13 Bin Cheng <bin.cheng@arm.com>
18059 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
18061 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
18063 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
18064 Generalize to complex and vector.
18065 * tree.c (build_all_ones_cst): New function.
18066 * tree.h (build_all_ones_cst): Declare it.
18068 2013-06-13 Alan Modra <amodra@gmail.com>
18070 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
18071 * config/rs6000/rs6000.md (signbittf2): New insn.
18072 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
18073 (abstf2_internal, cmptf_internal2): Likewise.
18074 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
18076 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
18077 Pat Haugen <pthaugen@us.ibm.com>
18078 Peter Bergner <bergner@vnet.ibm.com>
18080 * config/rs6000/rs6000.c (emit_load_locked): Add support for
18081 power8 byte, half-word, and quad-word atomic instructions.
18082 (emit_store_conditional): Likewise.
18083 (rs6000_expand_atomic_compare_and_swap): Likewise.
18084 (rs6000_expand_atomic_op): Likewise.
18086 * config/rs6000/sync.md (larx): Add new modes for power8.
18088 (AINT): New mode iterator to include TImode as well as normal
18089 integer modes on power8.
18090 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
18091 that VSX registers are not considered. Use AINT mode iterator
18092 instead of INT1 to allow inclusion of quad word atomic operations
18094 (load_locked<mode>): Likewise.
18095 (store_conditional<mode>): Likewise.
18096 (atomic_compare_and_swap<mode>): Likewise.
18097 (atomic_exchange<mode>): Likewise.
18098 (atomic_nand<mode>): Likewise.
18099 (atomic_fetch_<fetchop_name><mode>): Likewise.
18100 (atomic_nand_fetch<mode>): Likewise.
18101 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
18103 (ATOMIC): On power8, add QImode, HImode modes.
18104 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
18105 modes that promote to SImode.
18106 (load_lockedti): Convert TImode arguments to PTImode, so that we
18107 get a guaranteed even/odd register pair.
18108 (load_lockedpti): Likewise.
18109 (store_conditionalti): Likewise.
18110 (store_conditionalpti): Likewise.
18112 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
18113 atomic load/store instructions.
18116 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
18118 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
18120 (insn_count): New attribute, with most cases extracted from...
18121 (length): ...here. Redefine most cases in terms of insn_count.
18122 (single_insn): Delete.
18123 (can_delay): Use insn_count to check for single instructions.
18124 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
18125 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
18126 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
18127 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
18128 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
18129 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
18130 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
18131 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
18132 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
18133 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
18134 rather than "length".
18135 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
18136 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
18137 Use "insn_count" rather than "length".
18138 * config/mips/mips-dsp.md
18139 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
18140 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
18143 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
18145 PR tree-optimization/57361
18146 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
18148 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
18150 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
18152 (aarch64_simd_combine<mode>): New instruction expansion.
18153 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
18154 function prototype.
18155 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
18156 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
18158 2013-06-12 Jan Hubicka <jh@suse.cz>
18160 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
18161 decl has when in streaming stage.
18162 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
18163 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
18165 2013-06-12 Roland Stigge <stigge@antcom.de>
18168 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
18170 2013-06-12 Jakub Jelinek <jakub@redhat.com>
18172 PR tree-optimization/57537
18173 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
18174 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
18176 2013-06-12 Richard Biener <rguenther@suse.de>
18178 * data-streamer.h (streamer_write_char_stream): CSE
18179 obs->current_pointer.
18180 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
18181 streamer_write_char_stream manually and optimize the resulting loop.
18182 (streamer_write_hwi_stream): Likewise.
18184 2013-06-12 Jan Hubicka <jh@suse.cz>
18186 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
18187 * cgraph.h (varpool_create_empty_node): Declare.
18188 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
18190 * symtab.c (symtab_unregister_node): Be lax about missin entries
18192 (symtab_get_node): Update comment.
18193 * varpool.c (varpool_create_empty_node): Break out from ...
18194 (varpool_node_for_decl): ... here.
18195 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
18197 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
18199 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
18200 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
18201 part. Use straight-line flow at the end.
18202 <COMPONENT_REF>: Remove superfluous else.
18203 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
18205 2013-06-12 Jakub Jelinek <jakub@redhat.com>
18208 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
18209 target hook even for !TREE_PUBLIC decls. If no resolution info
18210 is available, return false for common and external decls.
18212 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18214 * config/rl78/constraints.md (U): New constraint.
18215 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
18218 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18221 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
18222 to allow returning address to AT_PLATFORM name.
18224 2013-06-11 Jan Hubicka <jh@suse.cz>
18226 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
18227 * cgraph.h (symtab_node_base): Add weakref flag.
18228 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
18229 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
18230 (output_weakrefs): Use weakref flag.
18231 * fold-const.c (simple_operand_p): Handle WEAK.
18232 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
18233 * ipa.c (varpool_externally_visible_p): Drop weakref.
18234 (function_and_variable_visibility): Update comment; fix weakref
18235 sanity checks; do not clear DECL_WEAK on them.
18236 * lto-cgraph.c (lto_output_node): update.
18237 (lto_output_varpool_node): Update.
18238 (input_overwrite_node): Update.
18239 (input_node): Update.
18240 (input_varpool_node): Update.
18241 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
18242 (lto_symtab_merge_symbols): Add sanity check.
18243 (lto_symtab_prevailing_decl): Do not special case weakrefs.
18244 * passes.c (rest_of_decl_compilation): Set static flag, too.
18245 * symtab.c (dump_symtab_base): Dump weakref.
18246 (verify_symtab_base): Sanity check weakrefs.
18247 (symtab_make_decl_local): Remove duplicated code.
18248 (symtab_alias_ultimate_target): Simplify.
18249 * varpool.c (varpool_create_variable_alias): Set weakref flag.
18251 2013-06-11 Tom de Vries <tom@codesourcery.com>
18253 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
18254 sequence_vect == NULL.
18256 2013-06-11 DJ Delorie <dj@redhat.com>
18258 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
18259 (rl78_unwind_word_mode): New.
18261 2013-06-11 David Malcolm <dmalcolm@redhat.com>
18263 * final.c (debug_prefix_maps): Make static.
18265 2013-06-11 David Malcolm <dmalcolm@redhat.com>
18267 * function.c (initial_trampoline): Remove stray copy.
18269 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
18271 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
18273 2013-06-11 Martin Jambor <mjambor@suse.cz>
18275 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
18276 within bounds at the beginning of the function.
18278 2013-06-11 Alan Modra <amodra@gmail.com>
18280 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
18282 (get_named_section): Don't NULL !DECL_P decl.
18284 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
18286 * doc/invoke.texi (core-avx2): Document.
18288 (atom): Updated with MOVBE.
18290 2013-06-11 Richard Biener <rguenther@suse.de>
18292 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
18294 2013-06-11 Anton Blanchard <anton@samba.org>
18296 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
18297 correct shift value in little-endian mode.
18299 2013-06-11 Jakub Jelinek <jakub@redhat.com>
18302 * varasm.c (get_variable_align): Move #endif to the right place.
18304 2013-06-10 Cary Coutant <ccoutant@google.com>
18306 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
18307 for hash so that hash table traversal order is deterministic.
18309 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
18310 Pat Haugen <pthaugen@us.ibm.com>
18311 Peter Bergner <bergner@vnet.ibm.com>
18313 * config/rs6000/vector.md (GPR move splitter): Do not split moves
18314 of vectors in GPRS if they are direct moves or quad word load or
18317 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
18319 (direct_move_p): Likewise.
18320 (quad_load_store_p): Likewise.
18322 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
18323 classes into bins based on the physical register type.
18324 (reg_class_to_reg_type): Likewise.
18325 (IS_STD_REG_TYPE): Likewise.
18326 (IS_FP_VECT_REG_TYPE): Likewise.
18327 (reload_fpr_gpr): Arrays to determine what insn to use if we can
18328 use direct move instructions.
18329 (reload_gpr_vsx): Likewise.
18330 (reload_vsx_gpr): Likewise.
18331 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
18332 information that is a simplification of register classes. Also
18333 precalculate direct move reload helpers.
18334 (direct_move_p): New function to return true if the operation can
18335 be done as a direct move instruciton.
18336 (quad_load_store_p): New function to return true if the operation
18337 is a quad memory operation.
18338 (rs6000_legitimize_address): If quad memory, only allow register
18339 indirect for TImode addresses.
18340 (rs6000_legitimate_address_p): Likewise.
18341 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
18342 (rs6000_reload_register_type): Likewise.
18343 (register_to_reg_type): Return register type.
18344 (rs6000_secondary_reload_simple_move): New helper function for
18345 secondary reload and secondary memory needed to identify anything
18346 that is a simple move, and does not need reloading.
18347 (rs6000_secondary_reload_direct_move): New helper function for
18348 secondary reload to identify cases that can be done with several
18349 instructions via the direct move instructions.
18350 (rs6000_secondary_reload_move): New helper function for secondary
18351 reload to identify moves between register types that can be done.
18352 (rs6000_secondary_reload): Add support for quad memory operations
18353 and for direct move.
18354 (rs6000_secondary_memory_needed): Likewise.
18355 (rs6000_debug_secondary_memory_needed): Change argument names.
18356 (rs6000_output_move_128bit): New function to return the move to
18357 use for 128-bit moves, including knowing about the various
18358 limitations of quad memory operations.
18360 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
18361 memory operations. call rs6000_output_move_128bit for the actual
18362 instruciton(s) to generate.
18363 (vsx_movti_64bit): Likewise.
18365 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
18366 (UNSPEC_P8V_MTVSRWZ): Likewise.
18367 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
18368 (UNSPEC_P8V_MTVSRD): Likewise.
18369 (UNSPEC_P8V_XXPERMDI): Likewise.
18370 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
18371 (UNSPEC_FUSION_GPR): Likewise.
18372 (FMOVE128_GPR): New iterator for direct move.
18373 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
18374 (f32_sv): Likewise.
18375 (f32_dm): Likewise.
18376 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
18377 loads and direct move instructions.
18378 (zero_extendsidi2_lfiwzx): Likewise.
18379 (extendsidi2_lfiwax): Likewise.
18380 (extendsidi2_nocell): Likewise.
18381 (floatsi<mode>2_lfiwax): Likewise.
18382 (lfiwax): Likewise.
18383 (floatunssi<mode>2_lfiwzx): Likewise.
18384 (lfiwzx): Likewise.
18385 (fix_trunc<mode>_stfiwx): Likewise.
18386 (fixuns_trunc<mode>_stfiwx): Likewise.
18387 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
18388 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
18389 (parity<mode>2_cmpb): Set length/type attr.
18390 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
18391 for 'mr.' to fast_compare.
18392 (bpermd_<mode>): Change type attr to popcnt.
18393 (p8_fmrgow_<mode>): New insns for power8 direct move support.
18394 (p8_mtvsrwz_1): Likewise.
18395 (p8_mtvsrwz_2): Likewise.
18396 (reload_fpr_from_gpr<mode>): Likewise.
18397 (p8_mtvsrd_1): Likewise.
18398 (p8_mtvsrd_2): Likewise.
18399 (p8_xxpermdi_<mode>): Likewise.
18400 (reload_vsx_from_gpr<mode>): Likewise.
18401 (reload_vsx_from_gprsf): Likewise.
18402 (p8_mfvsrd_3_<mode>): LIkewise.
18403 (reload_gpr_from_vsx<mode>): Likewise.
18404 (reload_gpr_from_vsxsf): Likewise.
18405 (p8_mfvsrd_4_disf): Likewise.
18406 (multi-word GPR splits): Do not split direct moves or quad memory
18409 2013-06-10 David Malcolm <dmalcolm@redhat.com>
18411 * tree-into-ssa.c (interesting_blocks): Make static.
18413 2013-06-10 Jakub Jelinek <jakub@redhat.com>
18416 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
18417 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
18418 Use DATA_ABI_ALIGNMENT for that case instead if defined.
18419 (get_variable_align): New function.
18420 (get_variable_section, emit_bss, emit_common,
18421 assemble_variable_contents, place_block_symbol): Use
18422 get_variable_align instead of DECL_ALIGN.
18423 (assemble_noswitch_variable): Add align argument, use it
18424 instead of DECL_ALIGN.
18425 (assemble_variable): Adjust caller. Use get_variable_align
18426 instead of DECL_ALIGN.
18427 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
18429 (DATA_ABI_ALIGNMENT): Define.
18430 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
18431 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
18432 opt is false, only return the psABI mandated alignment increase.
18433 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
18434 (DATA_ABI_ALIGNMENT): ... this.
18435 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
18436 (DATA_ABI_ALIGNMENT): ... this.
18437 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
18438 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
18439 (DATA_ABI_ALIGNMENT): ... this.
18440 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
18441 * doc/tm.texi: Regenerated.
18443 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
18445 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
18446 cmp_code to construct REG_EQUAL note.
18448 2013-06-09 Jakub Jelinek <jakub@redhat.com>
18451 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
18452 that operands[2] doesn't overlap with operands[0].
18454 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
18455 Jan Hubicka <jh@suse.cz>
18457 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
18458 hack to mark symbols as used.
18460 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
18462 PR rtl-optimization/57559
18463 * lra-constraints.c (process_alt_operands): Don't discourage
18464 memory with known offset for offsetable memory constraint.
18465 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
18467 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
18469 * varasm.c (struct oc_local_state): Reorder fields.
18470 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
18471 and adjust accordingly.
18472 (output_constructor): Reorder initialization code and adjust call to
18473 output_constructor_bitfield.
18475 2013-06-07 Jan Hubicka <jh@suse.cz>
18477 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
18479 2013-06-07 David Malcolm <dmalcolm@redhat.com>
18481 * tree-object-size.c (unknown): Make const.
18483 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18485 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
18486 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
18487 for last alternative in the cpu_facility attribute.
18489 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18492 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
18493 (xordi3): Change operand 2 constraint to arm_xordi_operand.
18494 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
18495 * config/arm/constraints.md (Dg): New constraint.
18496 * config/arm/neon.md (xordi3_neon): Remove.
18497 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
18498 * config/arm/predicates.md (arm_xordi_operand): New predicate.
18500 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18502 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
18503 Clean up alternatives.
18505 2013-06-07 Alan Modra <amodra@gmail.com>
18507 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
18510 2013-06-07 Alan Modra <amodra@gmail.com>
18512 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
18514 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18516 * config/arm/constraints.md (Df): New constraint.
18517 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
18518 Correct length attribute for last two alternatives.
18520 2013-06-07 Alan Modra <amodra@gmail.com>
18522 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
18523 override user -mfp-in-toc.
18524 (offsettable_ok_by_alignment): Consider just the current access
18525 rather than the whole object, unless BLKmode. Handle
18526 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
18527 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
18528 for -mcmodel=medium.
18529 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
18530 override user -mfp-in-toc or -msum-in-toc. Default to
18531 -mno-fp-in-toc for -mcmodel=medium.
18533 2013-06-06 DJ Delorie <dj@redhat.com>
18535 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
18536 TARGET_VALID_POINTER_MODE.
18538 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
18539 Pat Haugen <pthaugen@us.ibm.com>
18540 Peter Bergner <bergner@vnet.ibm.com>
18542 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
18543 Document new power8 builtins.
18545 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
18546 condition code register, to allow 128-bit logical operations to be
18547 done in the VSX or GPR registers.
18548 (nor<mode>3): Use the canonical form for nor.
18549 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
18550 vclz*, and vpopcnt* vector instructions.
18551 (nand<mode>3): Likewise.
18552 (orc<mode>3): Likewise.
18553 (clz<mode>2): LIkewise.
18554 (popcount<mode>2): Likewise.
18556 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
18557 that only the GPRs are recognized.
18559 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18560 support for new power8 builtins.
18562 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
18564 (xscvdpspn): Likewise.
18570 (vpopcnt): Likewise.
18571 (vpopcntb): Likewise.
18572 (vpopcnth): Likewise.
18573 (vpopcntw): Likewise.
18574 (vpopcntd): Likewise.
18576 (vmrgew): Likewise.
18577 (vmrgow): Likewise.
18579 (eqv_v16qi3): Likewise.
18580 (eqv_v8hi3): Likewise.
18581 (eqv_v4si3): Likewise.
18582 (eqv_v2di3): Likewise.
18583 (eqv_v4sf3): Likewise.
18584 (eqv_v2df3): Likewise.
18586 (nand_v16qi3): Likewise.
18587 (nand_v8hi3): Likewise.
18588 (nand_v4si3): Likewise.
18589 (nand_v2di3): Likewise.
18590 (nand_v4sf3): Likewise.
18591 (nand_v2df3): Likewise.
18593 (orc_v16qi3): Likewise.
18594 (orc_v8hi3): Likewise.
18595 (orc_v4si3): Likewise.
18596 (orc_v2di3): Likewise.
18597 (orc_v4sf3): Likewise.
18598 (orc_v2df3): Likewise.
18600 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
18601 allow power8 quad mode in 64-bit.
18602 (rs6000_builtin_vectorized_function): Add support to vectorize
18603 ISA 2.07 count leading zeros, population count builtins.
18604 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
18605 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
18606 (builtin_function_type): Add vgbbd builtin function which takes an
18608 (altivec_expand_vec_perm_const): Add support for new power8 merge
18611 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
18612 that does not include TImdoe for use with 32-bit.
18613 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
18615 (UNSPEC_VSX_CVDPSPN): Likewise.
18616 (vsx_xscvdpspn): Likewise.
18617 (vsx_xscvspdpn): Likewise.
18618 (vsx_xscvdpspn_scalar): Likewise.
18619 (vsx_xscvspdpn_directmove): Likewise.
18620 (vsx_and<mode>3): Split logical operations into 32-bit and
18621 64-bit. Add support to do logical operations on TImode as well as
18622 VSX vector types. Allow logical operations to be done in either
18623 VSX registers or in general purpose registers in 64-bit mode. Add
18624 splitters if GPRs were used. For AND, add clobber of CCmode to
18625 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
18627 (vsx_and<mode>3_32bit): Likewise.
18628 (vsx_and<mode>3_64bit): Likewise.
18629 (vsx_ior<mode>3): Likewise.
18630 (vsx_ior<mode>3_32bit): Likewise.
18631 (vsx_ior<mode>3_64bit): Likewise.
18632 (vsx_xor<mode>3): Likewise.
18633 (vsx_xor<mode>3_32bit): Likewise.
18634 (vsx_xor<mode>3_64bit): Likewise.
18635 (vsx_one_cmpl<mode>2): Likewise.
18636 (vsx_one_cmpl<mode>2_32bit): Likewise.
18637 (vsx_one_cmpl<mode>2_64bit): Likewise.
18638 (vsx_nor<mode>3): Likewise.
18639 (vsx_nor<mode>3_32bit): Likewise.
18640 (vsx_nor<mode>3_64bit): Likewise.
18641 (vsx_andc<mode>3): Likewise.
18642 (vsx_andc<mode>3_32bit): Likewise.
18643 (vsx_andc<mode>3_64bit): Likewise.
18644 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
18645 and xxlorc instructions.
18646 (vsx_eqv<mode>3_64bit): Likewise.
18647 (vsx_nand<mode>3_32bit): Likewise.
18648 (vsx_nand<mode>3_64bit): Likewise.
18649 (vsx_orc<mode>3_32bit): Likewise.
18650 (vsx_orc<mode>3_64bit): Likewise.
18652 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
18654 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
18656 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
18657 (p8_vmrgow): Likewise.
18658 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
18659 GPRs to be split under VSX.
18660 (p8v_clz<mode>2): Add power8 count leading zero support.
18661 (p8v_popcount<mode>2): Add power8 population count support.
18662 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
18665 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
18668 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
18670 (vec_nand): Likewise.
18671 (vec_vclz): Likewise.
18672 (vec_vclzb): Likewise.
18673 (vec_vclzd): Likewise.
18674 (vec_vclzh): Likewise.
18675 (vec_vclzw): Likewise.
18676 (vec_vgbbd): Likewise.
18677 (vec_vmrgew): Likewise.
18678 (vec_vmrgow): Likewise.
18679 (vec_vpopcnt): Likewise.
18680 (vec_vpopcntb): Likewise.
18681 (vec_vpopcntd): Likewise.
18682 (vec_vpopcnth): Likewise.
18683 (vec_vpopcntw): Likewise.
18685 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18687 PR rtl-optimization/57468
18688 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
18691 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18693 PR rtl-optimization/57459
18694 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
18695 type when setting live regs.
18697 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18699 * config/s390/s390.opt (mlra): New option.
18700 * config/s390/s390.c (s390_decompose_address): Check displacement
18701 for all registers for LRA.
18702 (s390_secondary_reload): Don't used secondary reloads for LRA.
18703 (s390_lra_p): New function.
18704 (TARGET_LRA_P): Define.
18705 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
18706 of attribute cpu_facility to zarch for the last alternative.
18707 (*cmpmem_short): Ditto.
18709 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
18711 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
18712 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
18713 (arm_expand_prologue): Likewise.
18715 2013-06-06 Teresa Johnson <tejohnson@google.com>
18718 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
18719 as this is now done by redirect_edge_and_branch_force.
18720 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
18721 barriers, and fix interaction with splitting.
18722 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
18723 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
18724 reflect changes made in the routine.
18725 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
18726 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
18727 since this is called in cfglayout mode, and replace partition fixup
18728 with assert as that is now done by force_nonfallthru_and_redirect.
18729 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
18730 already be marked with region crossing note.
18731 (insert_section_boundary_note): Make non-static, gate on flag
18732 has_bb_partition, rewrite to also check for multiple partitions.
18733 (rest_of_handle_reorder_blocks): Remove call to
18734 insert_section_boundary_note, now done later during free_cfg.
18735 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
18736 * bb-reorder.h (insert_section_boundary_note): Declare.
18737 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
18738 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
18739 invoke insert_section_boundary_note.
18740 (try_redirect_by_replacing_jump): Remove unnecessary
18741 check for region crossing note.
18742 (fixup_partition_crossing): New function.
18743 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
18744 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
18745 in non-cfglayout mode.
18746 (force_nonfallthru_and_redirect): Fixup partition boundaries,
18747 remove old code that tried to do this. Emit barrier correctly
18748 when we are in cfglayout mode.
18749 (last_bb_in_partition): New function.
18750 (rtl_split_edge): Correctly fixup partition boundaries.
18751 (commit_one_edge_insertion): Remove old code that tried to
18752 fixup region crossing edge since this is now handled in
18753 split_block, and set up insertion point correctly since
18754 block may now end in a jump.
18755 (verify_hot_cold_block_grouping): Guard against checking when not in
18756 linearized RTL mode.
18757 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
18759 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
18760 rtl_verify_flow_info, so not called in cfglayout mode.
18761 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
18762 (fixup_reorder_chain): Remove old code that attempted to fixup region
18763 crossing note as this is now handled in force_nonfallthru_and_redirect.
18764 (duplicate_insn_chain): Don't duplicate switch section notes.
18765 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
18767 * basic-block.h (emit_barrier_after_bb): Declare.
18769 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18771 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
18772 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
18773 arm_usatsihi): Adjust alternatives for arm_restrict_it.
18775 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18777 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
18779 * config/arm/ldmstm.md: Regenerate.
18781 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18783 * config/arm/sync.md (atomic_loaddi_1):
18784 Disable predication for arm_restrict_it.
18785 (arm_load_exclusive<mode>): Likewise.
18786 (arm_load_exclusivesi): Likewise.
18787 (arm_load_exclusivedi): Likewise.
18788 (arm_load_acquire_exclusive<mode>): Likewise.
18789 (arm_load_acquire_exclusivesi): Likewise.
18790 (arm_load_acquire_exclusivedi): Likewise.
18791 (arm_store_exclusive<mode>): Likewise.
18792 (arm_store_exclusive<mode>): Likewise.
18793 (arm_store_release_exclusivedi): Likewise.
18794 (arm_store_release_exclusive<mode>): Likewise.
18796 2013-06-06 Richard Biener <rguenther@suse.de>
18798 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
18800 (lto_tag_is_tree_code_p): Adjust.
18801 (lto_tag_is_gimple_code_p): Likewise.
18802 (lto_gimple_code_to_tag): Likewise.
18803 (lto_tag_to_gimple_code): Likewise.
18804 (lto_tree_code_to_tag): Likewise.
18805 (lto_tag_to_tree_code): Likewise.
18806 * data-streamer.h (streamer_write_hwi_in_range): Use
18807 uhwi streaming to stream the normalized range.
18808 (streamer_read_hwi_in_range): Likewise.
18810 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18812 * config/arm/arm.md (enabled_for_depr_it): New attribute.
18813 (predicable_short_it): Likewise.
18814 (predicated): Likewise.
18815 (enabled): Handle above.
18816 (define_cond_exec): Set predicated attribute to yes.
18818 2013-06-05 Mike Stump <mikestump@comcast.net>
18820 * gdbinit.in (__FUNCTION__): Add.
18822 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
18824 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
18825 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
18827 2013-06-05 Jan Hubicka <jh@suse.cz>
18829 * varasm.c (mark_decl_referenced): Revert the removal until targets
18832 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
18834 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
18835 instead of mark_decl_referenced.
18837 2013-06-05 Jan Hubicka <jh@suse.cz>
18839 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
18840 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
18841 and symtab_used_from_object_file_p.
18842 (cgraph_make_node_local_1): Clear forced_by_abi.
18843 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
18844 * cgraph.h (symtab_node_base): Add forced_by_abi.
18845 (decide_is_variable_needed): Remove.
18846 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
18847 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
18848 (decide_is_symbol_needed): ... this one; handle symbols in general;
18849 always analyze virtuals; honnor forced_by_abi.
18850 (cgraph_finalize_function): Update.
18851 (varpool_finalize_decl): Update.
18852 (symbol_defined_and_needed): Remove.
18853 (analyze_functions): Update.
18854 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18855 output_refs, input_overwrite_node): Handle forced_by_abi.
18856 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18857 (address_taken_from_non_vtable_p): ... this one.
18858 (comdat_can_be_unshared_p_1): New function.
18859 (cgraph_comdat_can_be_unshared_p): Rename to ...
18860 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18861 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18862 (function_and_variable_visibility): Clear forced_by_abi as needed.
18863 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18864 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18865 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18866 * varpool.c (decide_is_variable_needed): Remove.
18868 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18870 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18871 (arm_option_override): Override arm_restrict_it where appropriate.
18872 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18873 * config/arm/arm.opt (mrestrict-it): New command-line option.
18874 * doc/invoke.texi: Document -mrestrict-it.
18876 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18878 * tsan.c (tsan_atomic_table): Make const.
18880 2013-06-05 Richard Biener <rguenther@suse.de>
18882 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18883 index associated with the tree we are supposed to replace.
18884 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18885 TREE_ASM_WRITTEN as zero for everything but SSA names.
18887 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18889 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18891 2013-06-04 Jan Hubicka <jh@suse.cz>
18893 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18894 (input_node, input_varpool_node): Handle correctly external same
18896 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18897 nodes at ltrans stage.
18899 2013-06-04 Jan Hubicka <jh@suse.cz>
18901 * ipa-inline.c (update_caller_keys): Fix availability test.
18902 (update_callee_keys): Likewise.
18903 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18904 to follow ELF standard.
18906 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18908 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18909 (mips64r5900el-*-elf*): New configurations.
18910 * config/mips/mips-cpus.def (r5900): New processor.
18911 * config/mips/mips-tables.opt: Regenerate.
18912 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18913 (mips_issue_rate): Handle PROCESSOR_R5900.
18914 (mips_reorg_process_insns): Force reorder mode for the R5900.
18915 * config/mips/mips.h (TARGET_MIPS5900): Define.
18916 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18918 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18920 * config/mips/mips.md (processor): Add r5900.
18921 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18923 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18925 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18926 into function to generate MOVI instruction.
18927 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18928 (aarch64_preferred_simd_mode): Turn into wrapper.
18929 (aarch64_output_scalar_simd_mov_immediate): New function.
18930 * config/aarch64/aarch64-protos.h: Add prototype for above.
18932 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18934 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18935 element_char member.
18936 (sizetochar): Return signed char.
18937 (aarch64_simd_valid_immediate): Remove elchar and other
18938 unnecessary variables.
18939 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18940 Calculate element_char as required.
18941 * config/aarch64/aarch64-protos.h: Update and move prototype
18942 for aarch64_output_simd_mov_immediate.
18943 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18946 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18948 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18949 information completed by aarch64_simd_valid_immediate.
18950 (aarch64_legitimate_constant_p): Update arguments.
18951 (aarch64_simd_valid_immediate): Work with struct rather than many
18953 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18954 (aarch64_simd_make_constant): Update arguments.
18955 (aarch64_output_simd_mov_immediate): Work with struct rather than
18956 many pointers. Output immediate directly rather than as operand.
18957 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18959 * config/aarch64/constraints.md (Dn): Update arguments.
18961 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18963 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18965 (aarch64_simd_immediate_valid_for_move): Remove.
18966 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18967 (aarch64_simd_make_constant): Update call.
18968 (aarch64_output_simd_mov_immediate): Update call.
18969 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18971 * config/aarch64/constraints.md (Dn): Update call.
18973 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18975 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18976 return type to bool for prototype.
18977 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18978 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18979 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18981 2013-06-04 Catherine Moore <clm@codesourcery.com>
18983 * config/mips/mips.opt (meva): New.
18984 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18985 (ASM_SPEC): Handle -meva.
18986 * doc/invoke.texi (meva): Document.
18988 2013-06-04 Alan Modra <amodra@gmail.com>
18990 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18993 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18995 * rtl.def: Add extra fourth optional field to define_cond_exec.
18996 * gensupport.c (process_one_cond_exec): Process attributes from
18998 * doc/md.texi: Document fourth field in define_cond_exec.
19000 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
19002 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
19003 out the processing order as in store_bit_field_1.
19005 2013-06-04 Jan Hubicka <jh@suse.cz>
19007 PR middle-end/57500
19008 * cgraphunit.c (cgraph_process_same_body_aliases): Create
19009 non-VAR_DECL node if it does not exist yet.
19011 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
19013 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
19014 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
19015 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
19016 target_cpu_default setting.
19018 2013-06-03 Teresa Johnson <tejohnson@google.com>
19020 * dumpfile.c (opt_info_switch_p): Change -fopt-info
19021 default to -fopt-info=optimized instead of all.
19022 * doc/invoke.texi: Ditto.
19023 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
19024 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
19025 (execute_vect_slp): Emit BB vectorization success under
19026 MSG_OPTIMIZED_LOCATIONS.
19027 * tree-vect-slp.c (vect_slp_transform_bb): Change
19028 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
19029 * tree-vect-loop.c (vect_transform_loop): Ditto.
19031 2013-06-03 Jason Merrill <jason@redhat.com>
19034 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
19035 Use TARGET_EXPR for C++.
19037 2013-06-03 Jakub Jelinek <jakub@redhat.com>
19039 PR rtl-optimization/57268
19040 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
19041 if DEBUG_INSN_P (insn).
19044 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
19046 PR rtl-optimization/57268
19047 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19048 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19050 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19052 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
19053 (ix86_avoid_lea_for_addr): Likewise.
19054 (exact_dependency_1): Likewise.
19055 (ix86_adjust_cost): Likewise.
19056 (swap_top_of_ready_list): Fix formatting and !reload_completed check
19058 (do_reorder_for_imul): Fix typo, formatting and
19059 !reload_completed check removed.
19060 (ix86_sched_reorder): Fix typo and formatting.
19061 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
19064 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
19066 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
19068 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
19070 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
19071 <STRING_CST>: Likewise.
19072 <VECTOR_CST>: Likewise.
19074 2013-06-01 Janus Weil <janus@gcc.gnu.org>
19075 Mikael Morin <mikael@gcc.gnu.org>
19077 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
19078 * config.in: Regenerated.
19079 * configure: Regenerated.
19081 2013-06-01 Jan Hubicka <jh@suse.cz>
19083 PR middle-end/57366
19084 * cgraphunit.c (compile): When weakref is not supported,
19085 set up transparent aliases before final output pass.
19086 * varasm.c (assemble_alias): Do not try to do it here.
19088 2013-06-01 Jan Hubicka <jh@suse.cz>
19090 PR middle-end/57467
19091 * passes.c (for_per_function): Skip unanalyzed functions.
19093 2013-06-01 Jan Hubicka <jh@suse.cz>
19095 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
19096 (lto_symtab_merge_symbols_1): ... this one.
19097 (lto_symtab_merge_cgraph_nodes): Rename to ...
19098 (lto_symtab_merge_symbols): ... this one; simplify.
19099 * cgraph.c (same_body_aliases_done): Rename to ...
19100 (cpp_implicit_aliases_done): ... this one.
19101 (cgraph_create_function_alias): Update.
19102 (cgraph_same_body_alias): Update.
19103 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
19104 (verify_edge_corresponds_to_fndecl): Simplify.
19105 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
19106 (cgraph_node): Remove same_body_alias.
19107 (varpool_node): Remove alias_of and extra_name_alias.
19108 (same_body_aliases_done): Rename to ..
19109 (cpp_implicit_aliases_done): ... this one.
19110 (symtab_alias_ultimate_target): Add default parameter.
19111 (symtab_resolve_alias): New function.
19112 (fixup_same_cpp_alias_visibility): Declare.
19113 (cgraph_function_node): Add default parameter.
19114 (cgraph_node_asm_name): Likewise.
19115 (cgraph_function_or_thunk_node): Add default parameter; do
19116 not ICE when it is NULL.
19117 (varpool_variable_node): Likewise.
19118 * tree-emutls.c (create_emultls_var): Update.
19119 (ipa_lower_emutls): Update.
19120 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
19121 (cgraph_reset_node): Reset alias info.
19122 (cgraph_finalize_function): Update.
19123 (fixup_same_cpp_alias_visibility): Move to symtab.c.
19124 (analyze_function): Simplify.
19125 (cgraph_process_same_body_aliases): Simplify.
19126 (analyze_functions): Fixup same body aliases.
19127 (handle_alias_pairs): Simplify.
19128 (assemble_thunk): Update.
19129 (assemble_thunks_and_aliases): Update.
19130 (output_weakrefs): Rewrite.
19131 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
19132 (lto_output_varpool_node): Likewise.
19133 (compute_ltrans_boundary): Remve assert.
19134 (get_alias_symbol): New functoin.
19135 (input_node): Rewrite alias handling.
19136 (input_varpool_node): Likewise.
19137 * ipa-pure-const.c (propagate_pure_const): Fix formating.
19138 * ipa.c (process_references): Handle weakrefs correctly.
19139 (symtab_remove_unreachable_nodes): Likewise.
19140 * trans-mem.c (get_cg_data): Update.
19141 (ipa_tm_create_version_alias): Update.
19142 (ipa_tm_execute): Update.
19143 * symtab.c (dump_symtab_base): Dump aliases.
19144 (verify_symtab_base): Verify aliases.
19145 (symtab_node_availability): New function.
19146 (symtab_alias_ultimate_target): Simplify.
19147 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
19148 handle all the fixup cases.
19149 (symtab_resolve_alias): New function.
19150 * passes.c (ipa_write_summaries): Handle weakrefs.
19151 * varpool.c (varpool_analyze_node): Simplify.
19152 (assemble_aliases): Update.
19153 (varpool_create_variable_alias): Simplify.
19154 (varpool_extra_name_alias): Simplify.
19155 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
19156 (lto_symtab_merge_symbols): ... this one.
19158 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
19161 PR rtl-optimization/57268
19162 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19163 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19165 2013-06-01 Tobias Burnus <burnus@net-b.de>
19167 Partially reverted:
19168 2013-05-31 Tobias Burnus <burnus@net-b.de>
19170 PR middle-end/57073
19171 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
19174 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
19176 PR rtl-optimization/57268
19177 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19178 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19180 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
19182 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
19183 unordered comparison operators when -fno-trapping-math is in effect
19185 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
19186 and implement unordered comparison operators properly on the e500.
19188 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
19190 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
19191 for constant scalar integers.
19192 (simplify_relational_operation_1): Likewise.
19194 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
19196 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
19197 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
19200 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19201 Igor Zamyatin <igor.zamyatin@intel.com>
19203 Silvermont (SLM) architecture performance tuning.
19204 * config/i386/i386.h (enum ix86_tune_indices): Add
19205 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
19206 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
19208 * config/i386/i386.c (initial_ix86_tune_features)
19209 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
19210 (ix86_lea_outperforms): Handle Silvermont tuning.
19211 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
19213 (ix86_use_lea_for_mov): Likewise.
19214 (ix86_avoid_lea_for_addr): Likewise.
19215 (ix86_lea_for_add_ok): Likewise.
19216 (exact_dependency_1): New function.
19217 (exact_store_load_dependency): Likewise.
19218 (ix86_adjust_cost): Handle Silvermont tuning.
19219 (do_reoder_for_imul): Likewise.
19220 (swap_top_of_ready_list): New function.
19221 (ix86_sched_reorder): Changed to handle Silvermont tuning.
19223 * config/i386/i386.md (peepholes that split memory operand in fp
19226 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
19228 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
19229 Remove un-necessary braces.
19231 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
19233 * config/aarch64/aarch64.c (aarch64_classify_symbol):
19234 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
19236 2013-05-31 Tobias Burnus <burnus@net-b.de>
19238 PR middle-end/57073
19239 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
19241 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19244 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
19245 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
19246 * config/arm/neon.md (iordi3_neon): Remove.
19247 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
19248 * config/arm/predicates.md (imm_for_neon_logic_operand):
19249 Move to earlier in the file.
19250 (neon_logic_op2): Likewise.
19251 (arm_iordi_operand_neon): New predicate.
19253 2013-05-31 Richard Biener <rguenther@suse.de>
19255 PR tree-optimization/57478
19256 PR tree-optimization/57453
19257 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
19260 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19262 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
19263 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
19265 2013-05-30 Tobias Burnus <burnus@net-b.de>
19266 Thomas Koenig <tkoenig@gcc.gnu.org>
19268 PR middle-end/57073
19269 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
19270 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
19272 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
19274 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
19276 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
19278 * target.def (register_usage_leveling_p): New hook.
19279 * targhooks.c (default_register_usage_leveling_p): New.
19280 * targhooks.h (default_register_usage_leveling_p): New prototype.
19281 * lra-assigns.c (register_usage_leveling_p): Use the hook.
19282 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
19283 * doc/tm.texi: Update.
19284 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
19286 2013-05-30 Ian Bolton <ian.bolton@arm.com>
19288 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
19289 (*insv_reg<mode>): New define_insn.
19291 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
19293 PR rtl-optimization/57439
19294 * postreload.c (move2add_valid_value_p): Check that we have
19295 a zero subreg_regno_offset when accessing the register in
19296 the requested mode.
19298 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19299 Igor Zamyatin <igor.zamyatin@intel.com>
19301 Silvermont (SLM) architecture pipeline model, tuning and
19303 * config.gcc: Add slm config options and target.
19305 * config/i386/slm.md: New.
19307 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
19309 * config/i386/i386-c.c (ix86_target_macros_internal): New case
19311 (ix86_target_macros_internal): Likewise.
19313 * config/i386/i386.c (slm_cost): New cost.
19314 (m_SLM): New macro flag.
19315 (initial_ix86_tune_features): Set m_SLM.
19316 (x86_accumulate_outgoing_args): Likewise.
19317 (x86_arch_always_fancy_math_387): Likewise.
19318 (processor_target_table): Add slm cost.
19319 (cpu_names): Add slm cpu name.
19320 (x86_option_override_internal): Set SLM ISA.
19321 (ix86_issue_rate): New case PROCESSOR_SLM.
19322 (ia32_multipass_dfa_lookahead): Likewise.
19323 (fold_builtin_cpu): Add slm.
19325 * config/i386/i386.h (TARGET_SLM): New target macro.
19326 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
19327 (processor_type): Add PROCESSOR_SLM.
19329 * config/i386/i386.md (cpu): Add new value "slm".
19330 (slm.md): Include slm.md.
19332 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
19333 Zhenqiang Chen <zhenqiang.chen@linaro.org>
19335 * config/arm/arm-protos.h: Add and update function protos.
19336 * config/arm/arm.c (use_simple_return_p): New added.
19337 (thumb2_expand_return): Check simple_return flag.
19338 * config/arm/arm.md: Add simple_return and conditional simple_return.
19339 * config/arm/iterators.md: Add iterator for return and simple_return.
19341 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
19343 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
19344 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
19345 (arm_emit_vfp_multi_reg_pop): Likewise.
19346 (thumb2_emit_ldrd_pop): Likewise.
19347 (arm_expand_epilogue): Add misc REG_CFA notes.
19348 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
19350 2013-05-29 Lawrence Crowl <crowl@google.com>
19352 * config/arm/t-arm: Update for below.
19354 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
19355 Change type to hash_table. Update dependent calls and types.
19357 * config/i386/t-cygming: Update for below.
19359 * config/i386/t-interix: Update for below.
19361 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
19362 Change type to hash_table. Update dependent calls and types.
19363 (i386_find_on_wrapper_list::wrappers): Likewise.
19365 * config/ia64/t-ia64: Update for below.
19367 * config/ia64/ia64.c (bundle_state_table):
19368 Change type to hash_table. Update dependent calls and types.
19370 * config/mips/mips.c (mips_reorg_process_insns::htab):
19371 Change type to hash_table. Update dependent calls and types.
19373 * config/sol2.c (solaris_comdat_htab):
19374 Change type to hash_table. Update dependent calls and types.
19376 * config/t-sol2: Update for above.
19378 2013-05-29 Teresa Johnson <tejohnson@google.com>
19380 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
19381 functions are not yet marked as defined.
19383 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
19384 Pat Haugen <pthaugen@us.ibm.com>
19385 Peter Bergner <bergner@vnet.ibm.com>
19387 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
19391 (vrotl<mode>3): Likewise.
19392 (vashl<mode>3): Likewise.
19393 (vlshr<mode>3): Likewise.
19394 (vashr<mode>3): Likewise.
19396 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
19397 support for power8 V2DI builtins.
19399 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
19400 power8 V2DI builtins.
19401 (vupkhsw): Likewise.
19402 (vupklsw): Likewise.
19403 (vaddudm): Likewise.
19404 (vminsd): Likewise.
19405 (vmaxsd): Likewise.
19406 (vminud): Likewise.
19407 (vmaxud): Likewise.
19408 (vpkudum): Likewise.
19409 (vpksdss): Likewise.
19410 (vpkudus): Likewise.
19411 (vpksdus): Likewise.
19416 (vsubudm): Likewise.
19417 (vcmpequd): Likewise.
19418 (vcmpgtsd): Likewise.
19419 (vcmpgtud): Likewise.
19420 (vcmpequd_p): Likewise.
19421 (vcmpgtsd_p): Likewise.
19422 (vcmpgtud_p): Likewise.
19423 (vupkhsw): Likewise.
19424 (vupklsw): Likewise.
19425 (vaddudm): Likewise.
19426 (vmaxsd): Likewise.
19427 (vmaxud): Likewise.
19428 (vminsd): Likewise.
19429 (vminud): Likewise.
19430 (vpksdss): Likewise.
19431 (vpksdus): Likewise.
19432 (vpkudum): Likewise.
19433 (vpkudus): Likewise.
19438 (vsubudm): Likewise.
19440 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
19441 support for power8 V2DI instructions.
19443 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
19444 power8 V2DI instructions. Combine pack and unpack insns to use an
19445 iterator for each mode. Check whether a particular mode supports
19446 Altivec instructions instead of just checking TARGET_ALTIVEC.
19447 (UNSPEC_VPKUWUM): Likewise.
19448 (UNSPEC_VPKSHSS): Likewise.
19449 (UNSPEC_VPKSWSS): Likewise.
19450 (UNSPEC_VPKUHUS): Likewise.
19451 (UNSPEC_VPKSHUS): Likewise.
19452 (UNSPEC_VPKUWUS): Likewise.
19453 (UNSPEC_VPKSWUS): Likewise.
19454 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
19455 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
19456 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
19457 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
19458 (UNSPEC_VUPKHSB): Likewise.
19459 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
19460 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
19461 (UNSPEC_VUPKHSH): Likewise.
19462 (UNSPEC_VUPKLSB): Likewise.
19463 (UNSPEC_VUPKLSH): Likewise.
19465 (VI_char): Likewise.
19466 (VI_scalar): Likewise.
19467 (VI_unit): Likewise.
19469 (VP_small): Likewise.
19470 (VP_small_lc): Likewise.
19471 (VU_char): Likewise.
19472 (add<mode>3): Likewise.
19473 (altivec_vaddcuw): Likewise.
19474 (altivec_vaddu<VI_char>s): Likewise.
19475 (altivec_vadds<VI_char>s): Likewise.
19476 (sub<mode>3): Likewise.
19477 (altivec_vsubcuw): Likewise.
19478 (altivec_vsubu<VI_char>s): Likewise.
19479 (altivec_vsubs<VI_char>s): Likewise.
19480 (altivec_vavgs<VI_char>): Likewise.
19481 (altivec_vcmpbfp): Likewise.
19482 (altivec_eq<mode>): Likewise.
19483 (altivec_gt<mode>): Likewise.
19484 (altivec_gtu<mode>): Likewise.
19485 (umax<mode>3): Likewise.
19486 (smax<mode>3): Likewise.
19487 (umin<mode>3): Likewise.
19488 (smin<mode>3): Likewise.
19489 (altivec_vpkuhum): Likewise.
19490 (altivec_vpkuwum): Likewise.
19491 (altivec_vpkshss): Likewise.
19492 (altivec_vpkswss): Likewise.
19493 (altivec_vpkuhus): Likewise.
19494 (altivec_vpkshus): Likewise.
19495 (altivec_vpkuwus): Likewise.
19496 (altivec_vpkswus): Likewise.
19497 (altivec_vpks<VI_char>ss): Likewise.
19498 (altivec_vpks<VI_char>us): Likewise.
19499 (altivec_vpku<VI_char>us): Likewise.
19500 (altivec_vpku<VI_char>um): Likewise.
19501 (altivec_vrl<VI_char>): Likewise.
19502 (altivec_vsl<VI_char>): Likewise.
19503 (altivec_vsr<VI_char>): Likewise.
19504 (altivec_vsra<VI_char>): Likewise.
19505 (altivec_vsldoi_<mode>): Likewise.
19506 (altivec_vupkhsb): Likewise.
19507 (altivec_vupkhs<VU_char>): Likewise.
19508 (altivec_vupkls<VU_char>): Likewise.
19509 (altivec_vupkhsh): Likewise.
19510 (altivec_vupklsb): Likewise.
19511 (altivec_vupklsh): Likewise.
19512 (altivec_vcmpequ<VI_char>_p): Likewise.
19513 (altivec_vcmpgts<VI_char>_p): Likewise.
19514 (altivec_vcmpgtu<VI_char>_p): Likewise.
19515 (abs<mode>2): Likewise.
19516 (vec_unpacks_hi_v16qi): Likewise.
19517 (vec_unpacks_hi_v8hi): Likewise.
19518 (vec_unpacks_lo_v16qi): Likewise.
19519 (vec_unpacks_hi_<VP_small_lc>): Likewise.
19520 (vec_unpacks_lo_v8hi): Likewise.
19521 (vec_unpacks_lo_<VP_small_lc>): Likewise.
19522 (vec_pack_trunc_v8h): Likewise.
19523 (vec_pack_trunc_v4si): Likewise.
19524 (vec_pack_trunc_<mode>): Likewise.
19526 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
19528 (vec_vmaxsd): Likewise.
19529 (vec_vmaxud): Likewise.
19530 (vec_vminsd): Likewise.
19531 (vec_vminud): Likewise.
19532 (vec_vpksdss): Likewise.
19533 (vec_vpksdus): Likewise.
19534 (vec_vpkudum): Likewise.
19535 (vec_vpkudus): Likewise.
19536 (vec_vrld): Likewise.
19537 (vec_vsld): Likewise.
19538 (vec_vsrad): Likewise.
19539 (vec_vsrd): Likewise.
19540 (vec_vsubudm): Likewise.
19541 (vec_vupkhsw): Likewise.
19542 (vec_vupklsw): Likewise.
19544 2013-05-29 Jan Hubicka <jh@suse.cz>
19546 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
19547 flags; reorder rest of fields in more consistent way.
19548 (varpool_node): Remove analyzed, finalized and alias.
19549 (cgraph_ndoe): Likewise.
19550 (symtab_alias_ultimate_target): New function.
19551 (cgraph_function_node): Move offline.
19552 (cgraph_reset_node): Declare.
19553 (cgraph_comdat_can_be_unshared_p): Remove.
19554 (varpool_remove_initializer): Declare.
19555 (varpool_first_defined_variable, varpool_next_defined_variable
19556 cgraph_first_defined_function, cgraph_next_defined_function): Update.
19557 (cgraph_function_with_gimple_body_p): Update.
19558 (varpool_all_refs_explicit_p): Update.
19559 (symtab_alias_target): New function.
19560 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
19561 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
19562 (cgraph_function_or_thunk_node): Simplify using
19563 symtab_alias_ultimate_target.
19564 (varpool_variable_node): Likewise.
19565 * cgraph.c (cgraph_create_function_alias): Update.
19566 (cgraph_add_thunk): Update.
19567 (cgraph_remove_node): Update.
19568 (dump_cgraph_node): Do not dump removed flags.
19569 (cgraph_function_body_availability): Update.
19570 (cgraph_propagate_frequency): Update.
19571 (verify_cgraph_node): Check sanity of local flag.
19572 (cgraph_function_node): Move here from cgraph.h; revamp for
19573 cgraph_function_or_thunk_node.
19574 * lto-symtab.c (lto_varpool_replace_node): Update.
19575 (lto_symtab_resolve_can_prevail_p): Update.
19576 (lto_symtab_merge_cgraph_nodes): Update.
19577 * ipa-cp.c (determine_versionability, initialize_node_lattices,
19578 propagate_constants_accross_call, devirtualization_time_bonus,
19579 ipcp_propagate_stage): Update.
19580 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
19581 * ipa-inline-transform.c (clone_inlined_nodes,
19582 preserve_function_body_p): Update.
19583 * ipa-reference.c (propagate): Update.
19584 (write_node_summary_p): Update.
19585 * toplev.c (wrapup_global_declaration_2): Update.
19586 * cgraphunit.c (cgraph_analyze_function): Rename to ...
19587 (analyze_function) ... this one.
19588 (cgraph_process_new_functions): Update.
19589 (cgraph_reset_node): Export.
19590 (cgraph_finalize_function): Update.
19591 (cgraph_add_new_function): Update.
19592 (process_function_and_variable_attributes): Update.
19593 (varpool_finalize_decl): Update.
19594 (symbol_finalized): Remove.
19595 (symbol_finalized_and_needed): Rename to ...
19596 (symbol_defined_and_needed): ... update.
19597 (cgraph_analyze_functions): Update.
19598 (handle_alias_pairs): Update.
19599 (mark_functions_to_output): Update.
19600 (assemble_thunk): Update.
19601 (output_in_order): Update.
19602 (output_weakrefs): Update.
19603 (finalize_compilation_unit): Update.
19604 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
19605 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
19606 input_node, input_varpool_node): Update.
19607 * dbxout.c (dbxout_expand_expr): Update.
19608 * cgraphclones.c (cgraph_clone_node): Update.
19609 (cgraph_copy_node_for_versioning): Update.
19610 (cgraph_materialize_clone): Update.
19611 (cgraph_materialize_all_clones): Update.
19612 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
19613 propagate_pure_const, propagate_nothrow): Update.
19614 * lto-streamer-out.c (lto_output, write_symbol): Update.
19615 * ipa-utils.c (ipa_reverse_postorder): Update.
19616 * ipa-inline.c (can_inline_edge_p): Update.
19617 (update_caller_keys, ipa_inline): Update.
19618 * dwarf2out.c (reference_to_unused,
19619 premark_types_used_by_global_vars_helper): Update.
19620 * tree-eh.c (tree_could_trap_p): Update.
19621 * ipa-split.c (consider_split, execute_split_functions): Update.
19622 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
19623 has_addr_references_p): Update; move ahead in file for better
19625 (process_references): Simplify.
19626 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
19627 bodies are removed.
19628 (cgraph_comdat_can_be_unshared_p): Make static.
19629 (cgraph_externally_visible_p): Update.
19630 (varpool_externally_visible_p): Update.
19631 (function_and_variable_visibility): Update.
19632 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
19633 ipa_tm_mark_force_output_node): Update.
19634 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
19635 estimate_edge_devirt_benefit, inline_generate_summary,
19636 inline_write_summary): Update.
19637 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
19638 * ipa-prop.c (ipa_compute_jump_functions): Update.
19639 (ipa_print_node_params, ipa_prop_read_section,
19640 ipa_update_after_lto_read, read_replacements_section): Update.
19641 * varasm.c (mark_decl_referenced): Update.
19642 (assemble_alias, dump_tm_clone_pairs): Update.
19643 * tree-inline.c (copy_bb): Update.
19644 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
19646 * symtab.c (dump_symtab_base): Print new flags.
19647 (verify_symtab_base): Verify new flags.
19648 (symtab_alias_ultimate_target): New function.
19649 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
19650 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
19652 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
19654 * i386.c (ix86_get_function_versions_dispatcher,
19655 ix86_generate_version_dispatcher_body): Update.
19656 (fold_builtin_cpu): Use varpool_add_new_variable.
19657 * varpool.c (varpool_remove_initializer): Break out from ...
19658 (varpool_remove_node): ... this one.
19659 (dump_varpool_node, varpool_node_for_asm,
19660 cgraph_variable_initializer_availability, varpool_analyze_node,
19661 varpool_assemble_decl, varpool_remove_unreferenced_decls,
19662 varpool_finalize_named_section_flags, varpool_create_variable_alias):
19665 2013-05-29 Jan Hubicka <jh@suse.cz>
19667 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
19669 2013-05-29 Easwaran Raman <eraman@google.com>
19671 PR tree-optimization/57442
19672 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
19673 when control exits the main loop.
19675 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
19677 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
19679 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
19680 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
19681 * rx/t-rx: Add rx100 under multi library matches option for nofpu
19684 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19686 PR tree-optimization/57441
19687 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
19688 Don't limit size of incr_vec to number of candidates.
19690 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
19692 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
19693 and mips16 directories.
19694 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
19695 (MULTILIB_DIRNAMES): Ditto.
19696 (MULTILIB_EXCEPTIONS): Add new exceptions.
19697 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
19698 (MULTILIB_DIRNAMES): Ditto.
19699 (MULTILIB_EXCEPTIONS): Add new exceptions.
19701 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19702 Marcus Shawcroft <marcus.shawcroft@arm.com>
19704 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
19705 SYMBOL_TINY_ABSOLUTE.
19706 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
19707 SYMBOL_TINY_ABSOLUTE.
19708 (aarch64_expand_mov_immediate): Likewise.
19709 (aarch64_classify_symbol): Likewise.
19710 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
19711 Permit SYMBOL_TINY_ABSOLUTE.
19712 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
19714 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19715 Marcus Shawcroft <marcus.shawcroft@arm.com>
19717 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
19718 Refactor if/switch. Replace gcc_assert with if.
19720 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19722 * config/i386/i386.c (initial_ix86_tune_features): Enable
19723 FP Reassociation for AMD bdver1 and bdver2.
19725 2013-05-29 Martin Jambor <mjambor@suse.cz>
19727 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
19728 and IMAGPART_EXPR do not occur within other handled_components.
19730 2013-05-29 Richard Biener <rguenther@suse.de>
19732 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
19733 access on whether the use is in the BB we currently try to
19735 (vect_bb_vectorization_profitable_p): Pass the BB we currently
19736 vectorize to vect_bb_slp_scalar_cost.
19738 2013-05-29 Richard Biener <rguenther@suse.de>
19740 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
19741 computing scalar cost offsetted by stmts that are kept live
19743 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
19744 for computation of scalar cost.
19746 2013-05-28 Steve Ellcey <sellcey@mips.com>
19748 * config/mips/mips-cpus.def (mips32r2): Change processor type.
19750 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
19752 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
19753 array notation built-in reduction functions.
19754 * doc/passes.texi (Passes): Added documentation about changes done
19756 * doc/invoke.texi (C Dialect Options): Added documentation about
19757 the -fcilkplus flag.
19758 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
19759 (BUILTINS_DEF): Depend on cilkplus.def.
19760 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
19761 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
19762 * cilkplus.def: New file.
19764 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
19766 PR rtl-optimization/57439
19767 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
19769 2013-05-28 Easwaran Raman <eraman@google.com>
19771 PR tree-optimization/57337
19772 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
19773 (find_insert_point): Correctly identify the insertion point
19774 when two statements with the same UID is compared.
19776 2013-05-28 Richard Biener <rguenther@suse.de>
19778 PR tree-optimization/56787
19779 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
19780 from the list of data references.
19781 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
19783 (vect_analyze_loop_operations): Likewise.
19784 (vect_transform_loop): Remove clobbers.
19786 2013-05-28 Martin Jambor <mjambor@suse.cz>
19788 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
19789 and REALPART_EXPRs have scalar type.
19791 2013-05-28 Richard Biener <rguenther@suse.de>
19793 PR tree-optimization/57411
19794 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
19796 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
19797 virtual operand propagation.
19799 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19801 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
19802 destination register for bmasksi_vis.
19803 (vector_init_bshuffle): Likewise.
19804 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
19806 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19808 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
19809 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
19810 mode if the instruction isn't available in the original mode.
19811 * config/sparc/sparc.opt (mfix-ut699): New option.
19812 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
19813 (divdf3): Turn into expander.
19814 (divdf3_nofix): New insn.
19815 (divdf3_fix): Likewise.
19816 (divsf3): Disable if -mfix-ut699.
19817 (sqrtdf2): Turn into expander.
19818 (sqrtdf2_nofix): New insn.
19819 (sqrtdf2_fix): Likewise.
19820 (sqrtsf2): Disable if -mfix-ut699.
19822 2013-05-27 Richard Biener <rguenther@suse.de>
19824 PR middle-end/57412
19825 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
19826 block for the new loop.
19828 2013-05-27 Richard Biener <rguenther@suse.de>
19830 PR tree-optimization/57343
19831 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
19832 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
19833 (number_of_iterations_cond): Do not build the folded tree.
19835 2013-05-27 Richard Biener <rguenther@suse.de>
19838 PR middle-end/57381
19839 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19840 OEP_CONSTANT_ADDRESS_OF retained.
19842 PR tree-optimization/57417
19843 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
19844 for unchanged base.
19845 (set_ssa_val_to): Compare addresses using
19846 get_addr_base_and_unit_offset.
19848 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19850 PR rtl-optimization/56833
19851 * postreload.c (move2add_record_mode): New function.
19852 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19853 (move2add_use_add2_insn): Use move2add_record_sym_value.
19854 (move2add_use_add3_insn): Likewise.
19855 (reload_cse_move2add): Use move2add_valid_value_p and
19856 move2add_record_mode. Invalidate call-clobbered and REG_INC
19857 affected regs by setting reg_mode to VOIDmode.
19858 (move2add_note_store): Don't pretend the inside of a SUBREG is
19859 the actual destination. Invalidate single/leading registers by
19860 setting reg_mode to VOIDmode.
19861 Use move2add_record_sym_value, move2add_valid_value_p and
19862 move2add_record_mode.
19864 2013-05-27 Richard Biener <rguenther@suse.de>
19866 PR tree-optimization/57396
19867 * tree-affine.c (double_int_constant_multiple_p): Properly
19868 return false for val == 0 and div != 0.
19870 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19872 * config/mips/mips.h: Use #elif in preprocessor conditions.
19874 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19877 * config/mips/constraints.md (kl): New constraint.
19878 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19879 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19880 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19881 from LO for MIPS16.
19882 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19884 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19887 * config/mips/mips.c (mips_can_inline_p): New function.
19888 (TARGET_CAN_INLINE_P): Define.
19890 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19892 * sched-int.h (ds_t, dw_t): Make unsigned int.
19893 Fix documentation that describes how all the ds_t bits are used.
19894 Reserve the last bit for delayed-branch scheduling.
19895 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19896 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19897 (gen_dep_weak_1): Remove prototype.
19898 * sched-deps.c (get_dep_weak_1): Make static.
19899 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19900 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19901 * doc/tm.texi: Regenerate.
19902 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19904 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19907 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19909 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19910 Sandra Loosemore <sandra@codesourcery.com>
19912 * config.gcc (powerpc-*): Allow native for with-cpu.
19914 2013-05-24 Jeff Law <law@redhat.com>
19916 PR tree-optimization/57124
19917 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19918 conversion feeding a condition if the range has an overflow
19919 if -fstrict-overflow. Add warnings for when we do make the
19922 2013-05-24 Dehao Chen <dehao@google.com>
19924 * tree-cfg.c (locus_discrim_map): Fix the typo.
19925 (locus_discrim_hasher): Likewise.
19926 (locus_discrim_hasher::hash): Likewise.
19927 (locus_discrim_hasher::equal): Likewise.
19929 2013-05-24 Martin Jambor <mjambor@suse.cz>
19931 PR tree-optimization/57294
19932 * cgraph.h (ipa_record_stmt_references): Declare.
19933 * cgraphbuild.c (ipa_record_stmt_references): New function.
19934 (build_cgraph_edges): Use ipa_record_stmt_references.
19935 (rebuild_cgraph_edges): Likewise.
19936 (cgraph_rebuild_references): Likewise.
19937 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19938 associated with the old statement and build references from the
19939 newly built statements.
19940 * ipa-ref.c (ipa_remove_stmt_references): New function.
19941 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19943 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19945 * lra-constraints.c (emit_spill_move): Use smaller mode for
19947 (check_and_process_move): Consider mem-reg moves for secondary
19949 (curr_insn_transform): Don't lose insns emitted before for
19950 secondary memory moves.
19951 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19952 reg set up in the current insn.
19954 2013-05-24 Dehao Chen <dehao@google.com>
19956 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19958 (locus_descrim_hasher::equal): Likewise.
19959 (build_gimple_cfg): New discriminator assignment algorithm.
19960 (make_edges): Likewise.
19961 (next_discriminator_for_locus): Likewise.
19962 (same_line_p): Likewise.
19963 (assign_discriminators): Likewise.
19964 (make_cond_expr_edges): Likewise.
19965 (make_gimple_switch_edges): Likewise.
19966 (make_goto_expr_edges): Likewise.
19967 (make_gimple_asm_edges): Likewise.
19969 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19971 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19972 X format specifier to only display bottom 16 bits.
19973 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19974 immediate to match for operand 2, since it will be masked.
19976 2013-05-24 Richard Biener <rguenther@suse.de>
19978 PR tree-optimization/57287
19979 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19980 all SSA names that occur in abnormal PHIs.
19982 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19985 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19986 that index is not negative.
19988 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19990 PR rtl-optimization/55177
19991 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19992 (simplify_byte_swapping_operation): New.
19993 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
19994 (simplify_relational_operation_1): Deal with BSWAP.
19996 2013-05-23 Richard Henderson <rth@redhat.com>
19999 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
20000 (ix86_reorg): Call it.
20002 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
20005 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
20006 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
20007 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
20009 2013-05-23 Christian Bruel <christian.bruel@st.com>
20012 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
20014 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
20015 Marcus Shawcroft <marcus.shawcroft@arm.com>
20017 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
20018 * config/aarch64/constraints.md (Usa): Remove.
20019 * doc/md.texi (AArch64 Usa): Remove.
20021 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
20022 Marcus Shawcroft <marcus.shawcroft@arm.com>
20024 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
20025 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
20026 * config/aarch64/predicates.md (aarch64_const_address): Remove.
20027 (aarch64_mov_operand): Use aarch64_mov_operand_p.
20029 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
20031 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
20032 instruction (AdvSIMD).
20033 * config/aarch64/aarch64-builtins.c
20034 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
20035 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
20037 2013-05-23 Martin Jambor <mjambor@suse.cz>
20039 PR middle-end/57347
20040 * tree.h (contains_bitfld_component_ref_p): Declare.
20041 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
20042 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
20044 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
20045 not access a bit-field. Assert all final offsets are byte-aligned.
20047 2013-05-23 Richard Biener <rguenther@suse.de>
20049 PR tree-optimization/57380
20050 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
20051 least one invariant or re-used load.
20052 * passes.c (init_optimization_passes): Move pass_phiprop before
20055 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
20057 * config/aarch64/aarch64-simd.md
20058 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
20060 2013-05-23 Richard Biener <rguenther@suse.de>
20062 PR middle-end/57381
20063 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
20064 OEP_CONSTANT_ADDRESS_OF retained.
20066 2013-05-23 Jakub Jelinek <jakub@redhat.com>
20068 PR middle-end/57344
20069 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
20070 don't lower unit. Handle unit not being always BITS_PER_WORD.
20072 2013-05-23 Richard Biener <rguenther@suse.de>
20074 PR rtl-optimization/57341
20075 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
20076 instead of true_dependence.
20078 2013-05-22 David Malcolm <dmalcolm@redhat.com>
20080 * bb-reorder.c (branch_threshold): Make const.
20081 (exec_threshold): Ditto.
20083 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
20084 Pat Haugen <pthaugen@us.ibm.com>
20085 Peter Bergner <bergner@vnet.ibm.com>
20087 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
20088 documentation for the power8 crypto builtins.
20090 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
20092 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
20093 macros for defining power8 builtin functions.
20094 (BU_P8V_AV_2): Likewise.
20095 (BU_P8V_AV_P): Likewise.
20096 (BU_P8V_VSX_1): Likewise.
20097 (BU_P8V_OVERLOAD_1): Likewise.
20098 (BU_P8V_OVERLOAD_2): Likewise.
20099 (BU_CRYPTO_1): Likewise.
20100 (BU_CRYPTO_2): Likewise.
20101 (BU_CRYPTO_3): Likewise.
20102 (BU_CRYPTO_OVERLOAD_1): Likewise.
20103 (BU_CRYPTO_OVERLOAD_2): Likewise.
20104 (XSCVSPDP): Fix typo, point to the correct instruction.
20105 (VCIPHER): Add power8 crypto builtins.
20106 (VCIPHERLAST): Likewise.
20107 (VNCIPHER): Likewise.
20108 (VNCIPHERLAST): Likewise.
20109 (VPMSUMB): Likewise.
20110 (VPMSUMH): Likewise.
20111 (VPMSUMW): Likewise.
20112 (VPERMXOR_V2DI): Likewise.
20113 (VPERMXOR_V4SI: Likewise.
20114 (VPERMXOR_V8HI: Likewise.
20115 (VPERMXOR_V16QI: Likewise.
20116 (VSHASIGMAW): Likewise.
20117 (VSHASIGMAD): Likewise.
20118 (VPMSUM): Likewise.
20119 (VPERMXOR): Likewise.
20120 (VSHASIGMA): Likewise.
20122 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
20123 __CRYPTO__ if the crypto instructions are available.
20124 (altivec_overloaded_builtins): Add support for overloaded power8
20127 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
20128 support for power8 crypto builtins.
20129 (builtin_function_type): Likewise.
20130 (altivec_init_builtins): Add support for builtins that take vector
20131 long long (V2DI) arguments.
20133 * config/rs6000/crypto.md: New file, define power8 crypto
20136 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
20137 Pat Haugen <pthaugen@us.ibm.com>
20138 Peter Bergner <bergner@vnet.ibm.com>
20140 * doc/invoke.texi (Option Summary): Add power8 options.
20141 (RS/6000 and PowerPC Options): Likewise.
20143 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
20144 constraints.md instead of rs6000.h. Reorder w* constraints. Add
20145 wm, wn, wr documentation.
20147 * config/rs6000/constraints.md (wm): New constraint for VSX
20148 registers if direct move instructions are enabled.
20149 (wn): New constraint for no registers.
20150 (wq): New constraint for quad word even GPR registers.
20151 (wr): New constraint if 64-bit instructions are enabled.
20152 (wv): New constraint if power8 vector instructions are enabled.
20153 (wQ): New constraint for quad word memory locations.
20155 * config/rs6000/predicates.md (const_0_to_15_operand): New
20156 constraint for 0..15 for crypto instructions.
20157 (gpc_reg_operand): If VSX allow registers in VSX registers as well
20158 as GPR and floating point registers.
20159 (int_reg_operand): New predicate to match only GPR registers.
20160 (base_reg_operand): New predicate to match base registers.
20161 (quad_int_reg_operand): New predicate to match even GPR registers
20162 for quad memory operations.
20163 (vsx_reg_or_cint_operand): New predicate to allow vector logical
20164 operations in both GPR and VSX registers.
20165 (quad_memory_operand): New predicate for quad memory operations.
20166 (reg_or_indexed_operand): New predicate for direct move support.
20168 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
20169 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
20170 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
20171 (POWERPC_MASKS): Add power8 options.
20172 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
20175 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
20176 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
20178 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
20179 (-mpower8-fusion): New power8 options.
20180 (-mpower8-fusion-sign): Likewise.
20181 (-mpower8-vector): Likewise.
20182 (-mcrypto): Likewise.
20183 (-mdirect-move): Likewise.
20184 (-mquad-memory): Likewise.
20186 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
20187 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
20189 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
20190 (rs6000_debug_vector_unit): Add p8_vector.
20191 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
20192 definitions. Also print fusion state.
20193 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
20194 (rs6000_builtin_mask_calculate): Add power8 builtin support.
20195 (rs6000_option_override_internal): Add support for power8.
20196 (rs6000_common_init_builtins): Add debugging for skipped builtins
20197 if -mdebug=builtin.
20198 (rs6000_adjust_cost): Add power8 support.
20199 (rs6000_issue_rate): Likewise.
20200 (insn_must_be_first_in_group): Likewise.
20201 (insn_must_be_last_in_group): Likewise.
20202 (force_new_group): Likewise.
20203 (rs6000_register_move_cost): Likewise.
20204 (rs6000_opt_masks): Likewise.
20206 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
20207 power8 capable assembler, default to power7 options.
20208 (TARGET_DIRECT_MOVE): Likewise.
20209 (TARGET_CRYPTO): Likewise.
20210 (TARGET_P8_VECTOR): Likewise.
20211 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
20212 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
20213 (VECTOR_MEM_P8_VECTOR_P): Likewise.
20214 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
20215 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
20216 (TARGET_XSCVDPSPN): Likewise.
20217 (TARGET_XSCVSPDPN): Likewsie.
20218 (TARGET_SYNC_HI_QI): Likewise.
20219 (TARGET_SYNC_TI): Likewise.
20220 (MASK_CRYPTO): Likewise.
20221 (MASK_DIRECT_MOVE): Likewise.
20222 (MASK_P8_FUSION): Likewise.
20223 (MASK_P8_VECTOR): Likewise.
20224 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
20225 temporary used by some of the direct move instructions to get two FP
20226 temporary registers does not force creation of a stack frame.
20227 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
20228 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
20229 that any VSX registers are tieable, even if they are also an
20230 Altivec vector mode.
20231 (r6000_reg_class_enum): Add wm, wr, wv constraints.
20232 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
20233 (RS6000_BTM_CRYPTO): Likewise.
20234 (RS6000_BTM_COMMON): Likewise.
20236 * config/rs6000/rs6000.md (cpu attribute): Add power8.
20237 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
20238 (enum rs6000_vector): Add power8 vector support.
20240 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20244 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
20245 (any_sibcall_could_use_r3): this and handle indirect calls.
20246 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
20248 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20250 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
20252 2013-05-22 Richard Biener <rguenther@suse.de>
20254 PR middle-end/57349
20255 * profile.c (branch_prob): Do not split blocks that are
20256 abnormally receiving from ECF_RETURNS_TWICE functions.
20258 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
20260 * recog.c (offsettable_address_addr_space_p): Fix calculation of
20261 address mode. Move pointer mode initialization to the same place.
20263 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
20265 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
20266 while it has any effect.
20268 2013-05-21 Easwaran Raman <eraman@google.com>
20270 PR tree-optimization/57322
20271 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
20272 UID of the statement added to the BB to be 1.
20274 2013-05-21 Jakub Jelinek <jakub@redhat.com>
20276 PR tree-optimization/57331
20277 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
20278 of conversion from pointer type to integral type with integer.
20280 2013-05-21 Martin Jambor <mjambor@suse.cz>
20283 * ipa-prop.c (ipa_read_node_info): Process param_used and
20284 controlled_uses in the same order as when writing.
20286 2013-05-21 Magnus Granberg <baldrick@free.fr>
20289 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
20291 2013-05-21 Richard Biener <rguenther@suse.de>
20293 PR tree-optimization/57318
20294 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
20295 estimate stmts with side-effects as likely eliminated.
20297 2013-05-21 Richard Biener <rguenther@suse.de>
20299 PR tree-optimization/57330
20300 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
20301 preserve the call stmts fntype.
20303 2013-05-21 Richard Biener <rguenther@suse.de>
20305 PR tree-optimization/57303
20306 * tree-ssa-sink.c (statement_sink_location): Improve killing
20307 stmt detection and properly handle self-assignments.
20309 2013-05-21 Christian Bruel <christian.bruel@st.com>
20311 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
20312 spanning registers. LEAF_REG_REMAP is supported only for contiguous
20313 registers. Set register size out of the PARALLEL loop.
20315 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
20318 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
20319 (*fmasf4, *fmasf4_media): New insns.
20321 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20323 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
20324 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
20325 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
20326 (mips_idiv_insns): Update the comments to say that the returned
20327 instruction counts are in units of BASE_INSN_LENGTH.
20328 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
20329 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
20330 using 2 rather than 4 as the length of indirect MIPS16 and
20331 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
20332 length of a NOP. Don't divide MIPS16 lengths by 2.
20333 (mips16_split_long_branches): Assume a branch is long if the
20334 length is greater than 4 rather than 8.
20335 * config/mips/mips.md (length): Give MIPS16 lengths directly,
20336 rather than multiplying them by 2. Multiply instruction counts
20337 by BASE_INSN_LENGTH rather than 4.
20338 (*jump_mips16, tls_get_tp_mips16_<mode>)
20339 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
20341 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20343 * config/mips/mips.md (extended_mips16): Remove branch case.
20344 (length): Remove duplicated extended_mips16 test.
20346 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20348 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
20350 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
20352 * recog.h (Recog_data): Rename to...
20353 (recog_data_d): ...this.
20354 (recog_data): Update accordingly.
20355 * recog.c (recog_data): Likewise.
20356 * reload.c (save_recog_data): Likewise.
20357 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
20358 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
20360 2013-05-17 Julian Brown <julian@codesourcery.com>
20362 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
20363 found in a REG_EQUAL note, invalidate it.
20365 2013-05-17 Easwaran Raman <eraman@google.com>
20367 * tree-ssa-reassoc.c (find_insert_point): New function.
20368 (insert_stmt_after): Likewise.
20369 (get_def_stmt): Likewise.
20370 (ensure_ops_are_available): Likewise.
20371 (not_dominated_by): Likewise.
20372 (rewrite_expr_tree): Do not move statements beyond what is
20373 necessary. Remove call to swap_ops_for_binary_stmt...
20374 (reassociate_bb): ... and move it here.
20375 (build_and_add_sum): Assign UIDs for new statements.
20376 (linearize_expr): Likewise.
20377 (do_reassoc): Renumber gimple statement UIDs.
20379 2013-05-17 Jan Hubicka <jh@suse.cz>
20381 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
20383 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
20384 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
20386 (output_weakrefs): Update.
20388 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
20389 Martin Jambor <mjambor@suse.cz>
20391 PR middle-end/57276
20392 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
20393 value that corresponds to the given aggval is found in values vector.
20395 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
20397 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
20398 sse, sse2, sse3, ssse3 and sse4a flags to options.
20400 2013-05-17 David Malcolm <dmalcolm@redhat.com>
20402 * gengtype-state.c: (s_expr_writer): New class, to handle
20403 prettifying of output layout of s-expressions.
20404 (state_writer): New class, to write out gtype.state.
20405 (state_written_type_count): Move this variable into member data of
20407 (s_expr_writer::s_expr_writer): New code: constructor for new class
20408 (state_writer::state_writer(): ditto
20409 (s_expr_writer::write_new_line): New function
20410 (s_expr_writer::write_any_indent): ditto
20411 (s_expr_writer::begin_s_expr): ditto
20412 (s_expr_writer::end_s_expr): ditto
20413 (write_state_fileloc): convert to method of state_writer...
20414 (state_writer:: write_state_fileloc): ...and use methods of
20415 s_expr_writer to write indentation into the gtype.state output file
20416 to visually represent the hierarchical structure of the list
20418 (write_state_fields): ditto, renaming to...
20419 (state_writer::write_state_fields)
20420 (write_state_a_string): ditto, renaming to...
20421 (state_writer::write_state_a_string)
20422 (write_state_string_option): ditto, renaming to...
20423 (state_writer::write_state_string_option)
20424 (write_state_type_option): ditto, renaming to...
20425 (state_writer::write_state_type_option)
20426 (write_state_nested_option): ditto, renaming to...
20427 (state_writer::write_state_nested_option)
20428 (write_state_option): ditto, renaming to...
20429 (state_writer::write_state_option)
20430 (write_state_options): ditto, renaming to...
20431 (state_writer::write_state_options)
20432 (write_state_lang_bitmap): ditto, renaming to...
20433 (state_writer::write_state_lang_bitmap)
20434 (write_state_version): ditto, renaming to...
20435 (state_writer::write_state_version)
20436 (write_state_scalar_type): ditto, renaming to...
20437 (state_writer::write_state_scalar_type)
20438 (write_state_string_type): ditto, renaming to...
20439 (state_writer::write_state_string_type)
20440 (write_state_undefined_type): ditto, renaming to...
20441 (state_writer::write_state_undefined_type)
20442 (write_state_struct_union_type): ditto, renaming to...
20443 (state_writer::write_state_struct_union_type)
20444 (write_state_struct_type): ditto, renaming to...
20445 (state_writer::write_state_struct_type)
20446 (write_state_user_struct_type): ditto, renaming to...
20447 (state_writer::write_state_user_struct_type)
20448 (write_state_lang_struct_type): ditto, renaming to...
20449 (state_writer::write_state_lang_struct_type)
20450 (write_state_param_struct_type): ditto, renaming to...
20451 (state_writer::write_state_param_struct_type)
20452 (write_state_pointer_type): ditto, renaming to...
20453 (state_writer::write_state_pointer_type)
20454 (write_state_array_type): ditto, renaming to...
20455 (state_writer::write_state_array_type)
20456 (write_state_gc_used): ditto, renaming to...
20457 (state_writer::write_state_gc_used)
20458 (write_state_common_type_content): ditto, renaming to...
20459 (state_writer::write_state_common_type_content)
20460 (write_state_type): ditto, renaming to...
20461 (state_writer::write_state_type)
20462 (write_state_pair_list): ditto, renaming to...
20463 (state_writer::write_state_pair_list)
20464 (write_state_pair): ditto, renaming to...
20465 (state_writer::write_state_pair)
20466 (write_state_typedefs): ditto, renaming to...
20467 (state_writer::write_state_typedefs)
20468 (write_state_structures): ditto, renaming to...
20469 (state_writer::write_state_structures)
20470 (write_state_param_structs): ditto, renaming to...
20471 (state_writer::write_state_param_structs)
20472 (write_state_variables): ditto, renaming to...
20473 (state_writer::write_state_variables)
20474 (write_state_srcdir): ditto, renaming to...
20475 (state_writer::write_state_srcdir)
20476 (write_state_files_list): ditto, renaming to...
20477 (state_writer::write_state_files_list)
20478 (write_state_languages): ditto, renaming to...
20479 (state_writer::write_state_languages)
20480 (write_state): create a state_writer instance and use it when
20481 writing out the state file
20483 2013-05-17 Mike Stump <mikestump@comcast.net>
20485 PR rtl-optimization/57304
20486 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
20487 accessing DF_REF_REAL_LOC.
20489 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20491 PR rtl-optimization/57281
20492 PR rtl-optimization/57300
20493 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
20494 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
20495 what the other splitter did if the registers are dead.
20497 2013-05-17 Richard Biener <rguenther@suse.de>
20499 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
20502 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20504 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
20507 2013-05-17 Marek Polacek <polacek@redhat.com>
20509 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
20510 length when doing non-zero store of storing '\0' to '\0'.
20512 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20514 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
20515 vect_external_def oprnd1 with loop_vinfo, try to emit
20516 optional cast, negation and and stmts on the loop preheader
20517 edge instead of into the pattern def seq.
20519 PR tree-optimization/57051
20520 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
20521 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
20523 2013-05-16 Nick Clifton <nickc@redhat.com>
20525 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
20526 (rl78_is_naked_func): New function.
20527 (rl78_expand_prologue): Skip prologue generation for naked functions.
20528 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
20529 * doc/extend.texi (naked): Add RL78 to the list of processors
20530 that supports this attribute.
20532 2013-05-16 Jeff Law <law@redhat.com>
20534 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
20536 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20538 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
20539 cache parameters using detect_caches_amd also for CYRIX,
20540 NSC and TM2 signatures.
20542 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20543 Dzianis Kahanovich <mahatma@eu.by>
20547 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
20548 VIA/Centaur processors and determine their cache parameters
20549 using detect_caches_amd.
20551 2013-05-16 Teresa Johnson <tejohnson@google.com>
20553 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
20554 (rtl_verify_edges): New function.
20555 (rtl_verify_bb_insns): Ditto.
20556 (rtl_verify_bb_pointers): Ditto.
20557 (rtl_verify_bb_insn_chain): Ditto.
20558 (rtl_verify_fallthru): Ditto.
20559 (rtl_verify_bb_layout): Ditto.
20560 (rtl_verify_flow_info_1): Outline checks into new functions.
20561 (rtl_verify_flow_info): Ditto.
20563 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
20565 * cfghooks.c (copy_bbs): Add update_dominance argument.
20566 * cfghooks.h (copy_bbs): Update prototype.
20567 * tree-cfg.c (gimple_duplicate_sese_region):
20568 Add update_dominance argument.
20569 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
20570 * tree-ssa-loop-ch.c (copy_loop_headers): Update
20571 gimple_duplicate_sese_region call.
20572 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
20573 Update copy_bbs call.
20574 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
20575 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
20577 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20579 * tree-vectorizer.h (NUM_PATTERNS): Increment.
20580 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
20581 vect_recog_rotate_pattern.
20582 (vect_recog_rotate_pattern): New function.
20584 2013-05-16 Jason Merrill <jason@redhat.com>
20586 * Makefile.in (LLINKER): New variable.
20587 (mostlyclean): Remove link mutex.
20588 * configure.ac: Handle --enable-link-mutex.
20589 * lock-and-run.sh: New script.
20591 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20594 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
20597 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20599 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
20601 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
20603 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
20604 * config/arm/arm.c (next_consecutive_mem): New function.
20605 (gen_movmem_ldrd_strd): Likewise.
20606 * config/arm/arm.md (movmemqi): Update condition and code.
20607 (unaligned_loaddi, unaligned_storedi): New patterns.
20609 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20611 * config.gcc: Obsolete *-*-solaris2.9*.
20612 * doc/install.texi (Specific, *-*-solaris2*): Document it.
20614 2013-05-16 Richard Biener <rguenther@suse.de>
20616 * passes.c (init_optimization_passes): Move pass_parallelize_loops
20617 earlier, after GRAPHITE transforms and IV canonicalization.
20619 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20621 * omp-low.c (extract_omp_for_data): For collapsed loops,
20622 if at least one of the loops is known at compile time to
20623 iterate zero times, set count to 0.
20624 (expand_omp_regimplify_p): New function.
20625 (expand_omp_for_generic): For collapsed loops, if at least
20626 one of the loops isn't known to iterate at least once,
20627 add runtime check with setting count to 0.
20628 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
20629 For unsigned types if it isn't known at compile time that
20630 the loop will iterate at least once, add runtime check to bypass
20631 the whole loop if initial condition isn't true.
20633 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
20635 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
20637 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
20639 PR middle-end/57286
20640 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
20641 transformations to avoid an infinite loop.
20643 2013-05-16 Marek Polacek <polacek@redhat.com>
20645 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
20647 2013-05-15 Leif Ekblad <leif@rdos.net>
20649 * config/i386/i386.c (ix86_decompose_address): Use
20650 DEFAULT_TLS_SEG_REG to access TLS segment register.
20651 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
20652 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
20653 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
20655 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
20658 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
20659 sibling calls to functions that would normally be lazily bound,
20660 unless $gp is call-clobbered.
20662 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
20664 * config/i386/i386.c (ix86_option_override_internal): Update
20665 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
20666 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
20667 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
20668 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
20670 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
20672 2013-05-15 Andreas Schwab <schwab@suse.de>
20674 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
20675 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
20678 2013-05-15 Teresa Johnson <tejohnson@google.com>
20680 * loop-unroll.c (report_unroll_peel): Check decision before
20681 emitting unroll/peel message.
20683 2013-05-15 Teresa Johnson <tejohnson@google.com>
20685 * function.h (has_bb_partition): New rtl_data flag.
20686 (bb_reorder_complete): Ditto.
20687 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
20688 instead of flag_reorder_blocks_and_partition.
20689 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
20690 with some enhancements.
20691 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
20692 * bb-reorder.c (connect_traces): Check for has_bb_partition
20693 instead of flag_reorder_blocks_and_partition.
20694 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
20695 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
20696 verify_hot_cold_block_grouping.
20697 (partition_hot_cold_basic_blocks): Set has_bb_partition.
20699 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20702 * config/arm/predicates.md (call_insn_operand): New predicate.
20703 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
20704 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
20705 if insn is not a tail call.
20706 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
20708 * config/arm/arm.h (enum reg_class): New caller save register class.
20709 (REG_CLASS_NAMES): Likewise.
20710 (REG_CLASS_CONTENTS): Likewise.
20711 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
20714 2013-05-15 Richard Biener <rguenther@suse.de>
20716 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
20717 of MSG_OPTIMIZED_LOCATIONS.
20718 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
20719 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
20721 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
20722 of MSG_OPTIMIZED_LOCATIONS.
20723 (execute_vect_slp): Likewise.
20724 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
20725 (vect_create_cond_for_alias_checks): Likewise.
20726 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
20727 (vect_recog_widen_mult_pattern): Likewise.
20728 (vect_recog_widen_sum_pattern): Likewise.
20729 (vect_recog_over_widening_pattern): Likewise.
20730 (vect_recog_widen_shift_pattern): Likewise.
20731 (vect_recog_vector_vector_shift_pattern): Likewise.
20732 (vect_recog_divmod_pattern): Likewise.
20733 (vect_recog_mixed_size_cond_pattern): Likewise.
20734 (vect_recog_bool_pattern): Likewise.
20735 (vect_pattern_recog_1): Likewise.
20737 2013-05-15 Martin Jambor <mjambor@suse.cz>
20739 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
20740 non-functions to builtin_unreachable.
20741 * ipa-inline-transform.c (inline_call): Do not assert estimates were
20742 correct when new direct edges were discovered.
20744 2013-05-15 Martin Jambor <mjambor@suse.cz>
20746 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
20747 header, print symbol order instead of node uid, print more information
20748 about indirect edge targets.
20749 (ipa_make_edge_direct_to_target): Print symbol order instead of node
20751 (ipa_make_edge_direct_to_target): Likewise.
20752 (remove_described_reference): Likewise.
20753 (propagate_controlled_uses): Likewise.
20754 (ipa_print_node_params): Also print symbol order.
20755 (ipcp_transform_function): Print symbol order instead of node uids.
20756 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
20757 (cgraph_get_create_real_symbol_node): Likewise.
20758 * ipa-cp.c (print_lattice): Likewise.
20759 (print_all_lattices): Likewise.
20760 (determine_versionability): Likewise.
20761 (initialize_node_lattices): Likewise.
20762 (estimate_local_effects): Likewise.
20763 (update_profiling_info): Likewise.
20764 (create_specialized_node): Likewise.
20765 (perhaps_add_new_callers): Likewise.
20766 (decide_about_value): Likewise.
20767 (decide_whether_version_node): Likewise.
20768 (identify_dead_nodes): Likewise.
20769 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
20770 (dump_inline_summary): Likewise.
20771 (estimate_node_size_and_time): Likewise.
20772 (inline_analyze_function): Likewise.
20773 * ipa-inline.c (report_inline_failed_reason): Likewise.
20774 (want_early_inline_function_p): Likewise.
20775 (edge_badness): Likewise.
20776 (update_edge_key): Likewise.
20777 (inline_small_functions): Likewise. Add dumping of order to two other
20779 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
20780 instead of node uids.
20781 (propagate_pure_const): Likewise.
20782 (propagate_pure_const): Likewise.
20783 * ipa-utils.c (dump_cgraph_node_set): Likewise.
20784 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
20785 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
20787 * tree-pretty-print.c (dump_function_header): Likewise.
20788 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
20789 Print symbol order instead of node uids.
20791 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20793 * config/s390/s390.c (s390_register_move_cost): Don't impose the
20794 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
20796 2013-05-15 Richard Biener <rguenther@suse.de>
20798 PR tree-optimization/57275
20799 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
20800 return value for fail to do runtime alias checks for gather loads.
20802 2013-05-15 Jan Hubicka <jh@suse.cz>
20806 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
20807 weakrefs are not external.
20808 (lto_symtab_merge_decls): Fix thinko when dealing with
20809 non-lto_symtab decls.
20810 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
20811 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
20812 * varpool.c (dump_varpool_node): Dump more flags.
20814 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20816 * config/i386/i386.c (processor_alias_table): Add instruction
20817 FSGSBASE for AMD bdver3 architecture.
20819 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20821 * tree.c (warn_deprecated_use): Print file:line using locus color.
20822 * diagnostic.c (diagnostic_report_current_module): Print file:line
20823 and file:line:column using locus color.
20825 2013-05-14 Mike Stump <mikestump@comcast.net>
20827 * gdbinit.in: Add __null.
20829 2013-05-14 Mike Stump <mikestump@comcast.net>
20831 * recog.h: Rename struct recog_data to Recog_data.
20832 * recog.c: Likewise.
20833 * reload.c (can_reload_into): Likewise.
20834 * config/picochip/picochip.c: Likewise.
20836 2013-05-14 Mike Stump <mikestump@comcast.net>
20838 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
20840 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
20842 * resource.h (struct resources): Remove unch_memory member.
20843 (CLEAR_RESOURCE): Don't clear unch_memory.
20844 * resource.c (mark_referenced_resources): Don't set it.
20845 (mark_set_resources): Likewise.
20846 (mark_target_live_regs): Don't clear it.
20847 (init_resource_info): Likewise.
20848 * reorg.c (resource_conflicts_p): Don't compare it.
20849 (redundant_insn): Don't set it.
20851 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20853 * emit-rtl.c (next_label): Remove unused function.
20854 (skip_consecutive_labels, link_cc0_insns): Move to ...
20855 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20856 only place where these functions are used, and make them static.
20858 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20860 * fold-const.c (fold_negate_expr): Handle vectors.
20861 (fold_truth_not_expr): Make it static.
20862 (fold_invert_truthvalue): New static function.
20863 (invert_truthvalue_loc): Handle vectors. Do not call
20864 fold_truth_not_expr directly.
20865 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20866 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20867 (fold_comparison): Handle vector constants.
20868 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20869 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20870 * tree.h (fold_truth_not_expr): Remove declaration.
20872 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20874 * config/aarch64/aarch64-simd.md
20875 (aarch64_vcond_internal<mode>): Rename to...
20876 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20877 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20878 float modes. Clarify all iterator modes.
20879 (vcond<mode><mode>): Use new name for vcond expanders.
20880 (vcond<v_cmp_result><mode>): Likewise.
20881 (vcondu<mode><mode>: Likewise.
20882 * config/aarch64/iterators.md (VDQF_COND): New.
20884 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20887 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20888 variable for the shift amount. Check that we shift by non-negative
20891 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20894 * config/arm/arm.h (EPILOGUE_USES): Only return true
20895 for LR_REGNUM after epilogue_completed.
20897 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20899 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20900 is error_mark_node.
20902 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20905 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20906 and Solaris 11+/x86 with gld.
20907 * configure: Regenerate.
20909 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20911 * expmed.c (expand_shift_1): Canonicalize rotates by
20912 constant bitsize / 2 to bitsize - 1.
20913 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20914 case ROTATERT>: Likewise.
20917 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20919 * config/i386/i386.md (rotateinv): New code attr.
20920 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20921 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20922 roll $31, %eax, etc.
20924 2013-05-14 Richard Biener <rguenther@suse.de>
20926 PR middle-end/57235
20927 * tree-eh.c (sink_clobbers): Give up for successors with
20928 multiple predecessors and no virtual uses.
20930 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20932 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20933 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20935 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20937 PR middle-end/57251
20938 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20939 the case when both op0 and op1 have VOIDmode.
20941 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20943 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20944 in multiply-accumulate mode.
20946 2013-05-13 Guozhi Wei <carrot@google.com>
20948 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20950 2013-05-13 Kai Tietz <ktietz@redhat.com>
20953 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20954 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20955 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20956 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20957 (ix86_expand_prologue): Likewise.
20958 (ix86_expand_split_stack_prologue): Likewise.
20959 (legitimate_pic_address_disp_p): Likewise.
20960 (legitimize_pic_address): Likewise.
20961 (legitimize_tls_address): Likewise.
20962 (legitimize_pe_coff_symbol): Likewise.
20963 (output_pic_addr_const): Likewise.
20964 (construct_plt_address): Likewise.
20965 (ix86_expand_call): Likewise.
20966 (x86_output_mi_thunk): Likewise.
20967 (x86_function_profiler): Likewise.
20969 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20971 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20972 similar switch cases.
20973 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20974 (aarch64_simd_mov_to_<mode>low): Delete.
20975 (aarch64_simd_mov_to_<mode>high): Delete.
20976 (move_lo_quad_<mode>): Add w<-r alternative.
20977 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20978 (aarch64_simd_mov_from_*): Update type attribute.
20979 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20982 2013-05-13 Jan Hubicka <jh@suse.cz>
20984 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20985 * config/i386/i386.c (ix86_compute_frame_layout,
20986 ix86_expand_epilogue, emit_i387_cw_initialization,
20987 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20988 ix86_local_alignment): Fix use of size/speed predicates.
20990 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20992 PR tree-optimization/45216
20993 PR tree-optimization/57157
20994 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
20995 the (-Y) & (B - 1) variant if OP is |.
20996 * expmed.c (expand_shift_1): For rotations by const0_rtx just
20997 return shifted. Use (-op1) & (prec - 1) as other_amount
20998 instead of prec - op1.
21000 2013-05-13 Martin Jambor <mjambor@suse.cz>
21002 PR middle-end/42371
21003 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
21004 (ipa_constant_data): New type.
21005 (ipa_jump_func): Use ipa_constant_data to hold information about
21006 constant jump functions.
21007 (ipa_get_jf_constant): Adjust to jump function type changes.
21008 (ipa_get_jf_constant_rdesc): New function.
21009 (ipa_param_descriptor): New field controlled_uses.
21010 (ipa_get_controlled_uses): New function.
21011 (ipa_set_controlled_uses): Likewise.
21012 * ipa-ref.h (ipa_find_reference): Declare.
21013 * ipa-prop.c (ipa_cst_ref_desc): New type.
21014 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
21016 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
21017 New parameter cs. Adjust all callers.
21018 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
21019 (remove_described_reference): New function.
21020 (jfunc_rdesc_usable): Likewise.
21021 (try_make_edge_direct_simple_call): Decrement controlled use count,
21022 attempt to remove reference if it hits zero.
21023 (combine_controlled_uses_counters): New function.
21024 (propagate_controlled_uses): Likewise.
21025 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
21026 (ipa_edge_duplication_hook): Duplicate reference descriptions.
21027 (ipa_print_node_params): Print described use counter.
21028 (ipa_write_jump_function): Adjust to jump function type changes.
21029 (ipa_read_jump_function): New parameter CS, pass it to
21030 ipa_set_jf_constant. Adjust caller.
21031 (ipa_write_node_info): Stream controlled use count
21032 (ipa_read_node_info): Likewise.
21033 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
21035 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
21036 count. Remove cloning-added reference if it reaches zero.
21037 * ipa-ref.c (ipa_find_reference): New function.
21039 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
21041 * config/i386/i386.c (processor_target_table): Modified default
21042 alignment values for AMD BD and BT architectures.
21044 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
21046 * tree-vect-generic.c (uniform_vector_p): Move ...
21047 * tree.c (uniform_vector_p): ... here.
21048 * tree.h (uniform_vector_p): Declare it.
21049 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
21052 2013-05-13 Jakub Jelinek <jakub@redhat.com>
21054 PR tree-optimization/57230
21055 * tree-ssa-strlen.c (handle_char_store): Record length for
21056 array store from STRING_CST.
21058 PR tree-optimization/57230
21059 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
21062 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
21064 * config/epiphany/epiphany.c (epiphany_init): Check size of
21065 NUM_MODES_FOR_MODE_SWITCHING.
21066 (epiphany_expand_prologue):
21067 Remove CONFIG_REGNUM initial value handling code.
21068 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
21069 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
21070 (emit_set_fp_mode, epiphany_mode_after): Likewise.
21071 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
21072 Don't return 1 for FP_MODE_NONE.
21073 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
21074 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
21075 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
21076 * config/epiphany/epiphany.md (save_config): New pattern.
21078 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
21080 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
21082 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
21084 * config/i386/i386.md (memory): Handle sseishft1.
21085 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
21086 (*vec_extractv2di_1): Ditto.
21088 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
21090 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
21093 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
21095 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
21096 Add mthumb/march=armv7-a multilib.
21097 Add mthumb/march=armv7-r multilib.
21098 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
21100 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
21102 * config/v850/t-rtems: Add more multilibs.
21104 2013-05-10 Richard Biener <rguenther@suse.de>
21106 PR tree-optimization/57214
21107 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
21108 not propagate from SSA names that occur in abnormal PHI nodes.
21110 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
21112 * stor-layout.c (element_precision): New function.
21113 * machmode.h (element_precision): Declare it.
21114 * tree.c (build_minus_one_cst): New function.
21115 (element_precision): Likewise.
21116 * tree.h (build_minus_one_cst): Declare new function.
21117 (element_precision): Likewise.
21118 * fold-const.c (operand_equal_p): Use element_precision.
21119 (fold_binary_loc): Handle vector types.
21120 * convert.c (convert_to_integer): Use element_precision.
21121 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
21124 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
21126 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
21127 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
21128 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
21129 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
21130 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
21131 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
21132 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
21133 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
21134 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
21135 (Uuw8): New constraints.
21136 (Usb4): Move into alphabetical order.
21137 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
21138 (sd8_operand, ub8_operand, uw8_operand): New predicates.
21139 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
21140 previously unnamed patterns.
21141 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
21142 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
21143 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
21144 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
21145 of set_attr_alternative/if_then_else. Use extended_mips16 instead
21146 of specific lengths.
21148 2013-05-10 Jakub Jelinek <jakub@redhat.com>
21150 * config/i386/i386.md (rotateinv): New code attr.
21151 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
21152 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
21153 roll $31, %eax, etc.
21155 PR tree-optimization/45216
21156 PR tree-optimization/57157
21157 * tree-ssa-forwprop.c (simplify_rotate): New function.
21158 (ssa_forward_propagate_and_combine): Call it.
21160 2013-05-10 Richard Biener <rguenther@suse.de>
21162 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
21163 disable peeling when we version for aliasing.
21164 (vector_alignment_reachable_p): Honor explicit user alignment.
21165 (vect_supportable_dr_alignment): Likewise.
21166 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
21167 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
21168 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
21169 then peeling to arrange for the cost-model check to come first.
21171 2013-05-10 Alan Modra <amodra@gmail.com>
21173 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
21174 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
21175 * configure: Regenerate.
21177 2013-05-10 Alan Modra <amodra@gmail.com>
21180 * varasm.c (default_elf_select_section): Move !DECL_P check..
21181 (get_named_section): ..to here before calling get_section_name.
21183 (default_section_type_flags): Add DECL_P check.
21184 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
21185 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
21187 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
21189 * config/epiphany/epiphany.c (epiphany_expand_prologue):
21190 When using gen_stack_adjust_str with a register offset, add a
21191 REG_FRAME_RELATED_EXPR note.
21193 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
21195 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
21196 (*vec_extractv4si_zext_mem): Ditto.
21197 (*vec_extractv2di): Add 0->x and x->x alternatives.
21198 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
21199 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
21201 2013-05-09 Jason Merrill <jason@redhat.com>
21203 N3639 C++1y VLA support
21204 * gimplify.c (gimplify_vla_decl): Don't touch an existing
21207 * tree.c (build_constructor_va): New.
21208 * tree.h: Declare it.
21210 2013-05-09 Martin Jambor <mjambor@suse.cz>
21213 * gimple-fold.c (canonicalize_constructor_val): Call
21214 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
21216 2013-05-09 Jan Hubicka <jh@suse.cz>
21217 Richard Biener <rguenther@suse.de>
21220 * symtab.c (symtab_make_decl_local): Do not add private names.
21222 2013-05-09 Jan Hubicka <jh@suse.cz>
21225 * symtab.c (insert_to_assembler_name_hash): Handle clones.
21226 (unlink_from_assembler_name_hash): Likewise.
21227 (symtab_prevail_in_asm_name_hash, symtab_register_node,
21228 symtab_unregister_node, symtab_initialize_asm_name_hash,
21229 change_decl_assembler_name): Update.
21231 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
21233 * config/aarch64/aarch64.md: New movtf split.
21234 (*movtf_aarch64): Update.
21235 (aarch64_movdi_tilow): Handle TF modes and rename to
21236 aarch64_movdi_<mode>low.
21237 (aarch64_movdi_tihigh): Handle TF modes and rename to
21238 aarch64_movdi_<mode>high
21239 (aarch64_movtihigh_di): Handle TF modes and rename to
21240 aarch64_mov<mode>high_di
21241 (aarch64_movtilow_di): Handle TF modes and rename to
21242 aarch64_mov<mode>low_di
21243 (aarch64_movtilow_tilow): Remove spurious whitespace.
21244 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
21246 (aarch64_print_operand): Update.
21248 2013-05-09 Alan Modra <amodra@gmail.com>
21250 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
21252 * configure: Regenerate.
21254 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
21256 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
21257 splitter preparation statements.
21258 * config/i386/sse.md (*vec_extract* splitters): Ditto.
21259 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
21262 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21264 * gimple-ssa-strength-reduction.c (count_candidates): Change
21265 return value to int.
21266 (analyze_candidates_and_replace): Change type of length to int.
21268 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
21270 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
21271 (*vec_extract<mode>): Use VI12_128 mode iterator.
21272 (*vec_extract<mode>_mem): Ditto.
21273 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
21276 2013-05-08 Diego Novillo <dnovillo@google.com>
21281 2012-08-17 Diego Novillo <dnovillo@google.com>
21284 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
21285 * config.in: Regenerate.
21286 * configure: Regenerate.
21287 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
21289 2013-05-08 Jan Hubicka <jh@suse.cz>
21292 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
21293 * cgraph.h (symtab_node_base): Add unique_name.
21294 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
21295 input_overwrite_node, input_varpool_node): Stream unique_name.
21296 * cgraphclones.c (cgraph_create_virtual_clone,
21297 cgraph_function_versioning): Set unique_name.
21298 * ipa.c (function_and_variable_visibility): Set unique_name.
21300 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21302 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
21303 (alloc_cand_and_find_basis): Restrict conditional candidate
21304 processing to CAND_MULTs.
21306 2013-05-08 Jan Hubicka <jh@suse.cz>
21309 lto-symtab.c (lto_symtab_symbol_p): New function.
21310 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
21311 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
21312 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
21313 Skip static symbols.
21315 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
21317 PR tree-optimization/57200
21318 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
21319 Only call inform if the preceding warning_at returns true.
21321 2013-05-07 Han Shen <shenhan@google.com>
21323 * cfgexpand.c (record_or_union_type_has_array_p): New function.
21324 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
21325 * common.opt (fstack-protector-strong): New option.
21326 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
21327 * doc/invoke.texi (Optimization Options): Document
21328 "-fstack-protector-strong".
21329 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
21331 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
21333 * config/mips/mips.c (mips_machine_reorg2): Return 0.
21335 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
21337 * ira.c (update_equiv_regs): Add insn having equiv memory even if
21338 it is not lhs of the insn.
21339 (setup_reg_equiv): Remove insn having equiv memory which it is not
21341 * lra-constraints.c (process_address): Try to improve generation
21342 code for address base + disp.
21343 (lra_constraints): Make correct the code for checking insn setting
21344 up backward equivalence. Remove insn only if it is in the init
21346 * lra-eliminations.c (update_reg_eliminate): Change return value.
21347 (lra_eliminate): Use the result.
21349 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
21351 * config/i386/sse.md (ssescalarnummask): New mode attribute.
21352 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
21353 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
21354 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
21355 register target operands.
21356 (*vec_extractv8hi_sse2): New pattern.
21357 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
21358 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
21359 (*vec_extract<mode>_mem): New insn and split pattern.
21361 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
21363 * config/arm/arm.c (arm_asan_shadow_offset): New function.
21364 (TARGET_ASAN_SHADOW_OFFSET): Define.
21365 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
21366 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
21368 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21370 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
21371 (incr_vec_index): Return -1 if increment not found.
21372 (create_add_on_incoming_edge): Assert if increment not found.
21373 (record_increment): Limit number of increments recorded.
21374 (all_phi_incrs_profitable): Return false if an increment not found.
21375 (replace_profitable_candidates): Don't process increments that were
21377 (analyze_candidates_and_replace): Limit size of incr_vec.
21379 2013-05-07 Richard Biener <rguenther@suse.de>
21381 * calls.c (special_function_p): setjmp-like functions are leaf.
21382 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
21383 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
21385 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
21387 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
21388 (aarch64_simd_mov<mode>): New expander.
21389 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
21390 (aarch64_simd_mov_to_<mode>high): Likewise.
21391 (aarch64_simd_mov_from_<mode>low): Likewise.
21392 (aarch64_simd_mov_from_<mode>high): Likewise.
21393 (aarch64_dup_lane<mode>): Update.
21394 (aarch64_dup_lanedi): New instruction pattern.
21395 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
21396 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
21398 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21400 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
21401 (replace_mult_candidate): Remove unnecessary argument; remove
21402 unnecessary parameter from call to introduce_cast_before_cand.
21403 (replace_unconditional_candidate): Remove unnecessary parameter
21404 from call to replace_mult_candidate.
21405 (replace_conditional_candidate): Likewise.
21406 (insert_initializers): Use make_temp_ssa_name.
21407 (introduce_cast_before_cand): Remove unnecessary argument; use
21408 make_temp_ssa_name.
21409 (replace_one_candidate): Remove unnecessary argument; remove
21410 unnecessary parameter from calls to introduce_cast_before_cand.
21411 (replace_profitable_candidates): Remove unnecessary parameters
21412 from calls to replace_one_candidate.
21414 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21416 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
21417 phi def as possibly hiding a basis for a CAND_ADD whose operands
21418 have been commuted in the analysis.
21419 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
21421 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21423 * config/aarch64/aarch64.md
21424 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
21425 shift value between 0-4.
21427 2013-05-07 Richard Biener <rguenther@suse.de>
21429 * double-int.h (rshift): New overload.
21430 * double-int.c (rshift): New function.
21431 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
21432 (create_reference_ops_from_ref): Remove.
21433 (vn_reference_insert): Use shared ops for constructing the
21434 reference and copy it.
21436 2013-05-07 Richard Biener <rguenther@suse.de>
21438 PR middle-end/57190
21439 * tree-eh.c (sink_clobbers): Properly propagate
21440 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
21442 2013-05-07 Jakub Jelinek <jakub@redhat.com>
21444 PR tree-optimization/57149
21445 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
21446 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
21447 collect_phi_def_edges, execute_late_warn_uninitialized): Use
21448 uninit_undefined_value_p instead of ssa_undefined_value_p.
21451 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
21452 for modifier == EXPAND_INITIALIZER.
21454 2013-05-07 Anton Blanchard <anton@samba.org>
21456 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
21457 for powerpc64 little endian.
21458 * configure: Regenerate.
21460 2013-05-06 Graham Stott <grahams@btinternet.com>
21462 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
21463 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
21464 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
21465 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
21467 2013-05-06 Graham Stott <grahams@btinternet.com>
21469 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
21470 codes which allow non-lvalues.
21472 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
21474 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
21475 components are all 1s.
21476 (integer_minus_onep): New function.
21477 * tree.h (integer_minus_onep): Declare it.
21478 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
21479 integer_minus_onep instead of integer_all_onesp.
21481 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21484 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
21485 variations of these patterns.
21487 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21489 * config/i386/i386.md (isa): Add x64_sse4 member.
21490 (enabled): Handle x64_sse4.
21491 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
21492 instruction for 64bit SSE4_1 targets. Update insn attributes.
21493 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
21494 instruction for SSE4_1 targets. Update insn attributes.
21495 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21496 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
21497 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
21499 (*vec_extractv4si): Rename from *sse4_1_pextrd.
21500 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
21501 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
21503 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21506 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
21508 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
21510 * final.c (do_assembler_dialects): Don't handle curly braces and
21511 vertical bar escaped by % as dialect delimiters.
21512 (output_asm_insn): Print curly braces and vertical bar if escaped
21513 by % and ASSEMBLER_DIALECT defined.
21514 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
21515 * doc/tm.texi: Regenerated.
21517 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
21519 * config/mips/mips.c: Include tree-pass.h.
21520 (mips_reorg): Split in pre- and post-dbr_schedule parts.
21521 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
21522 (pass_mips_machine_reorg2): New machine specific pass.
21523 (insert_pass_mips_machine_reorg2): New pass plugin definition.
21524 (mips_option_override): Register the new pass.
21525 * rtl.h (cleanup_barriers): Remove prototype.
21526 (dbr_schedule): Likewise.
21527 * jump.c (cleanup_barriers): Make static.
21528 * reorg.c (dbr_schedule): Likewise.
21530 2013-05-06 Richard Biener <rguenther@suse.de>
21532 PR tree-optimization/57185
21533 * tree-parloops.c (add_field_for_reduction): Handle anonymous
21534 SSA names properly.
21536 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21539 * config/i386/i386.c (add_parameter_dependencies): Add dependence
21540 between "first_arg" and "insn", not "last" and "insn".
21542 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21544 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
21545 (find_candidates_in_block): Re-enable slsr_process_phi.
21546 (create_phi_basis): Fix double counting of candidate adjustment.
21548 2013-05-06 Richard Biener <rguenther@suse.de>
21550 PR middle-end/57147
21551 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
21552 the edge is also fallthru, preserve it and just clear the
21554 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
21555 also complex, preserve that and just clear the fallthru flag.
21556 * tree-inline.c (update_ssa_across_abnormal_edges): Also
21557 update virtual operands.
21559 2013-05-06 Alan Modra <amodra@gmail.com>
21561 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
21562 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
21563 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
21564 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
21565 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21566 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
21568 2013-05-06 Alan Modra <amodra@gmail.com>
21570 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
21571 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
21572 (DEFAULT_ASM_ENDIAN): Define.
21573 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21574 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
21575 Update -K PIC clause from sysv4.h.
21576 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
21577 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
21579 2013-05-06 Alan Modra <amodra@gmail.com>
21581 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
21582 twice for little-endian.
21583 (ashrdi3_no_power, ashrdi3): Support little-endian.
21585 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21588 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
21589 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
21591 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
21592 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
21593 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
21596 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
21597 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21599 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
21600 * config/pa/pa.opt: Make mbig-switch a no-op.
21601 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
21602 (CASE_VECTOR_MODE): Always return SImode.
21603 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
21604 for the !TARGET_BIG_SWITCH case.
21605 * config/pa/pa-linux.h: Likewise.
21606 * config/pa/pa-openbsd.h: Likewise.
21607 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
21608 * config/pa/pa.md (short_jump): Remove define_insn.
21609 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
21610 (casesi0): Remove define_insn.
21611 (type): Remove btable_branch.
21612 (pa_combine_type): Likewise.
21613 (in_nullified_branch_delay): Likewise.
21614 (in_call_delay): Likewise.
21615 (define_delay): Likewise.
21616 (define_insn_reservation "Z3"): Likewise.
21617 (define_insn_reservation "Z4"): Likewise.
21618 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
21619 (pa_adjust_insn_length): Remove adjustment for btable branches.
21620 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
21623 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
21625 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21626 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
21627 Add m->r,x alternatives.
21628 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
21629 splitters using SWI48x mode iterator.
21630 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
21631 TARGET_64BIT. Add m->x alternative.
21632 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
21633 Add o->x alternative. Enable for TARGET_SSE.
21634 (sse_storeq): Remove expander.
21635 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
21636 with memory input operand.
21637 (*vec_extractv2di_1 splitter): New.
21638 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
21639 * config/i386/i386.md (ssevecmodelower): New mode attribute.
21641 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21643 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
21644 (INT_LOWPART): Delete.
21645 (extract_MB): Adjust.
21646 (extract_ME): Adjust.
21647 (print_operand): Adjust.
21649 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21651 * config/rs6000/predicates.md (reg_or_add_cint_operand,
21652 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21653 (reg_or_logical_cint_operand, easy_fp_constant,
21654 logical_const_operand): Delete "CONST_DOUBLE" case.
21655 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
21656 "HOST_BITS_PER_WIDE_INT == 64" test.
21657 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
21658 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
21659 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
21660 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
21662 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
21663 CONST_DOUBLE DImode/VOIDmode case.
21664 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
21665 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21666 CONST_DOUBLE VOIDmode case.
21667 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21668 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
21669 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
21670 Delete CONST_DOUBLE case.
21671 (splitters for mov FMOVE64 const_double): Delete
21672 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21673 "HOST_BITS_PER_WIDE_INT >= 64" test.
21674 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
21676 (mov DI const_double): Delete.
21678 2013-05-04 Jakub Jelinek <jakub@redhat.com>
21680 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
21681 on op shows all bits zero in mode of a lowpart subreg, return zero.
21683 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
21686 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
21687 to save TFmode registers and DImode to save TImode registers for
21688 caller save operations.
21689 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
21690 mark being partially clobbered since they only use the first
21693 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
21694 and TDmode only use the upper 64-bits of each VSX register.
21696 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21698 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
21699 (find_candidates_in_block): Disable slsr_process_phi.
21701 2013-05-03 Guozhi Wei <carrot@google.com>
21703 * coverage.c (coverage_obj_init): Move the construction of gcov
21705 (build_init_ctor): ... here.
21707 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21709 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
21710 (slsr_cand_d): Redefine def_phi.
21711 (stride_status, phi_adjust_status, count_phis_status): New enums.
21712 (find_phi_def): New.
21713 (find_basis_for_base_expr): New.
21714 (find_basis_for_candidate): Handle hidden bases.
21715 (alloc_cand_and_find_basis): Handle phi candidates.
21716 (slsr_process_phi): New.
21717 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
21718 (create_mul_imm_cand): Likewise.
21719 (create_add_ssa_cand): Exclude phi base candidates.
21720 (create_add_imm_cand): Likewise.
21721 (slsr_process_cast): Likewise.
21722 (slsr_process_copy): Likewise.
21723 (find_candidates_in_block): Handle phi candidates.
21724 (dump_candidate): Likewise.
21725 (unconditional_cands): Delete.
21726 (unconditional_cands_with_known_stride_p): Delete.
21727 (phi_dependent_cand_p): New.
21728 (cand_increment): Handle phi-dependent candidates.
21729 (replace_dependent): Delete.
21730 (replace_mult_candidate): New.
21731 (replace_unconditional_candidate): New.
21732 (incr_vec_index): Move to avoid forward reference.
21733 (create_add_on_incoming_edge): New.
21734 (create_phi_basis): New.
21735 (replace_dependents): Delete.
21736 (replace_conditional_candidate): New.
21737 (phi_add_costs): New.
21738 (replace_uncond_cands_and_profitable_phis): New.
21739 (record_increment): Handle phi adjustments.
21740 (record_phi_increments): New.
21741 (record_increments): Handle phi adjustments.
21742 (phi_incr_cost): New.
21743 (lowest_cost_path): Handle phis.
21744 (total_savings): Likewise.
21745 (analyze_increments): Likewise.
21746 (ncd_with_phi): New.
21747 (ncd_of_cand_and_phis): New.
21748 (nearest_common_dominator_for_cands): Handle phi increments.
21749 (all_phi_incrs_profitable): New.
21750 (replace_profitable_candidates): Handle phi-dependent candidates.
21751 (analyze_candidates_and_replace): Likewise.
21753 2013-05-03 Teresa Johnson <tejohnson@google.com>
21756 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
21757 do not exceed REG_BR_PROB_BASE.
21759 2013-05-03 Jeff Law <law@redhat.com>
21761 PR tree-optimization/57144
21762 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
21763 operand of the condition will bit into the new type when eliminating
21764 a cast feeding a condition.
21766 2013-05-03 Jakub Jelinek <jakub@redhat.com>
21768 PR rtl-optimization/57130
21769 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
21770 of COMPARE as in_code to the recursive call if needed.
21772 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
21774 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
21775 (enabled): Handle new members.
21776 * config/i386/sse.md (*vec_concatv2si): Merge from
21777 *vec_concatv2si_sse2 and vec_concatv2si_sse.
21778 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
21780 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
21782 PR tree-optimization/57027
21783 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
21784 for fnms opportunity, check we got the prerequisite kind
21785 of tree / gimple before using accessor functions.
21787 2013-05-03 Richard Biener <rguenther@suse.de>
21789 * double-int.h (lshift): New overload without precision
21790 and arith argument.
21791 (operator *=, operator +=, operator -=): Move ...
21792 * double-int.c (operator *=, operator +=, operator -=): ... here
21793 and implement more efficiently.
21794 (mul_double_with_sign): Remove.
21795 (lshift_double): Adjust to take unsinged shift argument, push
21796 dispatching code to callers.
21797 (mul_double_wide_with_sign): Add early out for callers that
21798 are not interested in high parts or overflow.
21799 (lshift): New function.
21800 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
21801 dispatch code here.
21802 (lrotate, rrotate): Use logical shifts.
21803 * expr.c (get_inner_reference): Use lshift.
21804 * fixed-value.c (do_fixed_divide): Likewise.
21805 * tree-dfa.c (get_ref_base_and_extent): Likewise.
21806 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
21807 (indirect_refs_may_alias_p): Likewise.
21808 (stmt_kills_ref_p_1): Likewise.
21810 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21812 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
21814 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21816 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
21817 scalar form of FABD instruction.
21819 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
21821 * lra-constraints.c (process_alt_operands): Add checking alt
21822 number to choose the best alternative.
21824 2013-05-02 Richard Biener <rguenther@suse.de>
21826 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
21827 bitmap and its handling.
21828 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
21830 2013-05-02 Richard Biener <rguenther@suse.de>
21832 PR middle-end/57140
21833 * tree-inline.c (copy_loops): Properly handle removed loops.
21834 (copy_cfg_body): Mark destination loops for fixup if source
21835 loops needed fixup.
21837 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
21840 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
21841 generating simple_return for naked functions.
21843 2013-05-02 Martin Jambor <mjambor@suse.cz>
21845 PR middle-end/56988
21846 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
21847 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
21849 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21850 ipa_agg_replacement_value structures.
21851 (known_aggs_to_agg_replacement_list): Likewise.
21852 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21853 (read_agg_replacement_chain): Likewise.
21854 (ipcp_transform_function): Also check that by_ref flags match.
21856 2013-05-02 Richard Biener <rguenther@suse.de>
21858 * graphds.h (struct graph): Add obstack member.
21859 * graphds.c (new_graph): Initialize obstack and allocate
21861 (add_edge): Allocate edge from the obstack.
21862 (free_graph): Free the obstack instead of all edges and vertices.
21864 2013-05-02 Teresa Johnson <tejohnson@google.com>
21866 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21868 * cfg.c (update_bb_profile_for_threading): Ditto.
21869 * tree-inline.c (copy_bb): Ditto.
21870 (copy_edges_for_bb): Ditto.
21871 (initialize_cfun): Ditto.
21872 (copy_cfg_body): Ditto.
21873 (expand_call_inline): Ditto.
21874 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21875 (estimate_node_size_and_time): Ditto.
21876 (inline_merge_summary): Ditto.
21877 * cgraphclones.c (cgraph_clone_edge): Ditto.
21878 (cgraph_clone_node): Ditto.
21879 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21880 (compute_trg_info): Ditto.
21882 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21884 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21885 S reg when fp attribute set.
21886 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21888 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21890 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21892 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21893 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21894 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21896 2013-05-02 Richard Biener <rguenther@suse.de>
21898 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21899 (struct instantiate_cache_entry): New type.
21900 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21901 (struct instantiate_cache_type): New type.
21902 (set_instantiated_value, get_instantiated_value): Remove.
21903 (get_instantiated_value_entry): New function.
21904 (instantiate_scev_name): Use the new cache and adjust.
21905 (instantiate_scev_poly): Adjust.
21906 (instantiate_scev_binary): Likewise.
21907 (instantiate_array_ref): Likewise.
21908 (instantiate_scev_convert): Likewise.
21909 (instantiate_scev_not): Likewise.
21910 (instantiate_scev_3): Likewise.
21911 (instantiate_scev_2): Likewise.
21912 (instantiate_scev_r): Likewise.
21913 (instantiate_scev): Likewise.
21914 (resolve_mixers): Likewise.
21916 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21919 * lra-constraints.c (best_small_class_operands_num): Remove.
21920 (process_alt_operands): Remove small_class_operands_num. Take
21921 small classes operands into losers and only if the operand is not
21922 matched. Modify debugging output.
21923 (curr_insn_transform): Remove best_small_class_operands_num.
21926 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21928 * config/aarch64/aarch64-builtins.c
21929 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21930 * config/aarch64/aarch64-simd-builtins.def
21931 (reduc_splus_): Add new modes.
21932 (reduc_uplus_): New.
21933 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21934 (reduc_uplus_v4sf): Likewise.
21935 (reduc_splus_v4sf): Likewise.
21936 (aarch64_addv<mode>): Likewise.
21937 (reduc_uplus_<mode>): Likewise.
21938 (reduc_splus_<mode>): Likewise.
21939 (aarch64_addvv2di): Likewise.
21940 (reduc_uplus_v2di): Likewise.
21941 (reduc_splus_v2di): Likewise.
21942 (aarch64_addvv2si): Likewise.
21943 (reduc_uplus_v2si): Likewise.
21944 (reduc_splus_v2si): Likewise.
21945 (reduc_<sur>plus_<mode>): New.
21946 (reduc_<sur>plus_v2di): Likewise.
21947 (reduc_<sur>plus_v2si): Likewise.
21948 (reduc_<sur>plus_v4sf): Likewise.
21949 (aarch64_addpv4sf): Likewise.
21950 * config/aarch64/arm_neon.h
21951 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21952 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21953 add UNSPEC_SADDV, UNSPEC_UADDV.
21955 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21957 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21959 * config/aarch64/arm_neon.h
21960 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21962 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21964 * config/aarch64/aarch64-builtins
21965 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21967 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21969 * config/aarch64/aarch64-simd-builtins.def
21970 (reduc_smax_): New.
21971 (reduc_smin_): Likewise.
21972 (reduc_umax_): Likewise.
21973 (reduc_umin_): Likewise.
21974 (reduc_smax_nan_): Likewise.
21975 (reduc_smin_nan_): Likewise.
21978 (smax): Update for V2SF, V4SF and V2DF modes.
21981 (smin_nan): Likewise.
21982 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21983 (<su><maxmin><mode>3): ...This, refactor.
21984 (s<maxmin><mode>3): New.
21985 (<maxmin_uns><mode>3): Likewise.
21986 (reduc_<maxmin_uns>_<mode>): Refactor.
21987 (reduc_<maxmin_uns>_v4sf): Likewise.
21988 (reduc_<maxmin_uns>_v2si): Likewise.
21989 (aarch64_<fmaxmin><mode>: Remove.
21990 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21992 (vmin<q>_f<32,64>): Likewise.
21993 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21995 (su): Add mappings for smax, smin, umax, umin.
21997 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21998 (FMAXMIN): Rename as...
21999 (FMAXMIN_UNS): ...This.
22001 (fmaxminv): Likewise.
22002 (fmaxmin): Likewise.
22004 (maxmin_uns_op): Likewise.
22006 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22008 * config/aarch64/arm_neon.h
22009 (vac<ge, gt><sd>_f<32, 64>): Rename to...
22010 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
22011 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
22013 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22015 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
22016 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
22018 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22020 * config/aarch64/aarch64-simd.md
22021 (vcond<mode>_internal): Handle special cases for constant masks.
22022 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
22023 (vcondu<mode><mode>): Likewise.
22024 (vcond<v_cmp_result><mode>): New.
22026 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22028 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
22029 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
22030 * config/aarch64/aarch64-simd-builtins.def
22031 (cmeq): Update to BUILTIN_VALLDI.
22036 * config/aarch64/arm_neon.h
22037 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
22038 to builtins or C as appropriate.
22040 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22042 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
22044 (cmhi): Rename to...
22046 * config/aarch64/aarch64-simd.md
22047 (simd_mode): Add SF.
22048 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
22049 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
22050 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
22051 (cstore<mode>_neg): ...This.
22052 * config/aarch64/iterators.md
22054 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
22055 (COMPARISONS): New.
22056 (UCOMPARISONS): Likewise.
22057 (optab): Add missing comparisons.
22063 (VCMP_S): Likewise.
22064 (VCMP_U): Likewise.
22065 (V_cmp_result): Add DF, SF modes.
22066 (v_cmp_result): Likewise.
22068 (vmtype): Likewise.
22069 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
22071 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
22073 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
22074 define_insn to define_insn_and_split.
22075 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
22076 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
22077 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
22078 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
22079 (thumb2_negscc): Likewise.
22081 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
22083 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
22085 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
22087 * config/arm/thumb2.md: Remove trailing whitespaces.
22089 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
22091 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
22092 Use gen_int_mode rather than GEN_INT.
22094 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
22096 * value-prof.c (stream_in_histogram_value): Remove the strayed
22099 2013-04-30 Richard Biener <rguenther@suse.de>
22101 PR middle-end/57122
22102 * cfghooks.c (split_edge): Properly check for the loop latch edge.
22104 2013-04-30 Richard Biener <rguenther@suse.de>
22106 PR middle-end/57107
22107 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
22109 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
22111 PR rtl-optimization/56957
22112 PR rtl-optimization/57105
22113 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
22114 variable. Use just INSN_UID for determining whether an insn
22115 should be only disconnected from the insn stream.
22116 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
22118 2013-04-30 Jakub Jelinek <jakub@redhat.com>
22120 PR tree-optimization/57104
22121 * tsan.c (instrument_expr): Don't instrument accesses to
22122 DECL_HARD_REGISTER VAR_DECLs.
22124 2013-04-30 Richard Biener <rguenther@suse.de>
22126 * function.h (loops_for_fn): New inline function.
22127 (set_loops_for_fn): Likewise.
22128 * cfgloop.h (place_new_loop): Add struct function parameter.
22129 (get_loop): Likewise.
22130 (get_loops): Likewise.
22131 (number_of_loops): Likewise.
22132 (fel_next): Adjust.
22133 (fel_init): Likewise.
22134 * cfg.c (get_loop_copy): Adjust.
22135 * cfgloop.c (flow_loops_dump): Likewise.
22136 (record_loop_exits): Likewise.
22137 (verify_loop_structure): Likewise.
22138 * cfgloopanal.c (mark_irreducible_loops): Likewise.
22139 (estimate_reg_pressure_cost): Likewise.
22140 (mark_loop_exit_edges): Likewise.
22141 * cfgloopmanip.c (place_new_loop): Likewise.
22142 (add_loop): Likewise.
22143 (duplicate_loop): Likewise.
22144 * graph.c (draw_cfg_nodes): Likewise.
22145 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
22146 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
22147 (extract_affine_chrec): Likewise.
22148 (build_scop_iteration_domain): Likewise.
22149 * graphite.c (graphite_initialize): Likewise.
22150 * ira-build.c (create_loop_tree_nodes): Likewise.
22151 (more_one_region_p): Likewise.
22152 (rebuild_regno_allocno_maps): Likewise.
22153 (mark_loops_for_removal): Likewise.
22154 (mark_all_loops_for_removal): Likewise.
22155 (remove_unnecessary_regions): Likewise.
22156 (ira_build): Likewise.
22157 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
22158 * loop-init.c (fix_loop_structure): Likewise.
22159 (gate_rtl_move_loop_invariants): Likewise.
22160 (gate_rtl_unswitch): Likewise.
22161 (gate_rtl_unroll_and_peel_loops): Likewise.
22162 (rtl_doloop): Likewise.
22163 * lto-streamer-in.c (input_cfg): Likewise.
22164 * lto-streamer-out.c (output_cfg): Likewise.
22165 * modulo-sched.c (sms_schedule): Likewise.
22166 * predict.c (tree_estimate_probability): Likewise.
22167 (tree_estimate_probability_driver): Likewise.
22168 (estimate_loops): Likewise.
22169 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
22170 (move_sese_region_to_fn): Likewise.
22171 (debug_loop_num): Likewise.
22172 * tree-chrec.c (chrec_evaluate): Likewise.
22173 (hide_evolution_in_other_loops_than_loop): Likewise.
22174 (chrec_component_in_loop_num): Likewise.
22175 (reset_evolution_in_loop): Likewise.
22176 (evolution_function_is_invariant_rec_p): Likewise.
22177 * tree-if-conv.c (main_tree_if_conversion): Likewise.
22178 * tree-inline.c (copy_loops): Likewise.
22179 (copy_cfg_body): Likewise.
22180 (tree_function_versioning): Likewise.
22181 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
22182 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
22184 (add_to_evolution_1): Likewise.
22185 (scev_const_prop): Likewise.
22186 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
22187 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
22188 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
22189 (tree_ssa_lim_initialize): Likewise.
22190 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
22191 (verify_loop_closed_ssa): Likewise.
22192 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
22193 (tree_ssa_loop_im): Likewise.
22194 (tree_ssa_loop_unswitch): Likewise.
22195 (tree_vectorize): Likewise.
22196 (check_data_deps): Likewise.
22197 (tree_ssa_loop_ivcanon): Likewise.
22198 (tree_ssa_loop_bounds): Likewise.
22199 (tree_complete_unroll): Likewise.
22200 (tree_complete_unroll_inner): Likewise.
22201 (tree_parallelize_loops): Likewise.
22202 (tree_ssa_loop_prefetch): Likewise.
22203 (tree_ssa_loop_ivopts): Likewise.
22204 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
22205 * tree-vectorizer.c (vectorize_loops): Likewise.
22207 2013-04-29 Mike Frysinger <vapier@gentoo.org>
22209 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
22210 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
22211 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
22212 with EABI_LINK_SPEC.
22214 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22217 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
22220 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
22223 * lra-constraints.c (process_alt_operands): Discourage a bit more
22224 using memory for pseudos. Print cost dump for alternatives.
22225 Modify cost values for conflicts with early clobbers.
22226 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
22228 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22231 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
22233 2013-04-29 Ian Bolton <ian.bolton@arm.com>
22235 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
22236 from/to S register.
22237 (movdi_aarch64): Support LDR/STR from/to D register.
22239 2013-04-29 Ian Bolton <ian.bolton@arm.com>
22241 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
22242 or higher by default.
22244 2013-04-29 Richard Biener <rguenther@suse.de>
22246 PR middle-end/57075
22247 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
22248 even if not adding abnormal edges for calls that can make
22251 2013-04-29 Richard Biener <rguenther@suse.de>
22253 PR middle-end/57103
22254 * tree-cfg.c (move_stmt_op): Fix condition under which to update
22256 (move_stmt_r): Remove redundant checking.
22258 2013-04-29 Teresa Johnson <tejohnson@google.com>
22261 * basic-block.h (apply_scale): New function.
22262 (apply_probability): Use apply_scale.
22263 * gimple-streamer-in.c (input_bb): Ditto.
22264 * lto-streamer-in.c (input_cfg): Ditto.
22265 * lto-cgraph.c (merge_profile_summaries): Ditto.
22266 * tree-optimize.c (execute_fixup_cfg): Ditto.
22267 * tree-inline.c (copy_bb): Update comment to use apply_scale.
22268 (copy_edges_for_bb): Ditto.
22269 (copy_cfg_body): Ditto.
22271 2013-04-29 Tom de Vries <tom@codesourcery.com>
22273 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
22274 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
22275 (tail_merge_optimize): Handle current_loops == NULL.
22277 2013-04-26 Jeff Law <law@redhat.com>
22279 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
22280 (simplify_cond_using_ranges): Generalize code to simplify
22281 COND_EXPRs where one argument is a constant and the other
22282 is an SSA_NAME created by an integral type conversion.
22284 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22286 * config/arm/arm.md (store_minmaxsi): Use only when
22287 optimize_insn_for_size_p.
22289 2013-04-29 Christian Bruel <christian.bruel@st.com>
22292 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
22294 2013-04-29 Richard Biener <rguenther@suse.de>
22296 PR middle-end/57089
22297 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
22298 loop tree make sure to schedule a fixup for the child as well.
22299 (expand_omp_for_generic): Properly add loops.
22300 (expand_omp_for_static_nochunk): Likewise.
22301 (expand_omp_for_static_chunk): Likewise.
22302 (expand_omp_for): For the degenerate case fixup loops.
22303 (expand_omp_sections): Fix default bb placement in loops.
22304 (expand_omp_atomic_pipeline): Properly add loops.
22306 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22308 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
22310 2013-04-29 Tom de Vries <tom@codesourcery.com>
22312 * tree-ssa-tail-merge.c: Update header comment.
22314 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22316 * config/aarch64/arm_neon.h
22317 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
22318 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
22319 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
22320 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
22321 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
22322 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
22323 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
22324 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
22326 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22328 * config/aarch64/aarch64-simd.md
22329 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
22330 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
22331 fix_trunc, fixuns_trunc.
22332 (ftrunc<VDQF:mode>2): New.
22333 * config/aarch64/iterators.md (optab): Add fix, fixuns.
22334 (fix_trunc_optab): New.
22336 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22338 * config/aarch64/aarch64-builtins.c
22339 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
22340 iceilf, lround, iroundf.
22342 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22345 * config/i386/i386.h (enum ix86_tune_indices)
22346 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
22347 New, split from X86_TUNE_INTER_UNIT_MOVES.
22348 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
22349 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
22350 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
22351 (TARGET_INTER_UNIT_MOVES): Remove.
22352 * config/i386/i386.c (initial_ix86_tune_features): Update.
22353 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
22354 (ix86_expand_convert_uns_didf_sse): Use
22355 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
22356 (ix86_expand_vector_init_one_nonzero): Ditto.
22357 (ix86_expand_vector_init_interleave): Ditto.
22358 (inline_secondary_memory_needed): Return true for moves from SSE class
22359 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
22360 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
22361 * config/i386/constraints.md (Yi, Ym): Depend on
22362 TARGET_INTER_UNIT_MOVES_TO_VEC.
22363 (Yj, Yn): New constraints.
22364 * config/i386/i386.md (*movdi_internal): Change constraints of
22365 operand 1 from Yi to Yj and from Ym to Yn.
22366 (*movsi_internal): Ditto.
22367 (*movdf_internal): Ditto.
22368 (*movsf_internal): Ditto.
22369 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
22370 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
22371 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
22372 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
22373 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
22374 * config/i386/sse.md (movdi_to_sse): Ditto.
22375 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
22376 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
22377 TARGET_INTER_UNIT_MOVES.
22378 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
22379 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
22380 instead of TARGET_INTER_UNIT_MOVES.
22381 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
22382 operand 1 from Yi to Yj and from Ym to Yn.
22384 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22386 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
22387 (float_truncate_hi_): Likewise.
22388 (float_extend_lo_): Likewise.
22389 (float_truncate_lo_): Likewise.
22390 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
22391 (aarch64_float_extend_lo_v2df): Likewise.
22392 (vec_unpacks_hi_v4sf): Likewise.
22393 (aarch64_float_truncate_lo_v2sf): Likewise.
22394 (aarch64_float_truncate_hi_v4sf): Likewise.
22395 (vec_pack_trunc_v2df): Likewise.
22396 (vec_pack_trunc_df): Likewise.
22398 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22400 * config/aarch64/aarch64-builtins.c
22401 (aarch64_fold_builtin): Fold float conversions.
22402 * config/aarch64/aarch64-simd-builtins.def
22403 (floatv2si, floatv4si, floatv2di): New.
22404 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
22405 * config/aarch64/aarch64-simd.md
22406 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
22407 * config/aarch64/iterators.md (FLOATUORS): New.
22408 (optab): Add float, floatuns.
22409 (su_optab): Likewise.
22411 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22413 * config/aarch64/aarch64-builtins.c
22414 (aarch64_builtin_vectorized_function): Use new names for
22416 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
22417 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
22418 (fcvtzu): Split as...
22419 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
22420 (fcvtas): Split as...
22421 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
22422 (fcvtau): Split as...
22423 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
22424 (fcvtps): Split as...
22425 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
22426 (fcvtpu): Split as...
22427 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
22428 (fcvtms): Split as...
22429 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
22430 (fcvtmu): Split as...
22431 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
22432 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
22433 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
22434 (lfrintnusf, lfrintnudf): Likewise.
22435 * config/aarch64/aarch64-simd.md
22436 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
22438 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
22439 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
22440 (fcvt_pattern): Likewise.
22442 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22444 * config/aarch64/aarch64-simd.md
22445 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
22446 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
22448 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22450 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
22451 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
22452 (vrnd<a,m,n,p>_f32): Implement using builtins.
22453 (vrnd<i,x><q>_f<32, 64>): New.
22455 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22457 * config/aarch64/aarch64-builtins.c
22458 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
22459 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
22460 (frintz): Rename to...
22462 (frintp): Rename to...
22464 (frintm): Rename to...
22466 (frinti): Rename to...
22467 (nearbyint): ...this.
22468 (frintx): Rename to...
22470 (frinta): Rename to...
22472 * config/aarch64/aarch64-simd.md
22473 (aarch64_frint<frint_suffix><mode>): Delete.
22474 (<frint_pattern><mode>2): Convert to insn.
22475 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
22476 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
22477 (frint_pattern): Likewise.
22478 (frint_suffix): Likewise.
22480 2013-04-29 Richard Biener <rguenther@suse.de>
22482 PR tree-optimization/57081
22483 * loop-init.c: Include tree-flow.h.
22484 (loop_optimizer_finalize): Free number of iteration estimates.
22485 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
22487 2013-04-29 Jakub Jelinek <jakub@redhat.com>
22489 PR tree-optimization/57083
22490 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
22491 non-singleton shift count range, zero extend low_bound for uns case.
22493 * config/i386/predicates.md (general_vector_operand): New predicate.
22494 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
22495 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
22496 if they aren't nonimmediate operands. If their original values
22497 satisfy const_vector_equal_evenodd_p, don't shift them.
22498 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
22499 predicates. For the SSE4.1 case force operands[{1,2}] into registers
22500 if not nonimmediate_operand.
22501 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
22502 instead of register_operand.
22503 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
22505 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22507 * stor-layout.c (finalize_size_functions): Allocate a structure and
22508 reset cfun before dumping the functions.
22510 2013-04-27 Jakub Jelinek <jakub@redhat.com>
22512 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
22515 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
22516 use xop_pmacsdqh if uns_p.
22517 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
22518 the immediate rotate count.
22520 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
22522 * rtl.h (struct rtx_def): Add comment for field jump.
22523 (LRA_SUBREG_P): New macro.
22524 * recog.c (register_operand): Check LRA_SUBREG_P.
22525 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22527 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22528 (lra_final_code_change): Skip subreg change for operators.
22529 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22530 if there are no operand changes.
22531 * lra-constraints.c (curr_insn_set): New.
22532 (match_reload): Set LRA_SUBREG_P.
22533 (emit_spill_move): Ditto.
22534 (check_and_process_move): Use curr_insn_set. Process only single
22535 set insns. Don't initialize sec_mem_p and change_p.
22536 (simplify_operand_subreg): Use LRA_SUBREG_P.
22537 (reg_in_class_p): New function.
22538 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22539 of #ifdef. Add code to remove cycling.
22540 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22541 non-null disp. Reload inner instead of disp when base and index
22542 are null. Try to put lo_sum into register.
22543 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22544 (check_and_process_move): Move code for move cost check to
22545 simple_move_p. Remove equiv_substitution.
22546 (simple_move_p): New function.
22547 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22548 curr_insn_set. Call check_and_process_move only for single set
22549 insns. Use the new function. Move call of check_and_process_move
22550 after operand equiv substitution and address process.
22552 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22555 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
22556 with nonlocal goto receivers or returns twice calls, ignore
22557 unininitialized values from abnormal edges to nl goto receiver
22558 or returns twice call.
22560 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22562 PR tree-optimization/57051
22563 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
22564 and VEC_RSHIFT_EXPR if shift count is a multiple of element
22567 2013-04-26 Richard Biener <rguenther@suse.de>
22569 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
22570 (expand_omp_taskreg): Likewise. Mark loops for fixup.
22571 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
22572 (fixup_loop_arrays_after_move): New function.
22573 (move_sese_region_to_fn): Properly outline the loop tree parts
22574 of the SESE region.
22576 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
22578 * config/i386/i386.md (type, unit): Fix long lines.
22580 2013-04-26 Richard Biener <rguenther@suse.de>
22582 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
22583 (lto-streamer-out.o): Likewise.
22584 * cfgloop.c (init_loops_structure): Export, add struct function
22585 argument and adjust.
22586 (flow_loops_find): Adjust.
22587 * cfgloop.h (enum loop_estimation): Add EST_LAST.
22588 (init_loops_structure): Declare.
22589 * lto-streamer-in.c: Include cfgloop.h.
22590 (input_cfg): Input the loop tree.
22591 * lto-streamer-out.c: Include cfgloop.h.
22592 (output_cfg): Output the loop tree.
22593 (output_struct_function_base): Do not drop PROP_loops.
22595 2013-03-26 Richard Biener <rguenther@suse.de>
22597 * tree-cfg.c (execute_build_cfg): Build the loop tree.
22598 (pass_build_cfg): Provide PROP_loops.
22599 (move_sese_region_to_fn): Remove loops that are outlined into fn
22601 * tree-inline.c: Include cfgloop.h.
22602 (initialize_cfun): Do not drop PROP_loops.
22603 (copy_loops): New function.
22604 (copy_cfg_body): Copy loop structure.
22605 (tree_function_versioning): Initialize destination loop tree.
22606 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
22607 (pass_parallelize_loops): Do IL verification.
22608 * loop-init.c (loop_optimizer_init): Fixup loops if required.
22609 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
22610 the CFG make sure we fixup loops as well.
22611 * tree-ssa-tail-merge.c: Include cfgloop.h.
22612 (replace_block_by): When merging loop latches mark loops for fixup.
22613 * lto-streamer-out.c (output_struct_function_base): Drop
22614 PROP_loops for now.
22615 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
22616 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
22617 * ipa-split.c: Include cfgloop.h.
22618 (split_function): Add the new return block to the loop tree root.
22619 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
22620 whether we have removed the forwarder block.
22621 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
22622 * cfgloop.h (place_new_loop): Declare.
22623 * cfgloopmanip.c (place_new_loop): Export.
22624 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
22625 (tree-switch-conversion.o): Likewise.
22626 (tree-complex.o): Likewise.
22627 (tree-inline.o): Likewise.
22628 (tree-ssa-tailmerge.o): Likewise.
22629 (ipa-split.o): Likewise.
22630 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
22631 (tree-ssa-copy.o): Likewise.
22632 * tree-switch-conversion.c: Include cfgloop.h
22633 (process_switch): If we emit a bit-test cascade, schedule loops
22635 * tree-complex.c: Include cfgloop.h.
22636 (expand_complex_div_wide): Properly add new basic-blocks to loops.
22637 * asan.c: Include cfgloop.h.
22638 (create_cond_insert_point): Properly add new basic-blocks to
22639 loops, schedule loop fixup.
22640 * cfgloop.c (verify_loop_structure): Check that looks are not
22642 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
22644 (expand_omp_for_generic): Likewise.
22645 (expand_omp_sections): Likewise.
22646 (expand_omp_atomic_pipeline): Schedule loops for fixup.
22647 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
22648 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
22649 is initialized, not when loops are present.
22650 * tree-parloops.c (parallelize_loops): Remove checking here.
22651 * passes.c (init_optimization_passes): Schedule a copy-propagation
22652 pass before complete unrolling of inner loops.
22654 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22656 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
22657 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
22658 (colorize_init): Add argument to _WIN32 version.
22659 * toplev.c: Include diagnostic-color.h.
22660 (process_options): Default to -fdiagnostics-color=auto if
22661 GCC_COLORS env var is in the environment.
22662 * common.opt (fdiagnostics-color=): Add Var and Init.
22663 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
22664 env var is in the environment, the default is auto rather than never.
22666 * diagnostic.h (file_name_as_prefix): Add context argument.
22667 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
22668 the string as locus.
22669 * langhooks.c (lhd_print_error_function): Adjust caller.
22671 2013-04-25 Lawrence Crowl <crowl@google.com>
22673 * var-tracking.c (shared_hash_def::htab):
22674 Change type to hash_table. Update dependent calls and types.
22676 2013-04-25 Lawrence Crowl <crowl@google.com>
22678 * Makefile.in: Update as needed below.
22680 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
22681 Move declaration to after the type's method definitons.
22683 * attribs.c (htab_t scoped_attributes::attribute_hash):
22684 Change type to hash_table. Update dependent calls and types.
22686 * bitmap.c (htab_t bitmap_desc_hash):
22687 Change type to hash_table. Update dependent calls and types.
22689 * cselib.c (htab_t cselib_hash_table):
22690 Change type to hash_table. Update dependent calls and types.
22692 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
22693 (hash_string_slot_node): Move implementation into lto-streamer.h
22694 struct string_slot_hasher.
22695 (eq_string_slot_node): Likewise.
22697 * data-streamer-out.c: Update output_block::string_hash_table
22698 dependent calls and types.
22700 * dwarf2cfi.c (htab_t trace_index):
22701 Change type to hash_table. Update dependent calls and types.
22703 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
22704 Change type to hash_table. Update dependent calls and types.
22705 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
22706 (htab_t optimize_external_refs::map): Likewise.
22707 (htab_t output_comp_unit::extern_map): Likewise.
22708 (htab_t output_comdat_type_unit::extern_map): Likewise.
22709 (htab_t output_macinfo::macinfo_htab): Likewise.
22710 (htab_t optimize_location_lists::htab): Likewise.
22711 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
22713 * except.c (htab_t ehspec_hash_type):
22714 Change type to hash_table. Update dependent calls and types.
22715 (assign_filter_values::ttypes): Likewise.
22716 (assign_filter_values::ehspec): Likewise.
22717 (sjlj_assign_call_site_values::ar_hash): Likewise.
22718 (convert_to_eh_region_ranges::ar_hash): Likewise.
22720 * gcse.c (htab_t pre_ldst_table):
22721 Change type to hash_table. Update dependent calls and types.
22723 * ggc-common.c (htab_t saving_htab):
22724 Change type to hash_table. Update dependent calls and types.
22725 (htab_t loc_hash): Likewise.
22726 (htab_t ptr_hash): Likewise.
22727 (call_count): Rename ggc_call_count.
22728 (call_alloc): Rename ggc_call_alloc.
22729 (loc_descriptor): Rename make_loc_descriptor.
22730 (add_statistics): Rename ggc_add_statistics.
22732 * ggc-common.c (saving_htab):
22733 Change type to hash_table. Update dependent calls and types.
22735 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
22736 (push_gimplify_context): Likewise.
22737 (pop_gimplify_context): Likewise.
22738 (struct gimple_temp_hash_elt): Added.
22739 (struct gimplify_hasher): Likewise.
22740 (struct gimplify_ctx.temp_htab):
22741 Change type to hash_table. Update dependent calls and types.
22743 * gimple-fold.c: Include gimplify-ctx.h.
22745 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
22746 Change type to hash_table. Update dependent calls and types.
22747 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
22748 avoid potential global name collision.
22750 * gimplify.c: Include gimplify-ctx.h.
22751 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
22752 (htab_t gimplify_ctx::temp_htab):
22753 Update dependent calls and types for new type hash_table.
22754 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
22755 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
22757 * gimplify-ctx.h: New.
22758 (struct gimple_temp_hash_elt): Move from gimplify.c.
22759 (class gimplify_hasher): New.
22760 (struct gimplify_ctx): Move from gimple.h.
22761 (htab_t gimplify_ctx::temp_htab):
22762 Change type to hash_table. Update dependent calls and types.
22764 * graphite-clast-to-gimple.c: Include graphite-htab.h.
22765 (htab_t ivs_params::newivs_index):
22766 Change type to hash_table. Update dependent calls and types.
22767 (htab_t ivs_params::params_index): Likewise.
22768 (htab_t print_generated_program::params_index): Likewise.
22769 (htab_t gloog::newivs_index): Likewise.
22770 (htab_t gloog::params_index): Likewise.
22772 * graphite.c: Include graphite-htab.h.
22773 4htab_t graphite_transform_loops::bb_pbb_mapping):
22774 Change type to hash_table. Update dependent calls and types.
22776 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
22777 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
22778 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
22780 * graphite-dependences.c: Include graphite-htab.h.
22781 (loop_is_parallel_p): Change hash table type of parameter.
22783 * graphite-htab.h: New.
22784 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
22785 (extern find_pbb_via_hash): Move from graphite-poly.h.
22786 (extern loop_is_parallel_p): Move from graphite-poly.h.
22787 (extern get_loop_body_pbbs): Move from graphite-poly.h.
22789 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
22790 (extern loop_is_parallel_p): Move to graphite-htab.h.
22791 (extern get_loop_body_pbbs): Move to graphite-htab.h.
22793 * haifa-sched.c (htab_t delay_htab):
22794 Change type to hash_table. Update dependent calls and types.
22795 (htab_t delay_htab_i2): Likewise.
22797 * ira-color.c (htab_t allocno_hard_regs_htab):
22798 Change type to hash_table. Update dependent calls and types.
22800 * ira-costs.c (htab_t cost_classes_htab):
22801 Change type to hash_table. Update dependent calls and types.
22803 * loop-invariant.c (htab_t merge_identical_invariants::eq):
22804 Change type to hash_table. Update dependent calls and types.
22806 * loop-iv.c (htab_t bivs):
22807 Change type to hash_table. Update dependent calls and types.
22809 * loop-unroll.c (htab_t opt_info::insns_to_split):
22810 Change type to hash_table. Update dependent calls and types.
22811 (htab_t opt_info::insns_with_var_to_expand): Likewise.
22813 * lto-streamer.h (struct string_slot): Move from data-streamer.h
22814 (struct string_slot_hasher): New.
22815 (htab_t output_block::string_hash_table):
22816 Change type to hash_table. Update dependent calls and types.
22818 * lto-streamer-in.c (freeing_string_slot_hasher): New.
22819 (htab_t file_name_hash_table):
22820 Change type to hash_table. Update dependent calls and types.
22822 * lto-streamer-out.c: Update output_block::string_hash_table dependent
22825 * lto-streamer.c (htab_t tree_htab):
22826 Change type to hash_table. Update dependent calls and types.
22828 * omp-low.c: Include gimplify-ctx.h.
22830 * passes.c (htab_t name_to_pass_map):
22831 Change type to hash_table. Update dependent calls and types.
22832 (pass_traverse): Rename to passes_pass_traverse.
22834 * plugin.c (htab_t event_tab):
22835 Change type to hash_table. Update dependent calls and types.
22837 * postreload-gcse.c (htab_t expr_table):
22838 Change type to hash_table. Update dependent calls and types.
22839 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
22841 * sese.c (debug_rename_map_1): Make extern.
22842 (htab_t copy_bb_and_scalar_dependences::rename_map):
22843 Change type to hash_table. Update dependent calls and types.
22845 * sese.h (extern debug_rename_map): Move to .c file.
22847 * store-motion.c (htab_t store_motion_mems_table):
22848 Change type to hash_table. Update dependent calls and types.
22850 * trans-mem.c (htab_t tm_new_mem_hash):
22851 Change type to hash_table. Update dependent calls and types.
22853 * tree-browser.c (htab_t TB_up_ht):
22854 Change type to hash_table. Update dependent calls and types.
22856 * tree-cfg.c (htab_t discriminator_per_locus):
22857 Change type to hash_table. Update dependent calls and types.
22859 * tree-complex.c: Include tree-hasher.h
22860 (htab_t complex_variable_components):
22861 Change type to hash_table. Update dependent calls and types.
22863 * tree-eh.c (htab_t finally_tree):
22864 Change type to hash_table. Update dependent calls and types.
22866 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22867 struct int_tree_hasher.
22868 (extern int_tree_map_eq): Likewise.
22869 (uid_decl_map_hash): Removed.
22870 (extern decl_tree_map_eq): Likewise.
22872 * tree-hasher.h: New.
22873 (struct int_tree_hasher): New.
22874 (typedef int_tree_htab_type): New.
22876 * tree-inline.c: Include gimplify-ctx.h.
22878 * tree-mudflap.c: Include gimplify-ctx.h.
22880 * tree-parloops.c: Include tree-hasher.h.
22881 (htab_t eliminate_local_variables_stmt::decl_address):
22882 Change type to hash_table. Update dependent calls and types.
22883 (htab_t separate_decls_in_region::decl_copies): Likewise.
22885 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22886 Change type to hash_table. Update dependent calls and types.
22888 * tree-sra.c (candidates):
22889 Change type to hash_table. Update dependent calls and types.
22891 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22893 (int_tree_map_hash): Likewise.
22895 * tree-ssa-dom.c (htab_t avail_exprs):
22896 Change type to hash_table. Update dependent calls and types.
22898 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22899 Change type to hash_table. Update dependent calls and types.
22901 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22902 Change type to hash_table. Update dependent calls and types.
22904 * tree-ssa-phiopt.c (seen_ssa_names):
22905 Change type to hash_table. Update dependent calls and types.
22907 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22908 Change type to hash_table. Update dependent calls and types.
22910 * tree-ssa-uncprop.c (equiv):
22911 Change type to hash_table. Update dependent calls and types.
22913 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22915 PR rtl-optimization/57003
22916 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22917 call note_stores with kill_clobbered_value callback again after
22918 killing regs_invalidated_by_call.
22920 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22922 * config/aarch64/aarch64-simd.md
22923 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22924 (aarch64_simd_bsl<mode>): Likewise.
22925 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22927 2013-04-25 Marek Polacek <polacek@redhat.com>
22929 PR tree-optimization/57066
22930 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22932 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22934 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22936 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22938 * config/aarch64/aarch64-builtins.c
22939 (aarch64_fold_builtin): New.
22940 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22941 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22942 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22943 * config/aarch64/arm_neon.h
22944 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22946 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22947 Tejas Belagod <tejas.belagod@arm.com>
22949 * config/aarch64/aarch64-builtins.c
22950 (aarch64_gimple_fold_builtin): New.
22951 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22952 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22953 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22954 (addvv4sf): Update.
22955 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22957 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22959 * config/aarch64/aarch64.md
22960 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22962 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22964 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22965 (*ngcsi_uxtw): New pattern.
22967 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22968 Julian Brown <julian@codesourcery.com>
22970 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22971 (TB_DREG): Add T_V4HF.
22972 (v4hf_UP): New macro.
22973 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22974 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22975 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22977 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22978 (arm_vector_mode_supported_p): Handle V4HF.
22979 (arm_mangle_map): Handle V4HFmode.
22980 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22981 * config/arm/arm_neon_builtins.def: Add entries for
22982 vcvtv4hfv4sf, vcvtv4sfv4hf.
22983 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22984 (neon_vcvtv4hfv4sf): Likewise.
22985 * config/arm/neon-gen.ml: Handle half-precision floating point
22987 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22988 * config/arm/arm_neon.h: Regenerate.
22989 * config/arm/neon.ml (type elts): Add F16.
22990 (type vectype): Add T_float16x4, T_floatHF.
22991 (type vecmode): Add V4HF.
22992 (type features): Add Requires_FP_bit feature.
22993 (elt_width): Handle F16.
22994 (elt_class): Likewise.
22995 (elt_of_class_width): Likewise.
22996 (mode_of_elt): Refactor.
22997 (type_for_elt): Handle F16, fix error messages.
22998 (vectype_size): Handle T_float16x4.
22999 (vcvt_sh): New function.
23000 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
23001 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
23002 (string_of_mode): Handle V4HF.
23003 * doc/arm-neon-intrinsics.texi: Regenerate.
23005 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
23007 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
23008 format specifier in 'X' case.
23010 2013-04-25 Alan Modra <amodra@gmail.com>
23013 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
23014 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
23015 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
23016 Repeat for many other rotate/shift and mask patterns using subregs.
23017 Name lshiftrt insns.
23018 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
23019 on WORDS_BIG_ENDIAN.
23021 2013-04-25 Alan Modra <amodra@gmail.com>
23023 * config.gcc: Support little-endian powerpc-linux targets.
23024 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
23025 (LINK_OS_LINUX_SPEC): Define.
23026 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
23027 Preserve MASK_LITTLE_ENDIAN.
23028 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
23029 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
23030 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
23031 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
23032 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
23033 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
23034 Correct fp word order for little-endian. Don't shift toc entries
23035 smaller than a word for little-endian.
23036 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
23037 (bswapdi2 splits): Correct low-part subreg for little-endian.
23038 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
23039 low/high where such is correct only for be.
23040 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
23041 little-endian for -mcall-aixdesc.
23043 2013-04-25 Alan Modra <amodra@gmail.com>
23045 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
23046 replace_equiv_address_nv.
23048 2013-04-25 Alan Modra <amodra@gmail.com>
23050 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
23052 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23055 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23056 * rtl.h (struct rtx_def): ...
23058 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23060 PR rtl-optimizations/57046
23061 * lra-constraints (split_reg): Set up lra_risky_transformations_p
23062 for multi-reg splits.
23064 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
23066 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
23068 2013-04-24 Sterling Augustine <saugustine@google.com>
23070 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
23071 (comp_dir_string, debug_str_dwo_section): New.
23072 (DEBUG_STR_DWO_SECTION): Rename to ...
23073 (DEBUG_DWO_STR_SECTION): ... this.
23074 (DEBUG_NORM_STR_SECTION): Delete.
23075 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
23076 (DEBUG_STR_DWO_SECTION_FLAGS): New.
23077 (find_AT_string): Move most logic to ...
23078 (find_AT_string_in_table): ... here. New.
23079 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
23080 add_skeleton_AT_string. Delete logic.
23081 (output_skeleton_debug_sections): Remove call to
23082 add_top_level_skeleton_die_attrs.
23083 (add_comp_dir_attribute): Move logic to comp_dir_string.
23084 (dwarf2out_init): Initialize debug_str_dwo_section.
23085 (output_indirect_string): Call find_string_form.
23086 (output_indirect_strings): Rewrite.
23087 (prune_unused_types): Empty skeleton_debug_str_hash.
23088 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
23089 (dwarf2out_finish): Call output_indirect_strings.
23091 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
23093 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
23095 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23097 * rtl.h (struct rtx_def): Add comment for field jump.
23098 (LRA_SUBREG_P): New macro.
23099 * recog.c (register_operand): Check LRA_SUBREG_P.
23100 * lra.c (lra): Add note at the end of RTL code. Align non-empty
23102 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
23103 (lra_final_code_change): Skip subreg change for operators.
23104 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
23105 if there are no operand changes.
23106 * lra-constraints.c (curr_insn_set): New.
23107 (match_reload): Set LRA_SUBREG_P.
23108 (emit_spill_move): Ditto.
23109 (check_and_process_move): Use curr_insn_set. Process only single
23110 set insns. Don't initialize sec_mem_p and change_p.
23111 (simplify_operand_subreg): Use LRA_SUBREG_P.
23112 (reg_in_class_p): New function.
23113 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
23114 of #ifdef. Add code to remove cycling.
23115 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
23116 non-null disp. Reload inner instead of disp when base and index
23117 are null. Try to put lo_sum into register.
23118 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
23119 (check_and_process_move): Move code for move cost check to
23120 simple_move_p. Remove equiv_substitution.
23121 (simple_move_p): New function.
23122 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
23123 curr_insn_set. Call check_and_process_move only for single set
23124 insns. Use the new function. Move call of check_and_process_move
23125 after operand equiv substitution and address process.
23127 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
23129 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
23130 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
23131 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
23133 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
23135 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
23137 2013-04-24 Marek Polacek <polacek@redhat.com>
23139 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
23140 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
23141 (select_loops_exit_conditions): Likewise.
23142 (number_of_iterations_for_all_loops): Likewise.
23143 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
23144 (scev_analysis): Likewise.
23146 2013-04-02 Catherine Moore <clm@codesourcery.com>
23147 Chao-ying Fu <fu@mips.com>
23149 * config/mips/micromips.md (jraddiusp): New pattern.
23150 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
23151 instruction if possible.
23153 2013-04-24 Alan Modra <amodra@gmail.com>
23155 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
23157 2013-04-24 Julian Brown <julian@codesourcery.com>
23158 Chung-Lin Tang <cltang@codesourcery.com>
23160 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
23161 dependency behavior in enumeration type DIE generation. Add TODO note
23162 to comments about future DW_FORM_sdata/udata re-work of related code.
23164 2013-04-23 Lawrence Crowl <crowl@google.com>
23166 * Makefile.in: Update as needed below.
23168 * hash-table.h (class hash_table):
23169 Correct many methods with parameter types compare_type to the correct
23170 value_type. (Correct code was unlikely to notice the change.)
23171 (hash_table::elements_with_deleted) New.
23172 (class hashtable::iterator): New.
23173 (hashtable::begin()): New.
23174 (hashtable::end()): New.
23175 (FOR_EACH_HASH_TABLE_ELEMENT): New.
23177 * statistics.c (statistics_hashes):
23178 Change type to hash_table. Update dependent calls and types.
23180 * tree-into-ssa.c (var_infos):
23181 Change type to hash_table. Update dependent calls and types.
23183 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
23184 Change type to hash_table. Update dependent calls and types.
23186 * tree-ssa-loop-im.c (struct mem_ref.refs):
23187 Change type to hash_table. Update dependent calls and types.
23189 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
23190 Change type to hash_table. Update dependent calls and types.
23192 * tree-ssa-sccvn.c (vn_tables_s::nary):
23193 Change type to hash_table. Update dependent calls and types.
23194 (vn_tables_s::phis): Likewise.
23195 (vn_tables_s::references): Likewise.
23197 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
23198 (vn_reference_eq): Update parameter and return types.
23200 * tree-ssa-structalias.c (pointer_equiv_class_table):
23201 Change type to hash_table. Update dependent calls and types.
23202 (location_equiv_class_table): Likewise.
23204 * tree-vect-data-refs.c: Consequential changes for making
23205 peeling a hash_table.
23207 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
23208 (destroy_loop_vec_info): Dependent hash_table update.
23210 * tree-vectorizer.h (peeling_htab):
23211 Change type to hash_table. Update dependent calls and types.
23213 2013-04-23 Shiva Chen <shiva0217@gmail.com>
23215 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
23216 to check the register content is equal or not.
23217 * lra-constraints.c (match_reload): Use lra_assign_reg_val
23218 to assign register content record.
23219 * lra-eliminations.c (update_reg_eliminate): Use
23220 lra_update_reg_val_offset to update register content offset.
23221 * lra-int.h (struct lra_reg): Add offset member.
23222 (lra_reg_val_equal_p): New static inline function.
23223 (lra_update_reg_val_offset): New static inline function.
23224 (lra_assign_reg_val): New static inline function.
23225 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
23226 to assign register content record.
23227 (initialize_lra_reg_info_element): Initial offset to zero.
23229 2013-04-23 Catherine Moore <clm@codesourcery.com>
23231 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
23232 operands. Record compression.
23234 2013-04-23 Xinliang David Li <davidxl@google.com>
23236 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
23238 2013-04-23 Richard Biener <rguenther@suse.de>
23240 PR middle-end/57036
23241 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
23242 parameter, only add abnormal goto edges from the copied body
23243 if the call could perform abnormal gotos.
23244 (copy_cfg_body): Adjust.
23246 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
23248 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
23250 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
23252 * coretypes.h (gimple_stmt_iterator): Add struct to make
23255 2013-04-23 Richard Biener <rguenther@suse.de>
23257 PR tree-optimization/57026
23258 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
23259 from SSA names occuring in abnormal PHI nodes.
23261 2013-04-22 Andi Kleen <ak@linux.intel.com>
23263 * lto/lto.c (print_lto_report_1): Fix LTO report names.
23265 2013-04-22 Andi Kleen <ak@linux.intel.com>
23267 * lto/lto.c (print_lto_report_1): Declare early.
23268 (read_cgraph_and_symbols): Call print_lto_report_1 early.
23270 2013-04-22 Andi Kleen <ak@linux.intel.com>
23272 * common.opt (-flto-report-wpa): Add.
23273 * doc/invoke.texi (-flto-report-wpa): Add.
23274 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
23277 2013-04-22 Xinliang David Li <davidxl@google.com>
23279 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
23280 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
23281 * Makefile.in: New dependency
23283 David Daney <ddaney.cavm@gmail.com>
23285 * configure.ac (gcc_cv_as_micromips_support): Use the
23286 --fatal-warnings option.
23287 * configure: Regenerate.
23289 2013-04-22 Marek Polacek <polacek@redhat.com>
23292 * tsan.c (instrument_expr): Don't instrument expression
23293 in case its size is zero.
23295 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
23299 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23301 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23303 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23305 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
23306 (gimple_stmt_iterator): New typedef.
23307 * gimple.h (gimple_stmt_iterator): Rename to...
23308 (gimple_stmt_iterator_d): ... This.
23309 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
23310 trees be valid for GIMPLE and GENERIC.
23311 (TARGET_GIMPLE_FOLD_BUILTIN): New.
23312 * gimple-fold.c (gimple_fold_call): Call target hook
23313 gimple_fold_builtin.
23314 * hooks.c (hook_bool_gsiptr_false): New.
23315 * hooks.h (hook_bool_gsiptr_false): New.
23316 * target.def (fold_stmt): New.
23317 * doc/tm.texi: Regenerate.
23319 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
23322 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
23323 a set sp if no stack realignment.
23325 2013-04-22 Nick Clifton <nickc@redhat.com>
23327 * config.gcc (tilegx-linux): Extend extra_objs rather than
23329 (tilepro-linux): Likewise.
23331 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23333 * config/aarch64/aarch64-builtins.c
23335 (CF0, CF1, CF2, CF3, CF4, CF10): New.
23336 (VAR<1-12>): Add MAP parameter.
23337 (BUILTIN_*): Likewise.
23338 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
23339 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
23340 (aarch64_ushl_n<mode>): Likewise.
23341 (aarch64_sshr_n<mode>): Likewise.
23342 (aarch64_ushr_n<mode>): Likewise.
23343 (aarch64_<maxmin><mode>): Likewise.
23344 (aarch64_sqrt<mode>): Likewise.
23345 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
23346 (vshr<q>_n_*): Likewise.
23348 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23350 * config/aarch64/aarch64-builtins.c
23351 (aarch64_simd_builtin_type_mode): Handle SF types.
23353 (BUILTIN_GPF): Define.
23354 (aarch64_init_simd_builtins): Handle SF types.
23355 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
23356 (frecps): Likewise.
23357 (frecpx): Likewise.
23358 * config/aarch64/aarch64-simd.md
23359 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
23360 (aarch64_frecpe<mode>): New.
23361 (aarch64_frecps<mode>): Likewise.
23362 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
23363 (v8type): Add frecp<esx>.
23364 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
23365 (aarch64_frecps<mode>): Likewise.
23366 * config/aarch64/iterators.md (FRECP): New.
23367 (frecp_suffix): Likewise.
23368 * config/aarch64/arm_neon.h
23369 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
23371 2013-04-22 Christian Bruel <christian.bruel@st.com>
23374 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
23375 (REG_CLASS_NAMES): Idem.
23376 (REG_CLASS_CONTENTS): Idem.
23377 (REGCLASS_HAS_FP_REG): Idem.
23378 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
23379 (sh_conditional_register_usage): Idem.
23381 2013-04-21 Jeff Law <law@redhat.com>
23383 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
23384 (ssa_forward_propagate_and_combine): Use it.
23386 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
23388 * lra.c: Update the flow chart diagram.
23390 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
23392 PR rtl-optimization/56847
23393 * lra-constraints.c (process_alt_operands): Discourage alternative
23394 with non-matche doffsettable memory constraint fro memory with
23397 2013-04-19 Richard Biener <rguenther@suse.de>
23399 PR tree-optimization/56982
23400 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
23402 * gimplify.c (gimplify_call_expr): Notice special calls.
23403 (gimplify_modify_expr): Likewise.
23404 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
23405 abnormal control flow receivers.
23406 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
23407 in the same way as cfun->has_nonlocal_labels.
23408 (gimple_purge_dead_abnormal_call_edges): Likewise.
23409 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
23410 receivers start a basic-block.
23412 2013-04-19 Richard Biener <rguenther@suse.de>
23414 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
23416 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
23417 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
23418 (SLP_TREE_LOAD_PERMUTATION): Add.
23419 (vect_transform_slp_perm_load): Adjust prototype.
23420 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
23421 (vect_free_slp_instance): Likewise.
23422 (vect_create_new_slp_node): Likewise.
23423 (vect_supported_slp_permutation_p): Remove.
23424 (vect_slp_rearrange_stmts): Adjust.
23425 (vect_supported_load_permutation_p): Likewise. Inline
23426 vect_supported_slp_permutation_p here.
23427 (vect_analyze_slp_instance): Compute load permutations per
23428 slp node instead of per instance.
23429 (vect_get_slp_defs): Adjust.
23430 (vect_transform_slp_perm_load): Likewise.
23431 (vect_schedule_slp_instance): Remove redundant code.
23432 (vect_schedule_slp): Remove hack for PR56270, add it ...
23433 * tree-vect-stmts.c (vectorizable_load): ... here, do not
23434 CSE loads for SLP. Adjust.
23436 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
23438 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
23439 spelling in two comments.
23441 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
23444 * config/arm/arm.c (load_multiple_sequence): Require SP
23445 as base register for loads if SP is in the register list.
23447 2013-04-19 Martin Jambor <mjambor@suse.cz>
23449 PR tree-optimization/56718
23450 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
23451 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
23452 and made public. Adjusted all callers.
23453 (ipa_intraprocedural_devirtualization): New function.
23454 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
23455 (ipa_intraprocedural_devirtualization): Likewise.
23456 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
23458 2013-04-19 Richard Biener <rguenther@suse.de>
23460 PR tree-optimization/57000
23461 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
23463 2013-04-19 Terry Guo <terry.guo@arm.com>
23465 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
23467 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
23468 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
23469 (cortex_m4_fmacs): Use new reservations.
23470 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
23472 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
23474 PR rtl-optimization/56999
23475 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
23477 (lra_coalesce): Remove split_origin_bitmap and related code.
23478 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
23479 ranges if necessary.
23481 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
23483 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
23485 (ix86_expand_call): Remove clobbered_registers array and use
23486 x86_64_ms_sysv_extra_clobbered_registers instead.
23487 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
23489 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
23491 * config/i386/i386.md (*call_rex64_ms_sysv): Use
23492 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
23493 (*call_value_rex64_ms_sysv): Ditto.
23495 2013-04-18 Cary Coutant <ccoutant@google.com>
23497 * dwarf2out.c (output_pubnames): Check die_perennial_p of
23498 parent instead of die_mark.
23500 2013-04-18 Diego Novillo <dnovillo@google.com>
23502 * gimple.c (create_gimple_tmp): New.
23503 (get_expr_type): New.
23504 (build_assign): New.
23505 (build_type_cast): New.
23506 * gimple.h (enum ssa_mode): Define.
23507 (gimple_seq_set_location): New.
23508 * asan.c (build_check_stmt): Change some gimple_build_* calls
23509 to use build_assign and build_type_cast.
23511 2013-04-18 Richard Biener <rguenther@suse.de>
23513 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
23514 handle negative step. Remove redundant checks.
23515 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
23516 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
23517 for negative step and grouped loads fail to vectorize.
23519 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23521 * emit-rtl.c (reset_insn_used_flags): New function.
23522 (reset_all_used_flags): Use it.
23523 (verify_insn_sharing): New function.
23524 (verify_rtl_sharing): Fix verification for SEQUENCEs.
23526 2013-04-18 Jakub Jelinek <jakub@redhat.com>
23528 PR tree-optimization/56984
23529 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
23530 and (x >> M) >= N don't register any assertion if N << M is the
23533 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23535 * lower-subreg.c (resolve_simple_move): If called self-recursive,
23536 do not delete_insn insns that have not yet been emitted, only
23537 unlink them with remove_insn.
23538 * df-scan.c (df_insn_delete): Revert r197492.
23540 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
23542 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
23543 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
23545 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23547 * config/arm/arm.md (movsicc_insn): Convert define_insn into
23548 define_insn_and_split.
23549 (and_scc,ior_scc,negscc): Likewise.
23550 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
23552 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23554 * config/arm/arm.c (use_return_insn): Return 0 for targets that
23555 can benefit from using a sequence of LDRD instructions in epilogue
23556 instead of a single LDM instruction.
23558 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
23561 * doc/extend.texi: Fix typo.
23563 2013-04-17 Richard Biener <rguenther@suse.de>
23565 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
23566 (vect_build_slp_tree): ... here.
23567 (vect_build_slp_tree_1): Compute which stmts of the SLP group
23568 match. Remove special-casing of mismatched complex loads.
23569 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
23570 re-try the match with swapped commutative operands.
23571 (vect_supported_load_permutation_p): Remove special-casing of
23572 mismatched complex loads.
23573 (vect_analyze_slp_instance): Adjust.
23575 2013-04-17 Richard Biener <rguenther@suse.de>
23577 PR rtl-optimization/56921
23578 * cfgloop.h (struct loop): Add simple_loop_desc member.
23579 (struct niter_desc): Mark with GTY(()).
23580 (simple_loop_desc): Do not use aux field but simple_loop_desc.
23581 * loop-iv.c (get_simple_loop_desc): Likewise.
23582 (free_simple_loop_desc): Likewise.
23585 2013-04-16 Richard Biener <rguenther@suse.de>
23587 PR rtl-optimization/56921
23588 * loop-init.c (pass_rtl_move_loop_invariants): Add
23589 TODO_do_not_ggc_collect to todo_flags_finish.
23590 (pass_rtl_unswitch): Same.
23591 (pass_rtl_unroll_and_peel_loops): Same.
23592 (pass_rtl_doloop): Same.
23594 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
23596 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
23597 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
23598 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
23600 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
23601 * tree-streamer.c (record_common_node): Adjust reference in comment.
23603 2013-04-17 Terry Guo <terry.guo@arm.com>
23605 * config/arm/cortex-m4.md: Add a new bypass.
23607 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23609 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
23611 (*subs_<optab><mode>_multp2): New pattern.
23612 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23613 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23615 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23617 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
23618 (*subs_mul_imm_<mode>): New pattern.
23620 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
23623 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
23624 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
23625 (vsx_movti_32bit): Change j->wa to O->wa.
23627 2013-04-16 Richard Biener <rguenther@suse.de>
23629 PR rtl-optimization/56921
23630 * loop-init.c (pass_rtl_move_loop_invariants): Add
23631 TODO_do_not_ggc_collect to todo_flags_finish.
23632 (pass_rtl_unswitch): Same.
23633 (pass_rtl_unroll_and_peel_loops): Same.
23634 (pass_rtl_doloop): Same.
23636 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
23638 * config/arm/arm.c (emit_multi_reg_push): New declaration
23639 for an existing function.
23640 (arm_emit_strd_push): New function.
23641 (arm_expand_prologue): Used here.
23642 (arm_emit_ldrd_pop): New function.
23643 (arm_expand_epilogue): Used here.
23644 (arm_get_frame_offsets): Update condition.
23645 (arm_emit_multi_reg_pop): Add a special case for load of a single
23646 register with writeback.
23648 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
23650 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
23653 2013-04-16 Richard Biener <rguenther@suse.de>
23655 PR tree-optimization/56756
23656 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
23657 (first_mem_ref_loc): New.
23658 (execute_sm): Place the load temporarily before a previous
23659 access instead of in the latch edge to ensure its SSA dependencies
23660 are defined at points dominating the load.
23662 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
23664 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
23665 correct fix by moving header and footer insn to the footer of
23666 the merged basic block. Clear BB_END of the merged-away block.
23668 PR middle-end/43631
23669 * emit-rtl.c (make_note_raw): New function.
23670 (link_insn_into_chain): New static inline function.
23671 (add_insn): Use it.
23672 (add_insn_before, add_insn_after): Factor insn chain linking code...
23673 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
23674 using link_insn_into_chain.
23675 (note_outside_basic_block_p): New helper function for emit_note_after
23676 and emit_note_before.
23677 (emit_note_after): Use nobb variant of add_insn_after if the note
23678 should not be contained in a basic block.
23679 (emit_note_before): Use nobb variant of add_insn_before if the note
23680 should not be contained in a basic block.
23681 (emit_note_copy): Use make_note_raw.
23682 (emit_note): Likewise.
23683 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
23684 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
23685 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
23686 the moved barrier the tail of the basic block it follows.
23687 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
23689 2013-04-15 Jakub Jelinek <jakub@redhat.com>
23691 PR tree-optimization/56962
23692 * gimple-ssa-strength-reduction.c (record_increment): Only set
23693 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
23694 either rhs1 or rhs2 is equal to c->base_expr.
23696 2013-04-15 Richard Biener <rguenther@suse.de>
23698 PR tree-optimization/56933
23699 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
23701 (GROUP_READ_WRITE_DEPENDENCE): Remove.
23702 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
23703 * tree-vect-data-refs.c (vect_analyze_group_access): Move
23704 dependence check ...
23705 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
23707 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
23708 GROUP_READ_WRITE_DEPENDENCE.
23710 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23712 * emit-rtl.c (reset_all_used_flags): New function.
23713 (verify_rtl_sharing): Call reset_all_used_flags before and after
23714 performing the checks.
23716 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23718 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
23719 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
23720 * config/arm/constraints.md (De): New constraint.
23721 * config/arm/neon.md (anddi3_neon): Delete.
23722 (neon_vand<mode>): Expand to standard anddi3 pattern.
23723 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
23724 Move earlier in the file.
23725 (neon_inv_logic_op2): Likewise.
23726 (arm_anddi_operand_neon): New predicate.
23728 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23730 * configure.ac (gcc_cv_ld_as_needed): Set
23731 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
23732 Use -z ignore, -z record on *-*-solaris2*.
23733 (HAVE_LD_AS_NEEDED): Update comment.
23734 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
23735 * configure: Regenerate.
23736 * config.in: Regenerate.
23737 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
23738 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
23739 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
23740 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
23741 equivalents. Fix markup.
23742 * doc/tm.texi: Regenerate.
23744 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
23746 * config/i386/i386.opt: New option mstack-protector-guard=.
23747 * config/i386/i386-opts.h: Add enum stack_protector_guard.
23748 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
23749 TARGET_SSP_TLS_GUARD.
23750 * config/i386/i386.c (ix86_option_override_internal): Set
23751 ix86_stack_protector_guard.
23752 * config/i386/i386.md (stack_protect_set): Enable for
23753 TARGET_SSP_TLS_GUARD only.
23754 (stack_protect_set_<mode>): Ditto.
23755 (stack_protect_test): Ditto.
23756 (stack_protect_test_<mode>): Ditto.
23757 * doc/invoke.texi (i386 Option): Document.
23759 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
23762 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
23763 (S_MODES): Set H_MODE bit.
23764 (SF_MODES): Set only S_MODE and SF_MODE bits.
23765 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
23766 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
23767 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
23768 <MODE_FLOAT>: Likewise.
23770 2013-04-15 Joey Ye <joey.ye@arm.com>
23772 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
23774 2013-04-15 Joey Ye <joey.ye@arm.com>
23776 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
23778 (thumb_far_jump_used_p): Count instruction size and set
23781 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
23783 * reorg.c (fill_simple_delay_slots): Reindent block of code.
23784 * resource.c (mark_target_live_regs): Reformat conditional block.
23786 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23788 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
23789 notes, they are emitted only just before final.
23790 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
23792 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23794 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
23795 * cfgrtl.c (delete_insn): Call it here instead.
23796 * lra-spills.c (lra_final_code_change): Use delete_insn.
23797 * haifa-sched.c (sched_remove_insn): Likewise.
23798 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
23799 returning to the nop pool.
23800 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
23801 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
23803 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23805 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
23806 * doc/tm.texi: Regenerated.
23808 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
23810 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
23813 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23815 * df-core.c (df_find_def): Compare register numbers.
23816 (df_find_use): Likewise.
23818 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
23821 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
23822 lra_in_progress for return.
23824 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23826 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
23827 define_insn into define_insn_and_split and emit movsicc patterns.
23829 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23831 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
23833 2013-04-12 Richard Biener <rguenther@suse.de>
23835 * tree-pass.h (TODO_do_not_ggc_collect): New.
23836 * passes.c (execute_one_ipa_transform_pass): Honor
23837 TODO_do_not_ggc_collect.
23838 (execute_one_pass): Likewise.
23841 2013-04-10 Richard Biener <rguenther@suse.de>
23843 * passes.c (init_optimization_passes): Remove reload pass.
23844 * ira.c (do_reload): Merge into ...
23846 (rest_of_handle_reload): Remove.
23847 (pass_reload): Likewise.
23848 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23849 of reload for vzeroupper pass placement.
23851 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23853 PR tree-optimization/56918
23854 PR tree-optimization/56920
23855 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23856 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23857 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23858 use wide_mul_with_sign method.
23860 2013-04-12 Richard Biener <rguenther@suse.de>
23862 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23863 not be considered a gimple constant.
23865 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23867 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23868 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23871 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23872 Jakub Jelinek <jakub@redhat.com>
23874 * opts.c: Include diagnostic-color.h.
23875 (common_handle_option): Handle OPT_fdiagnostics_color_.
23876 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23877 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23878 (diagnostic-color.o): New.
23879 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23880 (diagnostic_color_rule): New enum.
23881 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23882 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23883 the location string.
23884 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23885 either NULL, or color kind.
23886 * diagnostic-color.c: New file.
23887 * diagnostic-color.h: New file.
23888 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23890 * doc/invoke.texi (-fdiagnostics-color): Document.
23891 * pretty-print.h (pp_show_color): Define.
23892 (struct pretty_print_info): Add show_color field.
23893 * diagnostic.c: Include diagnostic-color.h.
23894 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23895 macros. Colorize error:, warning: etc. strings and also the location
23897 (diagnostic_show_locus): Colorize the caret line.
23898 * pretty-print.c: Include diagnostic-color.h.
23899 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23900 inside of %< %> quotes or quoted through q format modifier.
23902 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23904 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23906 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23908 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23909 code in CC_NZ mode.
23910 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23913 2013-04-11 Marek Polacek <polacek@redhat.com>
23915 PR tree-optimization/48184
23916 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23918 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23920 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23921 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23922 (skip_simple_arithmetic): Tidy up.
23923 * tree.h (skip_simple_constant_arithmetic): Declare.
23925 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23927 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23929 2013-04-11 Richard Biener <rguenther@suse.de>
23931 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23932 generate vector constants.
23934 2013-04-11 Richard Biener <rguenther@suse.de>
23936 PR tree-optimization/56878
23937 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23938 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23940 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23941 Prefer to align the DR with the most invariant base address.
23943 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23945 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23948 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23950 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23951 floating-point vector comparisons against 0.
23953 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23955 PR tree-optimization/56899
23956 * fold-const.c (extract_muldiv_1): Apply distributive law
23957 only if TYPE_OVERFLOW_WRAPS (ctype).
23959 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23962 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23963 loading has side effect.
23965 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23967 * config/sparc/sparc.c: Include tree-pass.h.
23968 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23969 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23970 head of file. Change return type. Split off gate function.
23971 (sparc_gate_work_around_errata): New function.
23972 (pass_work_around_errata): New pass definition.
23973 (insert_pass_work_around_errata) New pass insert definition to
23974 insert pass_work_around_errata just after delayed-branch scheduling.
23975 (sparc_option_override): Insert the pass.
23976 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23978 2013-04-10 David S. Miller <davem@davemloft.net>
23980 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23981 or -mcpu=hypersparc.
23983 * target.def (cstore_mode): New hook.
23984 * target.h: Include insn-codes.h
23985 * targhooks.c: Likewise.
23986 (default_cstore_mode): New function.
23987 * targhooks.h: Declare it.
23988 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23989 * doc/tm.texi: Rebuild.
23990 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23991 target hook, rather than inspecting the insn_data.
23992 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23993 (TARGET_CSTORE_MODE): Redefine.
23994 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
23996 * config/sparc/predicates.md (cstore_result_operand): New special
23998 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
23999 Use it for operand 0.
24000 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
24001 (*snesi_special): Likewise.
24002 (*snesi_zero): Likewise.
24003 (*seqsi_zero): Likewise.
24004 (*sltu_insn): Likewise.
24005 (*sgeu_insn): Likewise.
24006 (*seqdi_special): Make operand 0 and comparison operation be of
24008 (*snedi_special): Likewise.
24009 (*snedi_special_vis3): Likewise.
24010 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
24011 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
24012 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
24013 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
24014 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
24015 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
24016 (*sltu_extend_sp64): Likewise.
24017 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
24018 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
24019 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
24020 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
24021 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
24023 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
24025 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
24026 (aarch64_start_file): Use the new function.
24028 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
24029 Jason Merrill <jason@redhat.com>
24031 * common.opt: Add -gdwarf.
24032 * opts.c (common_handle_option): Handle it.
24033 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
24035 2013-04-10 Richard Biener <rguenther@suse.de>
24037 * passes.c (execute_todo): Do not call ggc_collect conditional here.
24038 (execute_one_ipa_transform_pass): But unconditionally here.
24039 (execute_one_pass): And here.
24040 (init_optimization_passes): Remove reload pass.
24041 * tree-pass.h (TODO_ggc_collect): Remove.
24042 (pass_reload): Likewise.
24043 * ira.c (do_reload): Merge into ...
24045 (rest_of_handle_reload): Remove.
24046 (pass_reload): Likewise.
24047 * config/i386/i386.c (ix86_option_override): Refer to ira instead
24048 of reload for vzeroupper pass placement.
24049 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
24050 and todo_flags_finish of all passes.
24052 2013-04-10 Richard Biener <rguenther@suse.de>
24054 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
24055 first_const_oprnd field, rename first_def_type to first_op_type.
24056 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
24057 (vect_get_and_check_slp_defs): Always use the type of the
24058 operand. Allow mixed vect_external_def, vect_constant_def types.
24059 (vect_get_constant_vectors): Handle mixed vect_external_def,
24060 vect_constant_def types.
24062 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
24064 PR tree-optimization/55524
24065 * tree-ssa-math-opts.c
24066 (convert_mult_to_fma): Don't use an fms construct
24067 when we don't have an fms operation, but fnma, and it looks
24068 likely that we'll be able to use the latter.
24070 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
24072 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
24074 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
24075 inline fail caused by overwritable functions.
24077 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
24079 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
24080 unnecessary bits in the constant power of two case.
24082 2013-04-10 Richard Biener <rguenther@suse.de>
24084 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
24085 broken code swapping operands.
24086 (vect_build_slp_tree): Do not compute load permutations here.
24087 (vect_analyze_slp_instance): Compute load permutations here,
24088 after building the SLP tree.
24090 2013-04-09 Christian Bruel <christian.bruel@st.com>
24092 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
24093 of next/prev_real_insn.
24095 2013-04-09 Jan Hubicka <jh@suse.cz>
24097 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
24098 Drop aliased parameter.
24099 (function_and_variable_visibility): Do not handle alias pairs.
24100 * cgraph.c (varpool_externally_visible_p): Update prototype.
24101 * varpool.c (varpool_add_new_variable): Update.
24103 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24105 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
24107 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
24109 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
24111 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
24113 2013-04-09 Marek Polacek <polacek@redhat.com>
24115 PR tree-optimization/48762
24116 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
24118 2013-04-09 Richard Biener <rguenther@suse.de>
24120 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
24122 (vect_build_slp_tree): Likewise.
24123 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
24124 calculating the cost of a SLP instance.
24125 (vect_analyze_slp_instance): Use it from here, after building
24128 2013-04-09 Jakub Jelinek <jakub@redhat.com>
24130 PR middle-end/56883
24131 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
24132 expand_omp_for_static_chunk): Use simple_p = true in
24133 force_gimple_operand_gsi calls when assigning to addressable decls.
24135 2013-04-09 Jeff Law <law@redhat.com>
24137 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
24138 when the boolean was created by converting a wider object which
24139 had a boolean range.
24141 2013-04-09 Richard Biener <rguenther@suse.de>
24143 * tree-vectorizer.h (slp_void_p): Remove.
24144 (slp_tree): Typedef before _slp_tree declaration.
24145 (struct _slp_tree): Use a vector of slp_tree as children.
24146 (vect_get_place_in_interleaving_chain): Remove.
24147 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
24149 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
24151 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
24152 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
24153 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
24154 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
24155 Use slp_node instead of slp_void_p and adjust.
24157 2013-04-09 Richard Biener <rguenther@suse.de>
24159 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
24160 work that is not necessary.
24162 2013-04-09 Jakub Jelinek <jakub@redhat.com>
24164 PR tree-optimization/56854
24165 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
24166 forward into clobber stmts if it would change MEM_REF lhs into
24169 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
24171 * tree.c (type_hash_lookup, type_hash_add): Make static.
24172 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
24174 2013-04-09 Richard Biener <rguenther@suse.de>
24176 * tree.h (unsave_expr_now): Remove.
24177 * tree-inline.c (mark_local_for_remap_r): Remove.
24178 (unsave_expr_1): Likewise.
24179 (unsave_r): Likewise.
24180 (unsave_expr_now): Likewise.
24181 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
24182 (propagate_tree_value): Likewise.
24184 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
24186 * doc/rtl.texi (sequence): Rewrite documentation to match the
24187 current use of SEQUENCE rtl objects.
24188 * rtl.def (SEQUENCE): Likewise.
24190 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
24191 Update documentation.
24192 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
24193 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
24195 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
24197 2013-04-08 Teresa Johnson <tejohnson@google.com>
24199 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
24200 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
24202 (estimate_edge_size_and_time): Add comment to suggest using rounding
24204 (estimate_node_size_and_time): Ditto.
24205 (remap_edge_change_prob): Use helper rounding divide methods.
24206 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
24207 (gimple_mod_pow2_value_transform): Ditto.
24208 (gimple_mod_subtract_transform): Ditto.
24209 (gimple_ic_transform): Ditto.
24210 (gimple_stringops_transform): Ditto.
24211 * stmt.c (conditional_probability): Ditto.
24212 (emit_case_dispatch_table): Ditto.
24213 * lto-cgraph.c (merge_profile_summaries): Ditto.
24214 * tree-optimize.c (execute_fixup_cfg): Ditto.
24215 * cfgcleanup.c (try_forward_edges): Ditto.
24216 * cfgloopmanip.c (scale_loop_profile): Ditto.
24218 (duplicate_loop_to_header_edge): Ditto.
24219 (lv_adjust_loop_entry_edge): Ditto.
24220 * tree-vect-loop.c (vect_transform_loop): Ditto.
24221 * profile.c (compute_branch_probabilities): Ditto.
24222 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
24223 * lto-streamer-in.c (input_cfg): Ditto.
24224 * gimple-streamer-in.c (input_bb): Ditto.
24225 * ipa-cp.c (update_profiling_info): Ditto.
24226 (update_specialized_profile): Ditto.
24227 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
24228 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
24230 * sched-rgn.c (compute_dom_prob_ps): Ditto.
24231 (compute_trg_info): Ditto.
24232 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
24233 (purge_dead_edges): Ditto.
24234 * loop-unswitch.c (unswitch_loop): Ditto.
24235 * cgraphclones.c (cgraph_clone_edge): Ditto.
24236 (cgraph_clone_node): Ditto.
24237 * tree-inline.c (copy_bb): Ditto.
24238 (copy_edges_for_bb): Ditto.
24239 (initialize_cfun): Ditto.
24240 (copy_cfg_body): Ditto.
24241 (expand_call_inline): Ditto.
24243 2013-04-08 Kai Tietz <ktietz@redhat.com>
24245 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
24246 TARGET_CYGWIN64 by TARGET_64BIT.
24248 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
24250 * config/epiphany/epiphany.md (GPR_1): New constant.
24251 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
24252 * config/epiphany/epiphany.c (gen_compare_reg):
24253 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
24254 is already in place.
24255 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
24256 Don't require being called during rtl expansion; If y operlaps r0,
24258 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
24259 (epiphany_expand_epilogue): Likewise.
24261 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
24262 Don't use CC_FPmode for ORDERED / UNORDERED.
24263 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
24265 * config/epiphany/constraints.md (CnL): New constraint.
24266 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
24267 * config/epiphany/predicates.md (add_operand): Allow 1024.
24269 * config/epiphany/epiphany.md (logical_op): New code iterator.
24270 (op_mnc): New code attribute.
24271 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
24272 (mov_f+1, mov_f+2): New peephole2 patterns.
24274 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
24275 (cstoresi4): Also allow re-use of zero result when doing a NE
24276 comparison to a non-zero operand.
24277 Use (clobber (scratch)) for first insn if the gpr output is not needed.
24279 * config/epiphany/epiphany.md (<insn_opname>v2si3):
24280 Use gen_addsi3_i / gen_subsi3_i.
24282 2013-04-08 Jakub Jelinek <jakub@redhat.com>
24286 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
24287 contain anything but clobbers, at most one __builtin_stack_restore,
24288 optionally debug stmts and final resx, and if it has at least one
24289 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
24290 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
24291 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
24292 which isn't defaut definition, remove them.
24293 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
24294 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
24295 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
24296 with MEM_REF LHS with SSA_NAME address.
24298 2013-04-08 Jeff Law <law@redhat.com>
24300 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
24302 2013-04-08 Richard Biener <rguenther@suse.de>
24304 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
24306 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
24307 determined vector type.
24308 (vect_analyze_data_refs): Likewise.
24309 (vect_get_new_vect_var): Adjust.
24310 (vect_create_destination_var): Preserve SSA name versions.
24311 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
24312 not dump anything here.
24314 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
24316 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
24317 Add member lr_slot_known.
24318 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
24320 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
24321 Remove code that sets lr_slot_offset according to what a previous
24322 version of epiphany_emit_save_restore used to do.
24323 (epiphany_emit_save_restore): When doing an lr save or restore,
24324 set/verify lr_slot_known and lr_slot_offset.
24326 2013-04-08 Xinyu Qi <xyqi@marvell.com>
24329 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
24332 2013-04-08 Richard Biener <rguenther@suse.de>
24334 * alias.c (find_base_term): Fix thinko in previous change.
24336 2013-04-08 Jakub Jelinek <jakub@redhat.com>
24338 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
24339 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
24340 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
24341 if possible to compute val.
24342 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
24343 For QImode integers don't require anything about precision. Use
24344 const_with_all_bytes_same to find out if the constant doesn't have
24345 repeated bytes in it.
24347 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24349 * config/s390/s390.c (s390_expand_insv): Only accept insertions
24352 2013-04-08 Marek Polacek <polacek@redhat.com>
24354 PR rtl-optimization/48182
24355 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
24358 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24361 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
24362 nuses, make sure we have a label.
24364 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24367 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
24368 (rs6000_emit_swdiv_low_precision): Remove.
24369 (rs6000_emit_swdiv): Rewrite to handle between one and four
24370 iterations of Newton-Raphson generally; modify required number of
24371 iterations for some cases.
24372 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
24374 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
24376 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
24377 set-but-unused variable.
24379 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
24380 basic blocks of released function bodies garbage-collectable.
24382 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
24383 (struct rtl_opt_pass): Add TODO_df_finish.
24385 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
24387 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24389 * config/arm/constraints.md (q): New constraint.
24390 * config/arm/ldrdstrd.md: New file.
24391 * config/arm/arm.md (ldrdstrd.md) New include.
24392 (arm_movdi): Use "q" instead of "r" constraint
24393 for double-word memory access.
24394 (movdf_soft_insn): Likewise.
24395 * config/arm/vfp.md (movdi_vfp): Likewise.
24396 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
24397 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
24398 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
24399 (mem_ok_for_ldrd_strd): Likewise.
24400 (output_move_double): Update assertion.
24402 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24404 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
24406 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24408 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
24409 define_insn_and_split.
24410 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
24412 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24414 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
24415 define_insn_and_split.
24416 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
24417 (shiftsi3_compare): New pattern.
24418 (rrx): New pattern.
24419 * config/arm/unspecs.md (UNSPEC_RRX): New.
24421 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24423 * config/arm/arm.md (negdi_extendsidi): New pattern.
24424 (negdi_zero_extendsidi): Likewise.
24426 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24428 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
24429 define_insn_and_split.
24430 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
24431 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
24433 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24435 * config/arm/arm.md (arm_subdi3): Convert define_insn into
24436 define_insn_and_split.
24437 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
24438 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
24440 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24442 * config/arm/arm.md (subsi3_carryin): New pattern.
24443 (subsi3_carryin_const): Likewise.
24444 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
24445 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
24447 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24449 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
24451 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24453 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
24454 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
24456 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24458 * config/arm/arm.c (arm_expand_builtin): Change fcode
24459 type to unsigned int.
24461 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24463 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
24465 2013-04-04 Ian Lance Taylor <iant@google.com>
24467 * doc/standards.texi (Standards): The Go frontend supports the Go 1
24470 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
24472 PR middle-end/56729
24473 * df-scan.c (df_insn_delete): Disable failing assert.
24475 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24477 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
24478 New function prototype.
24479 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
24480 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
24481 (arm_builtin_vectorized_function): New function.
24483 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24485 * config/arm/arm_neon_builtins.def: New file.
24486 * config/arm/arm.c (neon_builtin_data): Move contents to
24487 arm_neon_builtins.def.
24488 (enum arm_builtins): Include neon builtin definitions.
24489 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
24490 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
24492 2013-04-04 Marek Polacek <polacek@redhat.com>
24494 PR tree-optimization/48186
24495 * predict.c (maybe_hot_frequency_p): Return false if
24496 HOT_BB_FREQUENCY_FRACTION is 0.
24497 (cgraph_maybe_hot_edge_p): Likewise.
24499 2013-04-04 Richard Biener <rguenther@suse.de>
24501 PR tree-optimization/56826
24502 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
24505 2013-04-04 Richard Biener <rguenther@suse.de>
24507 PR tree-optimization/56213
24508 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
24509 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
24511 2013-04-04 Richard Biener <rguenther@suse.de>
24513 PR tree-optimization/56837
24514 * tree-loop-distribution.c (classify_partition): For non-zero
24515 values require that the value has the same precision as its
24516 mode to be useful as memset value.
24518 2013-04-03 Nick Clifton <nickc@redhat.com>
24520 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
24521 (fmssf4): Use fmsf.s on E3V5 architectures.
24522 (fnmasf4): Use fnmaf.s on E3V5 architectures.
24523 (fnmssf4): Use fnmsf.s on E3V5 architectures.
24525 2013-04-03 Jeff Law <law@redhat.com>
24527 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
24528 (lra-eliminations.o): Likewise.
24530 2013-04-03 Teresa Johnson <tejohnson@google.com>
24532 * gcov-io.c (compute_working_sets): Moved most of body of old
24533 compute_working_sets here from profile.c.
24534 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
24535 (gcov_working_set_t): Moved typedef here from basic-block.h
24536 (compute_working_set): Declare.
24537 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
24538 (get_working_sets): Renamed from compute_working_set,
24539 replace most of body with call to new compute_working_sets.
24540 (get_exec_counts): Replace call to compute_working_sets
24541 to get_working_sets.
24542 * profile.h (get_working_sets): Renamed from compute_working_set.
24543 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
24544 to get_working_sets.
24545 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
24546 * gcov-dump.c (dump_working_sets): New function.
24548 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
24550 * hwint.c (sext_hwi, zext_hwi): New functions.
24551 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
24552 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
24553 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
24554 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
24555 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
24556 (sext_hwi, zext_hwi): New functions.
24558 2013-04-03 Jeff Law <law@redhat.com>
24560 PR tree-optimization/56799
24561 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
24562 back test for widening conversion erroneously dropped in prior change.
24564 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24567 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
24568 instead of next_real_insn.
24570 2013-04-03 Marek Polacek <polacek@redhat.com>
24573 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
24575 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24578 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
24580 (thumb1_output_casesi): Likewise.
24581 (thumb2_output_casesi): Likewise.
24583 2013-04-03 Richard Biener <rguenther@suse.de>
24585 PR tree-optimization/56817
24586 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
24588 (tree_unroll_loops_completely_1): ... new function to manually
24589 walk the loop tree, properly defering outer loops of unrolled
24590 loops to later iterations.
24592 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24594 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
24595 (vectorizable_load): Likewise.
24596 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
24597 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
24599 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24601 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
24604 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
24606 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
24608 2013-04-03 Bin Cheng <bin.cheng@arm.com>
24610 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
24612 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24614 PR tree-optimization/56790
24615 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
24618 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24620 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
24622 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
24623 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
24626 2013-04-03 Jakub Jelinek <jakub@redhat.com>
24629 * tree.h (force_folding_builtin_constant_p): New decl.
24630 * builtins.c (force_folding_builtin_constant_p): New variable.
24631 (fold_builtin_constant_p): Fold immediately also if
24632 force_folding_builtin_constant_p.
24634 2013-04-03 Richard Biener <rguenther@suse.de>
24636 PR tree-optimization/56812
24637 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
24638 DRs of the same interleaving chain are independent.
24640 2013-04-02 Jason Merrill <jason@redhat.com>
24642 * gdbinit.in (pbb): Use debug fn.
24644 2013-04-02 Lawrence Crowl <crowl@google.com>
24646 * sese.h (struct ivtype_map_elt_s): Remove unused.
24647 (extern debug_ivtype_map): Remove unused.
24648 (extern eq_ivtype_map_elts): Remove unused.
24649 * sese.c (debug_ivtype_map): Removed unused.
24650 (debug_ivtype_map_1): Removed unused.
24651 (debug_ivtype_elt): Remove unused.
24652 (eq_ivtype_map_elts): Remove unused.
24654 2013-04-02 Kai Tietz <ktietz@redhat.com>
24657 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
24658 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
24659 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
24661 (legitimize_pe_coff_symbol): Likewise.
24662 (is_imported_p): New helper-function.
24663 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
24664 for Windows x64 targets.
24665 (ix86_expand_prologue): Optimize for pe-coff targets.
24666 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
24667 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
24668 medium/large code-model.
24669 (legitimize_pic_address): Likewise.
24670 (legitimize_tls_address): Likewise.
24671 (ix86_expand_call): Likewise.
24672 (x86_output_mi_thunk): Likewise.
24673 (get_dllimport_decl): Add new beimport argument.
24674 (construct_plt_address): Don't assert for x64 pe-coff targets.
24675 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
24677 (SYMBOL_FLAG_STUBVAR): New macro.
24678 (SYMBOL_REF_STUBVAR_P): Likewise.
24679 * config/i386/winnt.c (stub_list): New structure.
24680 (stub_head): New local variable.
24681 (i386_pe_record_stub): New function.
24682 (i386_pe_file_end): Emit refptr-stubs.
24684 2013-04-02 Jakub Jelinek <jakub@redhat.com>
24686 PR rtl-optimization/56745
24687 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
24688 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
24691 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
24692 and both of them are MEM_REFs, just compare first argument for
24693 equality and attempt to deal even with differing offsets.
24696 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
24697 of gimple_clobber_p to be MEM_REF.
24698 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
24699 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
24700 after gimplification.
24701 * asan.c (get_mem_ref_of_assignment): Don't instrument
24702 gimple_clobber_p stmts.
24703 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
24704 gimple_clobber_p stmt if they have MEM_REF lhs and
24705 are dead because of another gimple_clobber_p stmt.
24706 * tree-ssa-live.c (clear_unused_block_pointer): Treat
24707 gimple_clobber_p stmts like debug stmts.
24708 (remove_unused_locals): Remove clobbers with MEM_REF lhs
24709 that refer to unused VAR_DECLs or uninitialized values.
24710 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
24711 gimple_clobber_p stmts if they refer to removed parameters.
24712 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
24715 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
24717 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
24718 using SWI48 mode attribute.
24720 2013-04-02 Wei Mi <wmi@google.com>
24722 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
24723 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
24724 *<rotate_insn><mode>3_mask in i386.md.
24726 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
24728 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
24730 2013-04-02 Richard Biener <rguenther@suse.de>
24732 PR tree-optimization/56778
24733 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
24734 Runtime alias tests are not supported for gather loads.
24735 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
24736 stmts referenced from SSA operands before updating SSA form.
24738 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
24739 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24741 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
24742 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
24743 * config/arm/cortex-a53.md: New file.
24744 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
24745 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
24746 * config/arm/arm.c (arm_issue_rate): Likewise.
24747 * config/arm/arm-tune.md: Regenerate
24748 * config/arm/arm-tables.opt: Regenerate.
24749 * config/arm/arm-cores.def: Add cortex-a53.
24751 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
24753 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
24756 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24758 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
24759 scalar load/store operations using B/H registers.
24760 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
24762 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24764 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
24766 * config/aarch64/aarch64.c
24767 (aarch64_simd_scalar_immediate_valid_for_move): New.
24768 * config/aarch64/aarch64-protos.h
24769 (aarch64_simd_scalar_immediate_valid_for_move): New.
24770 * config/aarch64/constraints.md (Dh, Dq): New.
24771 * config/aarch64/iterators.md (hq): New.
24773 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
24775 * reorg.c (get_branch_condition): Deal with conditional returns.
24776 (fill_simple_delay_slots): Remove dead code dealing with jumps.
24778 2013-04-01 Wei Mi <wmi@google.com>
24780 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
24781 Truncate operand 2 using %b asm operand modifier.
24782 (*<shift_insn><mode>3_mask): Ditto.
24783 (*<rotate_insn><mode>3_mask): Ditto.
24785 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
24787 PR middle-end/56798
24788 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
24790 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
24792 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
24794 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
24796 2013-03-30 Lawrence Crowl <crowl@google.com>
24798 * dse.c (clear_alias_sets): Remove never set.
24799 (disqualified_clear_alias_sets): Remove never set.
24800 (clear_alias_mode_pool): Remove never set.
24801 (dse_step0): Remove condition that is never true.
24802 (canon_address): Remove condition that is never true.
24803 (dse_step7): Remove condition that is never true.
24804 (rest_of_handle_dse): Remove condition that is never true.
24805 (rest_of_handle_dse::did_global): Remove never read from above.
24806 (dse_step2_spill): Remove never called from above.
24807 (dse_step5_spill): Remove never called from above.
24809 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
24811 * doc/md.texi (Standard Names) <casesi>: Update documentation for
24812 JUMP_TABLE_DATA changes.
24813 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
24814 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
24815 (Insns) <jump_table_data>: New entry.
24816 * doc/tm.texi: Regenerate.
24818 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
24820 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
24821 for table jump at the end of a basic block using tablejump_p.
24822 * targhooks.c (default_invalid_within_doloop): Likewise.
24823 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
24824 target hook implementation that is identical to the default hook.
24825 (rs6000_invalid_within_doloop): Remove.
24827 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
24828 unused variable from tablejump_p call.
24830 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
24831 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
24832 (INSN_DELETED_P): Likewise.
24833 (emit_jump_table_data): New prototype.
24834 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
24835 after 4th as unused.
24836 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
24837 * sched-vis.c (print_insn): Likewise.
24838 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
24839 insn for compatibility with back ends that use next_active_insn to
24840 identify jump table data.
24841 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
24842 (remove_insn): Likewise.
24843 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
24845 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
24846 (emit_jump_table_data): New function.
24848 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24849 basic block, a JUMP_TABLE_DATA never is.
24850 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24851 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24852 off from code handling real insns.
24853 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24854 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24855 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24856 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24857 is not a NONDEBUG_INSN_P.
24858 * ira-costs.c (scan_one_insn): Likewise.
24859 * jump.c (mark_all_labels): Likewise.
24860 (mark_jump_label_1): Likewise.
24861 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24862 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24864 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24865 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24866 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24867 JUMP_TABLE_DATA_P insns.
24868 (calculate_elim_costs_all_insns): Likewise.
24869 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24870 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24871 (delete_output_reload): Code style fixups.
24872 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24873 insn flags on this non-insn.
24874 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24875 as scheduling barriers, for pre-change compatibility.
24876 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24877 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24879 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24880 redundant JUMP_TABLE_DATA_P test.
24881 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24882 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24883 (frv_for_each_packet): Likewise.
24884 * config/i386/i386.c (min_insn_size): Likewise.
24885 (ix86_avoid_jump_mispredicts): Likewise.
24886 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24887 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24888 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24889 (mips16_insn_length): Robustify.
24890 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24891 (mips16_split_long_branches): Likewise.
24892 * config/pa/pa.c (pa_combine_instructions): Likewise.
24893 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24894 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24895 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24896 as contributing to pool range lengths.
24897 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24898 Remove redundant JUMP_TABLE_DATA_P test.
24899 (sh_loop_align): Likewise.
24900 (split_branches): Likewise.
24901 (sh_insn_length_adjustment): Likewise.
24902 * config/spu/spu.c (get_branch_target): Likewise.
24904 2013-03-29 Jan Hubicka <jh@suse.cz>
24906 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24907 gcov streaming; stream hot bb threshold to ltrans.
24908 * predict.c (get_hot_bb_threshold): Break out from ....
24909 (maybe_hot_count_p): ... here.
24910 (set_hot_bb_threshold): New function.
24911 * lto-section-in.c (lto_section_name): Add profile.
24912 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24913 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24914 and data-streamer.h
24915 (histogram_entry): New structure.
24916 (histogram, histogram_pool): New global vars.
24917 (histogram_hash): New structure.
24918 (histogram_hash::hash): New method.
24919 (histogram_hash::equal): Likewise.
24920 (account_time_size): New function.
24921 (cmp_counts): New function.
24922 (dump_histogram): New function.
24923 (ipa_profile_generate_summary): New function.
24924 (ipa_profile_write_summary): New function.
24925 (ipa_profile_read_summary): New function.
24926 (ipa_profile): Decide on threshold.
24927 (pass_ipa_profile): Add ipa_profile_write_summary and
24928 ipa_profile_read_summary.
24929 * Makefile.in (ipa.o): Update dependencies.
24930 * lto-streamer.h (LTO_section_ipa_profile): New section.
24932 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24934 * tree.h (VAR_P): New.
24936 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24939 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24941 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24943 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24944 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24945 (control_flow_insn_p): Likewise.
24946 * cfgrtl.c (duplicate_insn_chain): Likewise.
24947 * final.c (get_attr_length_1): Likewise.
24948 (shorten_branches): Likewise.
24949 (final_scan_insn): Likewise.
24950 * function.c (instantiate_virtual_regs): Likewise.
24951 * gcse.c (insert_insn_end_basic_block): Likewise.
24952 * ira-costs.c (scan_one_insn): Likewise.
24953 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24954 * lra.c (check_rtl): Likewise.
24955 * reload1.c (elimination_costs_in_insn): Likewise.
24956 * reorg.c (follow_jumps): Likewise.
24958 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24959 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24960 (thumb_far_jump_used_p): Likewise.
24961 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24962 (workaround_speculation): Likewise.
24963 (add_sched_insns_for_speculation): Likewise.
24964 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24965 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24966 (frv_for_each_packet): Likewise.
24967 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24968 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24969 (final_emit_insn_group_barriers): Likewise.
24970 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24971 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24972 (mips16_insn_length): Likewise.
24973 * config/pa/pa.c (pa_reorg): Likewise.
24974 (pa_combine_instructions): Likewise.
24975 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24976 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24977 (sh_reorg): Likewise.
24978 (split_branches): Likewise.
24979 * config/spu/spu.c (get_branch_target): Likewise.
24981 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24984 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24986 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24987 Fix declaration name.
24989 2013-03-28 Lawrence Crowl <crowl@google.com>
24991 * graphds.h (struct graph.indicies): Remove unused.
24992 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24993 (SCOP_ORIGINAL_PDDRS): Remove unused.
24994 * sese.h (extern insert_loop_close_phis): Removed unused.
24995 (extern insert_guard_phis): Removed unused.
24996 (extern ivtype_map_elt_info): Removed unused.
24997 (new_ivtype_map_elt): Removed unused.
24998 * sese.c (ivtype_map_elt_info): Removed unused.
25000 2013-03-28 Lawrence Crowl <crowl@google.com>
25002 * Makefile.in: Add several missing include dependences.
25004 (test-dump.o): New. This object is not added to any executable,
25005 but is present for ad-hoc testing.
25007 (debug (const bitmap_head_def &)): New.
25008 (debug (const bitmap_head_def *)): New.
25010 (extern debug (const bitmap_head_def &)): New.
25011 (extern debug (const bitmap_head_def *)): New.
25013 (debug (edge_def &)): New.
25014 (debug (edge_def *)): New.
25016 (debug (basic_block_def &)): New.
25017 (debug (basic_block_def *)): New.
25019 (dump_node (const_tree, int, FILE *)): Correct source file.
25021 (debug (die_struct &)): New.
25022 (debug (die_struct *)): New.
25024 (extern debug (die_struct &)): New.
25025 (extern debug (die_struct *)): New.
25026 * gimple-pretty-print.c
25027 (debug (gimple_statement_d &)): New.
25028 (debug (gimple_statement_d *)): New.
25029 * gimple-pretty-print.h
25030 (extern debug (gimple_statement_d &)): New.
25031 (extern debug (gimple_statement_d *)): New.
25033 (debug (ira_allocno_copy &)): New.
25034 (debug (ira_allocno_copy *)): New.
25035 (debug (ira_allocno &)): New.
25036 (debug (ira_allocno *)): New.
25038 (extern debug (ira_allocno_copy &)): New.
25039 (extern debug (ira_allocno_copy *)): New.
25040 (extern debug (ira_allocno &)): New.
25041 (extern debug (ira_allocno *)): New.
25043 (debug (live_range &)): New.
25044 (debug (live_range *)): New.
25046 (debug (lra_live_range &)): New.
25047 (debug (lra_live_range *)): New.
25049 (debug (lra_live_range &)): New.
25050 (debug (lra_live_range *)): New.
25052 (debug (omega_pb_d &)): New.
25053 (debug (omega_pb_d *)): New.
25055 (extern debug (omega_pb_d &)): New.
25056 (extern debug (omega_pb_d *)): New.
25058 (debug (const rtx_def &)): New.
25059 (debug (const rtx_def *)): New.
25061 (debug_tree (tree): Move within file.
25062 (debug_raw (const tree_node &)): New.
25063 (debug_raw (const tree_node *)): New.
25064 (dump_tree_via_hooks (const tree_node *, int)): New.
25065 (debug (const tree_node &)): New.
25066 (debug (const tree_node *)): New.
25067 (debug_verbose (const tree_node &)): New.
25068 (debug_verbose (const tree_node *)): New.
25069 (debug_head (const tree_node &)): New.
25070 (debug_head (const tree_node *)): New.
25071 (debug_body (const tree_node &)): New.
25072 (debug_body (const tree_node *)): New.
25073 (debug_vec_tree (tree): Move and reimplement in terms of dump.
25074 (debug (vec<tree, va_gc> &)): New.
25075 (debug (vec<tree, va_gc> *)): New.
25077 (extern debug (const rtx_def &)): New.
25078 (extern debug (const rtx_def *)): New.
25080 (debug_raw (simple_bitmap_def &)): New.
25081 (debug_raw (simple_bitmap_def *)): New.
25082 (debug (simple_bitmap_def &)): New.
25083 (debug (simple_bitmap_def *)): New.
25085 (extern debug (simple_bitmap_def &)): New.
25086 (extern debug (simple_bitmap_def *)): New.
25087 (extern debug_raw (simple_bitmap_def &)): New.
25088 (extern debug_raw (simple_bitmap_def *)): New.
25090 (debug (vinsn_def &)): New.
25091 (debug (vinsn_def *)): New.
25092 (debug_verbose (vinsn_def &)): New.
25093 (debug_verbose (vinsn_def *)): New.
25094 (debug (expr_def &)): New.
25095 (debug (expr_def *)): New.
25096 (debug_verbose (expr_def &)): New.
25097 (debug_verbose (expr_def *)): New.
25098 (debug (vec<rtx> &)): New.
25099 (debug (vec<rtx> *)): New.
25101 (extern debug (vinsn_def &)): New.
25102 (extern debug (vinsn_def *)): New.
25103 (extern debug_verbose (vinsn_def &)): New.
25104 (extern debug_verbose (vinsn_def *)): New.
25105 (extern debug (expr_def &)): New.
25106 (extern debug (expr_def *)): New.
25107 (extern debug_verbose (expr_def &)): New.
25108 (extern debug_verbose (expr_def *)): New.
25109 (extern debug (vec<rtx> &)): New.
25110 (extern debug (vec<rtx> *)): New.
25112 (_list_iter_cond_expr): Make inline instead of static.
25114 (debug (sreal &)): New.
25115 (debug (sreal *)): New.
25117 (extern debug (sreal &)): New.
25118 (extern debug (sreal *)): New.
25120 (extern debug_raw (const tree_node &)): New.
25121 (extern debug_raw (const tree_node *)): New.
25122 (extern debug (const tree_node &)): New.
25123 (extern debug (const tree_node *)): New.
25124 (extern debug_verbose (const tree_node &)): New.
25125 (extern debug_verbose (const tree_node *)): New.
25126 (extern debug_head (const tree_node &)): New.
25127 (extern debug_head (const tree_node *)): New.
25128 (extern debug_body (const tree_node &)): New.
25129 (extern debug_body (const tree_node *)): New.
25130 (extern debug (vec<tree, va_gc> &)): New.
25131 (extern debug (vec<tree, va_gc> *)): New.
25133 (debug (struct loop &)): New.
25134 (debug (struct loop *)): New.
25135 (debug_verbose (struct loop &)): New.
25136 (debug_verbose (struct loop *)): New.
25137 * tree-dump.c: Add header dependence.
25139 (extern debug (struct loop &)): New.
25140 (extern debug (struct loop *)): New.
25141 (extern debug_verbose (struct loop &)): New.
25142 (extern debug_verbose (struct loop *)): New.
25144 (debug (data_reference &)): New.
25145 (debug (data_reference *)): New.
25146 (debug (vec<data_reference_p> &)): New.
25147 (debug (vec<data_reference_p> *)): New.
25148 (debug (vec<ddr_p> &)): New.
25149 (debug (vec<ddr_p> *)): New.
25151 (extern debug (data_reference &)): New.
25152 (extern debug (data_reference *)): New.
25153 (extern debug (vec<data_reference_p> &)): New.
25154 (extern debug (vec<data_reference_p> *)): New.
25155 (extern debug (vec<ddr_p> &)): New.
25156 (extern debug (vec<ddr_p> *)): New.
25158 (debug (pt_solution &)): New.
25159 (debug (pt_solution *)): New.
25161 (extern debug (pt_solution &)): New.
25162 (extern debug (pt_solution *)): New.
25164 (debug (_var_map &)): New.
25165 (debug (_var_map *)): New.
25166 (debug (tree_live_info_d &)): New.
25167 (debug (tree_live_info_d *)): New.
25169 (extern debug (_var_map &)): New.
25170 (extern debug (_var_map *)): New.
25171 (extern debug (tree_live_info_d &)): New.
25172 (extern debug (tree_live_info_d *)): New.
25174 2013-03-28 Jan Hubicka <jh@suse.cz>
25176 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
25178 2013-03-28 Ian Bolton <ian.bolton@arm.com>
25180 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
25181 record only when desired or required.
25183 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
25185 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
25186 *vec_extractv2di_1_rex64. Use x64 isa attribute.
25188 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25190 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
25191 (*andsi3_compare0_uxtw): New pattern.
25192 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
25193 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
25195 2013-03-28 Jan Hubicka <jh@suse.cz>
25197 * data-streamer-in.c (streamer_read_gcov_count): New function.
25198 * gimple-streamer-out.c: Include value-prof.h.
25199 (output_gimple_stmt): Output histogram.
25200 (output_bb): Use streamer_write_gcov_count.
25201 * value-prof.c: Include data-streamer.h
25202 (dump_histogram_value): Add HIST_TYPE_MAX.
25203 (stream_out_histogram_value): New function.
25204 (stream_in_histogram_value): New function.
25205 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
25206 (stream_out_histogram_value, stream_in_histogram_value): Declare.
25207 * data-streamer-out.c (streamer_write_gcov_count): New function.
25208 (streamer_write_gcov_count_stream): New function.
25209 * lto-cgraph.c (lto_output_edge): Update counter streaming.
25210 (lto_output_node): Likewise.
25211 (input_node, input_edge): Likewise.
25212 * lto-streamer-out.c (output_cfg): Update streaming.
25213 * lto-streamer-in.c (input_cfg): Likewise.
25214 * data-streamer.h (streamer_write_gcov_count,
25215 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
25216 * gimple-streamer-in.c: Include value-prof.h
25217 (input_gimple_stmt): Input histograms.
25218 (input_bb): Update profile streaming.
25220 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
25222 * genmodes.c (emit_max_int): New function.
25223 (emit_insn_modes_h): Added call to emit_max_function.
25224 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
25226 * machmode.def: Fixed comment.
25228 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
25230 * combine.c (try_combine): Removed useless assert.
25231 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
25233 2013-03-28 Marek Polacek <polacek@redhat.com>
25234 Richard Biener <rguenther@suse.de>
25236 PR tree-optimization/56695
25237 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
25238 build signed result of a vector comparison.
25239 * tree-cfg.c (verify_gimple_comparison): Check that a result
25240 of a vector comparison has signed type.
25242 2013-03-28 Richard Biener <rguenther@suse.de>
25244 PR tree-optimization/37021
25245 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
25246 do not restrict gaps between groups.
25247 * tree-vect-stmts.c (vectorizable_load): Properly account for
25248 a gap between groups.
25250 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
25252 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
25253 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
25256 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
25258 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
25259 * doc/extend.texi (Named Address Spaces): Ditto.
25260 (Variable Attributes): Ditto.
25262 2013-03-27 Kai Tietz <ktietz@redhat.com>
25264 * config.build: Add support for cygwin x64 target.
25265 * config.gcc: Likewise.
25266 * config.host: Likewise.
25267 * configure.ac: Likewise
25268 * configure: Regenerated.
25270 2013-03-27 Kai Tietz <ktietz@redhat.com>
25272 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
25273 * config/i386/t-cygwin-w64: New file.
25274 * config/i386/cygwin-w64.h: New file.
25275 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
25276 and add support for x64-cygwin target.
25277 (CPP_SPEC): Likewise.
25278 (CXX_WRAP_SPEC_LIST): Undefine before define.
25279 (LIBGCJ_SONAME): Use 15 as version.
25281 2013-03-27 Richard Biener <rguenther@suse.de>
25283 PR tree-optimization/56716
25284 * tree-ssa-structalias.c (perform_var_substitution): Adjust
25285 dumping for ref nodes.
25287 2013-03-27 Martin Jambor <mjambor@suse.cz>
25289 PR tree-optimization/55334
25290 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
25291 restricted pointers to arrays.
25293 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
25295 * Makefile.in (.SUFFIXES): Add .cc.
25296 (.c.o): Apply same recipe for implicit rule .cc.o.
25298 2013-03-27 Richard Biener <rguenther@suse.de>
25300 PR tree-optimization/37021
25301 * tree-vect-data-refs.c (vect_check_strided_load): Allow
25302 REALPART/IMAGPART_EXPRs around the supported refs.
25303 * tree-ssa-structalias.c (find_func_aliases): Assume that
25304 floating-point values are not used to transfer pointers.
25306 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
25308 * target.def (TARGET_HAS_IFUNC_P): New target hook.
25309 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
25310 * doc/tm.texi: Regenerate.
25311 * targhooks.h (default_has_ifunc_p): New.
25312 * targhooks.c (default_has_ifunc_p): Ditto.
25313 * config/linux-protos.h: New file.
25314 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
25315 hook for linux which disables support of indirect functions in android.
25316 * config/linux-android.c: New file.
25317 * config/t-linux-android.c: Ditto.
25318 * config.gcc: Added new object file linux-android.o.
25319 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
25320 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
25321 * varasm.c (do_assemble_alias): Likewise.
25322 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
25323 doesn't support indirect functions.
25324 * configure: Regenerate.
25326 2013-03-27 Bin Cheng <bin.cheng@arm.com>
25329 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
25330 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
25333 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25335 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
25337 2013-03-27 Terry Guo <terry.guo@arm.com>
25339 * config/arm/arm-cores.def: Added core cortex-r7.
25340 * config/arm/arm-tune.md: Regenerated.
25341 * config/arm/arm-tables.opt: Regenerated.
25342 * doc/invoke.texi: Added entry for core cortex-r7.
25344 2013-03-27 Walter Lee <walt@tilera.com>
25346 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
25347 double-decrement of next_scratch_regno.
25349 2013-03-27 Walter Lee <walt@tilera.com>
25351 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
25353 (insn_v1mulus): Ditto.
25354 (insn_v2muls): Ditto.
25356 2013-03-27 Walter Lee <walt@tilera.com>
25358 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
25359 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
25361 2013-03-27 Walter Lee <walt@tilera.com>
25363 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
25364 (*sibcall_value): Ditto.
25366 2013-03-27 Walter Lee <walt@tilera.com>
25368 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
25369 (insn_mnz_v8qi): ... this ...
25370 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
25372 (insn_v<n>mnz): Replaced by ...
25373 (insn_v1mnz): ... this ...
25374 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
25376 (insn_mz_<mode>): Replaced by ...
25377 (insn_mz_v8qi): ... this ...
25378 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
25380 (insn_v<n>mz): Replaced by ...
25381 (insn_v1mz): ... this ...
25382 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
25385 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
25387 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
25389 2013-03-26 Roland McGrath <mcgrathr@google.com>
25391 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
25392 than fprintf with a non-constant, non-format string.
25394 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
25396 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
25397 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
25398 operand 0 predicate.
25399 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
25400 attribute. Use general_x64nomem_operand as operand 1 predicate.
25401 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
25402 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
25403 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
25404 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
25405 (mov<mode>_insv_1): Remove expander. Merge insn with
25406 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
25407 Use general_x64nomem_operand as operand 1 predicate.
25408 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
25409 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
25410 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
25411 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
25412 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
25413 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
25414 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
25415 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
25416 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
25417 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
25418 (general_x64nomem_operand): Ditto.
25420 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
25422 * config/rtems.opt: Add -pthread option.
25424 2013-03-26 Richard Biener <rguenther@suse.de>
25426 * alias.c (find_base_term): Avoid redundant and not used recursion.
25427 (base_alias_check): Get the initial base term from the caller.
25428 (true_dependence_1): Compute and pass base terms to base_alias_check.
25429 (write_dependence_p): Likewise.
25430 (may_alias_p): Likewise.
25432 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
25434 * config/aarch64/aarch64.c (aarch64_classify_address): Support
25435 PC-relative load in SI modes and above only.
25437 2013-03-26 Xinyu Qi <xyqi@marvell.com>
25439 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
25440 * config/arm/iwmmxt.md (WCGR0): Update.
25441 (WCGR1, WCGR2, WCGR3): Likewise.
25443 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
25445 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
25446 Use x64 and nox64 isa attributes.
25448 2013-03-26 Richard Biener <rguenther@suse.de>
25450 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
25451 alignment computations and rely on get_object_alignment_1
25452 for the !TYPE_P case.
25453 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
25455 2013-03-26 Walter Lee <walt@tilera.com>
25457 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
25458 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
25460 2013-03-25 Jeff Law <law@redhat.com>
25462 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
25463 check for INTEGRAL_TYPE_P that was missing due to checking in
25464 wrong version of prior patch.
25466 2013-03-25 Walter Lee <walt@tilera.com>
25468 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
25469 TILEGX_INSN_SHUFFLEBYTES1.
25470 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
25472 (tilegx_builtins): Ditto.
25473 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
25475 2013-03-25 Walter Lee <walt@tilera.com>
25477 * config/tilegx/tilegx.md (floatsisf2): New pattern.
25478 (floatunssisf2): New pattern.
25479 (floatsidf2): New pattern.
25480 (floatunssidf2): New pattern.
25482 2013-03-25 Walter Lee <walt@tilera.com>
25484 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
25485 tests for constraint J, K, N, P.
25487 2013-03-25 Walter Lee <walt@tilera.com>
25489 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
25490 Use indirect/pcrel encoding.
25491 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
25494 2013-03-25 Steve Ellcey <sellcey@mips.com>
25496 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
25497 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
25498 * config/mips/mips.c (mips_option_override): Set IMADD default.
25499 * config/mips/mips.h (PTF_AVOID_IMADD): New.
25500 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
25501 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
25502 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
25503 * doc/invoke.texi (-mimadd/-mno-imadd): New.
25505 2013-03-25 Jeff Law <law@redhat.com>
25507 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
25508 slightly to avoid creating and folding useless trees. Simplify
25509 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
25511 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
25513 * config/i386/i386.md (*zero_extendsidi2): Merge with
25514 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
25515 * config/i386/predicates.md (x86_64_zext_operand): Rename from
25516 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
25517 targets. Clarify comment.
25519 2013-03-25 Martin Jambor <mjambor@suse.cz>
25521 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
25522 pass-through jump functions differently.
25523 (ipa_read_jump_function): Likewise. Also use setter functions to set
25526 2013-03-25 Martin Jambor <mjambor@suse.cz>
25528 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
25529 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
25531 (ipa_get_indirect_edge_target): New function.
25532 (devirtualization_time_bonus): New parameter known_aggs, pass it to
25533 ipa_get_indirect_edge_target. Update all callers.
25534 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
25535 ipa_get_indirect_edge_target_1 instead of calling
25536 ipa_get_indirect_edge_target.
25537 (create_specialized_node): Pass aggvlas to
25538 ipcp_discover_new_direct_edges.
25540 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25542 * config/arm/arm.md (f_sels, f_seld): New types.
25543 (*cmov<mode>): New pattern.
25544 * config/arm/predicates.md (arm_vsel_comparison_operator): New
25547 2013-03-25 Kai Tietz <ktietz@redhat.com>
25549 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
25550 POSIX-printf for mingw-hosted builds.
25552 2013-03-25 Richard Biener <rguenther@suse.de>
25554 PR middle-end/56694
25555 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
25556 must-not-throw stmt location.
25558 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25560 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
25561 Emit load-acquire versions when acq is true.
25562 (arm_emit_store_exclusive): Add rel parameter.
25563 Emit store-release versions when rel is true.
25564 (arm_split_compare_and_swap): Use acquire-release instructions
25566 of barriers when appropriate.
25567 (arm_split_atomic_op): Likewise.
25568 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
25569 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
25570 (VUNSPEC_SLX): Likewise.
25571 (VUNSPEC_LDA): Likewise.
25572 (VUNSPEC_STL): Likewise.
25573 * config/arm/sync.md (atomic_load<mode>): New pattern.
25574 (atomic_store<mode>): Likewise.
25575 (arm_load_acquire_exclusive<mode>): Likewise.
25576 (arm_load_acquire_exclusivesi): Likewise.
25577 (arm_load_acquire_exclusivedi): Likewise.
25578 (arm_store_release_exclusive<mode>): Likewise.
25580 2013-03-25 Catherine Moore <clm@codesourcery.com>
25582 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
25583 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
25584 * config/mip/predicates.md (lwsp_swsp_operand,
25585 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
25586 sb16_operand, db4_operand, db7_operand, ib3_operand,
25587 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
25588 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
25589 andi16_operand): New predicates.
25590 * config/mips/mips.md (compression): New attribute.
25591 (enabled): New attribute.
25592 (length): Consider compression in computing length.
25593 (shift_compression): New code attribute.
25594 (*add<mode>3): New operands. Record compression.
25595 (sub<mode>3): Likewise.
25596 (one_cmpl<mode>2): Likewise.
25597 (*and<mode>3): Likewise.
25598 (*ior<mode>3): Likewise.
25599 (unnamed pattern for xor): Likewise.
25600 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
25601 (*<optab><mode>3): Likewise.
25602 (*mov<mode>_internal: Likewise.
25603 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
25604 (mips_unsigned_immediate_p): New.
25605 (umips_lwsp_swsp_address_p): New.
25606 (m16_based_address_p): New.
25607 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
25608 (mips_unsigned_immediate_p): New prototype.
25609 (lwsp_swsp_address_p): New prototype.
25610 (m16_based_address_p): New prototype.
25611 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
25612 (mips_signed_immediate_p): New function.
25613 (m16_based_address_p): New function.
25614 (lwsp_swsp_address_p): New function.
25615 (mips_print_operand_punctuation): Recognize short delay slot insns
25616 for microMIPS.add<mode>3"
25618 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25621 * config/arm/iterators.md (v_cmp_result): New mode attribute.
25622 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
25624 2013-03-25 Richard Biener <rguenther@suse.de>
25626 PR tree-optimization/56689
25627 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
25630 2013-03-25 Richard Biener <rguenther@suse.de>
25632 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
25634 (memory_references): Likewise.
25635 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
25636 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
25637 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
25638 (gather_mem_refs_in_loops): Fold into ...
25639 (analyze_memory_references): ... this. Move initialization
25640 to tree_ssa_lim_initialize.
25641 (fill_always_executed_in): Rename to ...
25642 (fill_always_executed_in_1): ... this.
25643 (fill_always_executed_in): Move contains_call computation to
25644 this new function from ...
25645 (tree_ssa_lim_initialize): ... here.
25646 (tree_ssa_lim): Call fill_always_executed_in.
25648 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25650 * postreload.c (reload_combine): Fix code detecting returns.
25652 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25654 * function.c (emit_use_return_register_into_block): On cc0 targets,
25655 do not emit the sequence between cc0 setter and user.
25657 2013-03-25 Kai Tietz <ktietz@redhat.com>
25659 * config/i386/predicates.md (local_symbolic_operand): Interpret
25660 dll-imported symbols as none-local.
25662 2013-03-25 Richard Biener <rguenther@suse.de>
25664 * tree-ssa-loop-im.c (struct depend): Remove.
25665 (struct lim_aux_data): Make depends a vec of gimples.
25666 (free_lim_aux_data): Adjust.
25667 (add_dependency): Likewise.
25668 (set_level): Likewise.
25670 2013-03-25 Richard Biener <rguenther@suse.de>
25672 PR middle-end/56434
25673 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
25674 the pointer returned by calls with ECF_MALLOC set.
25676 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25678 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
25680 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25682 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
25683 using MMXMODE mode iterator.
25684 (*move<mode>_internal): Merge with *movv2sf_internal and
25685 *movv2sf_internal_rex64 using MMXMODE mode iterator.
25687 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25689 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
25690 (record_last_mem_set_info): Likewise.
25692 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
25693 of XNEWVEC followed by memset.
25694 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
25696 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25698 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
25699 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
25700 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
25701 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
25702 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
25703 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
25704 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
25705 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
25706 BARRIER_P instead of GET_CODE.
25708 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25710 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
25711 inaccuracy in the probing code.
25713 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
25714 (ctrapdi4): Likewise.
25716 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25718 * calls.c (expand_call): Add missing guard to code handling return
25719 of non-BLKmode structures in MSB.
25720 * function.c (expand_function_end): Likewise.
25722 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25724 * combine.c (try_combine): Adjust comment. Do not add the set of
25725 insn #0 if the destination indirectly is set or dies in insn #2.
25726 Tidy up code to distribute a new note.
25728 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25730 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
25731 also for alternatives 16 and 17.
25733 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25735 * config/i386/sse.md (*mov<mode>_internal): Merge with
25736 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
25737 Emit insn template depending on type attribute. Use
25738 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25739 movd instead of movq mnemonic for interunit moves. Rewrite mode
25740 attribute calculation. Remove unit attribute calculation.
25741 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
25742 Set prefix_data16 attribute for DImode ssemov types.
25743 Use Ym instead of y for SSE-MMX conversion alternatives.
25744 Reorder operand constraints.
25746 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
25748 * df.h (df_insn_delete): Adjust prototype.
25749 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
25750 and let it decide whether mark the basic block dirty.
25751 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
25752 * df-scan.c (df_insn_info_delete): New helper function, split
25753 off from df_insn_delete.
25754 (df_scan_free_bb_info): Use it.
25755 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
25757 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
25758 that the insn is actually an insn and it has a non-NULL basic block.
25759 Do not mark basic block dirty if only deleting a DEBUG_INSN.
25761 2013-03-22 Richard Biener <rguenther@suse.de>
25763 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
25765 (mem_ref_alloc): Do not allocate them.
25766 (refs_independent_p): Do not query or maintain a cache.
25768 2013-03-22 Richard Biener <rguenther@suse.de>
25770 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
25771 (gather_mem_refs_in_loops): Do not compute it.
25772 (analyze_memory_references): Do not allocate it.
25773 (tree_ssa_lim_finalize): Do not free it.
25774 (for_all_locs_in_loop): Do not query all_refs_in_loop.
25776 2013-03-22 Richard Biener <rguenther@suse.de>
25778 * is-a.h (as_a): Use gcc_checking_assert.
25780 2013-03-22 Ian Bolton <ian.bolton@arm.com>
25782 * config/aarch64/aarch64.c (aarch64_print_operand): New
25783 format specifier for printing a constant in hex.
25784 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
25785 format specifier for printing second operand.
25787 2013-03-22 Richard Biener <rguenther@suse.de>
25789 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
25791 (gather_mem_refs_in_loops): Perform store accumulation here.
25792 (create_vop_ref_mapping_loop): Remove.
25793 (create_vop_ref_mapping): Likewise.
25794 (analyze_memory_references): Initialize refs_stored_in_loop.
25795 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
25796 (record_indep_loop): Remove.
25797 (record_dep_loop): New function.
25798 (ref_indep_loop_p_1): Adjust to only walk over references
25799 in the loop, not its subloops.
25800 (ref_indep_loop_p): Rename to ...
25801 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
25802 maintaining a more fine-grained cache.
25803 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
25804 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
25806 2013-03-22 Richard Biener <rguenther@suse.de>
25808 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
25809 (struct mem_ref): Make accesses_in_loop a vec of a vec of
25810 aggregate mem_ref_loc.
25811 (free_mem_ref_locs): Inline into ...
25812 (memref_free): ... this and adjust.
25813 (mem_ref_alloc): Adjust.
25814 (mem_ref_locs_alloc): Remove.
25815 (record_mem_ref_loc): Adjust.
25816 (get_all_locs_in_loop): Rewrite into ...
25817 (for_all_locs_in_loop): ... this iterator.
25818 (rewrite_mem_ref_loc): New functor.
25819 (rewrite_mem_refs): Use for_all_locs_in_loop.
25820 (sm_set_flag_if_changed): New functor.
25821 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
25822 (ref_always_accessed): New functor.
25823 (ref_always_accessed_p): Use for_all_locs_in_loop.
25825 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25827 * tree-pass.h (PROP_gimple_lvec): New.
25828 * passes.c (dump_properties): Handle PROP_gimple_lvec.
25829 (init_optimization_passes): Move pass_lower_vector.
25830 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
25832 (pass_lower_vector): Provide PROP_gimple_lvec.
25833 (pass_lower_vector_ssa): Likewise.
25834 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
25836 2013-03-21 Mark Wielaard <mjw@redhat.com>
25838 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
25840 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25842 * config/i386/i386.md (*movdi_internal): Disparage slightly
25843 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
25844 conversion alternatives.
25846 2013-03-21 Jakub Jelinek <jakub@redhat.com>
25848 PR middle-end/48087
25849 * diagnostic.def (DK_WERROR): New kind.
25850 * diagnostic.h (werrorcount): Define.
25851 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25852 promoted to DK_ERROR, increment DK_WERROR counter instead of
25854 * toplev.c (toplev_main): Call print_ignored_options even if
25855 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25856 even if just werrorcount is non-zero.
25859 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25861 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25862 (string_cst_pool_decl): New function.
25863 (optimize_one_addr_into_implicit_ptr): New function.
25864 (resolve_addr_in_expr): Optimize DWARF location expression
25865 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25866 which doesn't live in memory, but has DW_AT_location or
25867 DW_AT_const_value, or refers to a string literal, into
25868 DW_OP_GNU_implicit_pointer.
25869 (optimize_location_into_implicit_ptr): New function.
25870 (resolve_addr): If removing DW_AT_location of a variable because
25871 it was DW_OP_addr of address of the variable, but the variable doesn't
25872 live in memory, try to emit const value attribute for the initializer.
25874 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25876 * tree.h (VECTOR_TYPE_P): New macro.
25877 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25878 TYPE_MODE): Use it.
25879 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25880 VEC_COND_EXPR cannot be lvalues.
25881 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25883 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25885 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25886 Restrict the transformation to equal modes.
25888 2013-03-21 Richard Biener <rguenther@suse.de>
25890 PR tree-optimization/39326
25891 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25892 (MEM_ANALYZABLE): Adjust.
25893 (record_mem_ref_loc): Move bitmap ops ...
25894 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25895 unanalyzable refs, do not record locations for it.
25896 (analyze_memory_references): Allocate ref zero as shared
25898 (refs_independent_p): Do not test for unanalyzed mems here.
25899 (ref_indep_loop_p_1): Special-case disambiguation against
25900 the unanalyzed ref.
25901 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25903 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25905 * config/arm/arm-protos.h (tune_params): Add
25906 prefer_neon_for_64bits field.
25907 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25908 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25909 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25910 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25911 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25912 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25913 (arm_option_override): Handle -mneon-for-64bits new option.
25914 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25915 (prefer_neon_for_64bits): Declare new variable.
25916 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25917 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25918 (arch_enabled): Handle new arch types. Remove support for onlya8
25920 (one_cmpldi2): Use new arch names.
25921 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25922 * config/arm/arm.opt (mneon-for-64bits): Add option.
25923 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25924 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25925 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25927 * doc/invoke.texi (-mneon-for-64bits): Document.
25929 2013-03-21 Richard Biener <rguenther@suse.de>
25931 PR tree-optimization/39326
25932 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25933 (sort_bbs_in_loop_postorder_cmp): New function.
25934 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25936 2013-03-21 Richard Biener <rguenther@suse.de>
25938 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25939 (vect_insert_into_interleaving_chain): Likewise.
25940 (vect_drs_dependent_in_basic_block): Inline ...
25941 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25943 (vect_analyze_data_ref_dependence): ... here. Simplify.
25944 (vect_check_interleaving): Simplify.
25945 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25946 (vect_slp_analyze_data_ref_dependences): ... this new function.
25947 (dr_group_sort_cmp): New function.
25948 (vect_analyze_data_ref_accesses): Compute data-reference groups
25949 here instead of in vect_analyze_data_ref_dependence. Use
25950 a more efficient algorithm.
25951 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25952 vect_slp_analyze_data_ref_dependences. Call
25953 vect_analyze_data_ref_accesses earlier.
25954 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25955 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25956 (vect_slp_analyze_data_ref_dependences): New prototype.
25958 2013-03-21 Richard Biener <rguenther@suse.de>
25960 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25961 ref is stored in the loop.
25962 (find_refs_for_sm): Walk only over all stores.
25963 (store_motion_loop): Allocate from lim_bitmap_obstack.
25964 (store_motion): Likewise.
25966 2013-03-21 Richard Biener <rguenther@suse.de>
25968 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25969 Update virtual SSA form.
25971 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25973 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25974 * configure: Regenerate.
25975 * config.in: Regenerate.
25976 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25977 if !HAVE_LD_EH_FRAME_CIEV3.
25979 2013-03-21 Richard Biener <rguenther@suse.de>
25981 * tree-cfg.c (verify_expr_no_block): New function.
25982 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25983 nor DECL_VALUE_EXPR have locations with associated blocks.
25984 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25985 (clear_unused_block_pointer): Remove code dealing with
25986 blocks in DECL_DEBUG_EXPR locations.
25988 2013-03-21 Richard Biener <rguenther@suse.de>
25990 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25991 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25992 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25993 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
25994 instead of DECL_DEBUG_EXPR_IS_FROM.
25995 * gimplify.c (gimplify_modify_expr): Likewise.
25996 * tree-cfg.c (verify_expr_location_1): Likewise.
25997 * tree-complex.c (create_one_component_var): Likewise.
25998 * tree-sra.c (create_access_replacement): Likewise.
25999 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
26000 (clear_unused_block_pointer): Likewise.
26001 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
26002 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
26003 * var-tracking.c (var_debug_decl): Likewise.
26004 (track_expr_p): Likewise.
26005 * tree-inline.c (add_local_variables): Likewise. Set
26006 DECL_HAS_DEBUG_EXPR_P after copying it.
26007 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
26008 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
26010 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
26013 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
26014 * configure: Regenerate.
26015 * config.in: Regenerate.
26016 * config/i386/i386.md (*movdf_internal): Use
26017 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
26018 movd instead of movq mnemonic for interunit moves.
26019 (*movdi_internal): Ditto.
26021 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26023 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
26024 (abd<mode>_3): New pattern.
26025 (aba<mode>_3): New pattern.
26026 (fabd<mode>_3): New pattern.
26028 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26030 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
26031 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
26032 occurrence of REGISTER_PREFIX as its empty string.
26034 2013-03-20 Jeff Law <law@redhat.com>
26036 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
26037 addititional equivalences for equality comparisons between an SSA_NAME
26038 and a constant where the SSA_NAME was set from a widening conversion.
26040 2013-03-20 Walter Lee <walt@tilera.com>
26042 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
26044 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26046 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
26047 depending on type attribute.
26048 (*movti_internal): Ditto.
26049 (*movtf_internal): Ditto.
26050 (*movxf_internal): Ditto.
26051 (*movdf_internal): Ditto.
26052 (*movsf_internal): Ditto.
26054 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26056 * config/i386/i386.md (*movti_internal): Set prefix attribute to
26057 maybe_vex for sselog1 and ssemov types.
26058 (*movdi_internal): Reorder operand constraints.
26059 (*movsi_internal): Ditto. Set prefix attribute to
26060 maybe_vex for sselog1 and ssemov types.
26061 (*movtf_internal): Set prefix attribute to maybe_vex
26062 for sselog1 and ssemov types.
26063 (*movdf_internal): Ditto. Set prefix_data16 attribute for
26064 DImode ssemov types. Reorder operand constraints.
26065 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
26066 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
26067 attribute for SImode ssemov types. Reorder operand constraints.
26069 2013-03-20 Martin Jambor <mjambor@suse.cz>
26071 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
26072 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
26074 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
26076 * config/rs6000/predicates.md (indexed_address, update_address_mem
26077 update_indexed_address_mem): New predicates.
26078 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
26079 attribute for load/store instructions.
26080 * config/rs6000/dfp.md (movsd_store): Likewise.
26081 (movsd_load): Likewise.
26082 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
26083 (unnamed HI->DI extend define_insn): Likewise.
26084 (unnamed SI->DI extend define_insn): Likewise.
26085 (unnamed QI->SI extend define_insn): Likewise.
26086 (unnamed QI->HI extend define_insn): Likewise.
26087 (unnamed HI->SI extend define_insn): Likewise.
26088 (unnamed HI->SI extend define_insn): Likewise.
26089 (extendsfdf2_fpr): Likewise.
26090 (movsi_internal1): Likewise.
26091 (movsi_internal1_single): Likewise.
26092 (movhi_internal): Likewise.
26093 (movqi_internal): Likewise.
26094 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
26095 attribute for load/store instructions.
26096 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
26098 (mov<mode>_softfloat): Likewise.
26099 (mov<mode>_hardfloat32): Likewise.
26100 (mov<mode>_hardfloat64): Likewise.
26101 (mov<mode>_softfloat64): Likewise.
26102 (movdi_internal32): Likewise.
26103 (movdi_internal64): Likewise.
26104 (probe_stack_<mode>): Likewise.
26106 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
26108 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
26109 floating point, and decimal floating point to reload iterator.
26111 * config/rs6000/constraints.md (wl constraint): New constraints to
26112 return FLOAT_REGS if certain options are used to reduce the number
26113 of separate patterns that exist in the file.
26114 (wx constraint): Likewise.
26115 (wz constraint): Likewise.
26117 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
26118 -mdebug=reg, print wg, wl, wx, and wz constraints.
26119 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
26120 Initialize the reload functions for 64-bit binary/decimal floating
26122 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
26123 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
26124 create the buffer on the stack to overcome not having a 32-bit
26126 (rs6000_emit_move): Likewise.
26127 (rs6000_secondary_memory_needed_rtx): Likewise.
26128 (rs6000_alloc_sdmode_stack_slot): Likewise.
26129 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
26130 via xxlxor, just like DFmode 0.0.
26132 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
26133 define as 1 if we are running on a power7 or newer.
26134 (enum r6000_reg_class_enum): Add new constraints.
26136 * config/rs6000/dfp.md (movsd): Delete, combine with binary
26137 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
26138 with other moves by using conditional constraits (wg). Use LFIWZX
26139 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
26140 (movsd splitter): Likewise.
26141 (movsd_hardfloat): Likewise.
26142 (movsd_softfloat): Likewise.
26144 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
26145 binary and decimal floating point moves.
26146 (fmove_ok): New attributes to combine binary and decimal floating
26147 point moves, and to combine power6x (mfpgpr) moves along normal
26149 (real_value_to_target): Likewise.
26150 (f32_lr): Likewise.
26151 (f32_lm): Likewise.
26152 (f32_li): Likewise.
26153 (f32_sr): Likewise.
26154 (f32_sm): Likewise.
26155 (f32_si): Likewise.
26156 (movsf): Combine binary and decimal floating point moves. Combine
26157 power6x (mfpgpr) moves with other moves by using conditional
26158 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
26159 (mov<mode> for SFmode/SDmode); Likewise.
26160 (SFmode/SDmode splitters): Likewise.
26161 (movsf_hardfloat): Likewise.
26162 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
26163 (movsf_softfloat): Likewise.
26164 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
26166 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
26167 wx and wz constraints.
26169 * config/rs6000/constraints.md (wg constraint): New constraint to
26170 return FLOAT_REGS if -mmfpgpr (power6x) was used.
26172 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
26175 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
26176 -mdebug=reg, print wg, wl, wx, and wz constraints.
26177 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
26178 Initialize the reload functions for 64-bit binary/decimal floating
26180 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
26181 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
26182 create the buffer on the stack to overcome not having a 32-bit
26184 (rs6000_emit_move): Likewise.
26185 (rs6000_secondary_memory_needed_rtx): Likewise.
26186 (rs6000_alloc_sdmode_stack_slot): Likewise.
26187 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
26188 via xxlxor, just like DFmode 0.0.
26190 * config/rs6000/dfp.md (movdd): Delete, combine with binary
26191 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
26192 with other moves by using conditional constraits (wg). Use LFIWZX
26193 and STFIWX for loading SDmode on power7.
26194 (movdd splitters): Likewise.
26195 (movdd_hardfloat32): Likewise.
26196 (movdd_softfloat32): Likewise.
26197 (movdd_hardfloat64_mfpgpr): Likewise.
26198 (movdd_hardfloat64): Likewise.
26199 (movdd_softfloat64): Likewise.
26201 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
26202 64-bit binary and decimal floating point moves.
26203 (FMOVE64X): Likewise.
26204 (movdf): Combine 64-bit binary and decimal floating point moves.
26205 Combine power6x (mfpgpr) moves with other moves by using
26206 conditional constraits (wg).
26207 (mov<mode> for DFmode/DDmode): Likewise.
26208 (DFmode/DDmode splitters): Likewise.
26209 (movdf_hardfloat32): Likewise.
26210 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
26211 (movdf_softfloat32): Likewise.
26212 (movdf_hardfloat64_mfpgpr): Likewise.
26213 (movdf_hardfloat64): Likewise.
26214 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
26215 (movdf_softfloat64): Likewise.
26216 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
26217 (reload_<mode>_load): Move to later in the file so they aren't in
26218 the middle of the floating point move insns.
26219 (reload_<mode>_store): Likewise.
26221 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
26224 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
26225 constraint if -mdebug=reg.
26226 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
26227 Enable using dd reload support if needed.
26229 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
26230 binary and decimal floating point moves in rs6000.md.
26231 (movtd_internal): Likewise.
26233 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
26234 decimal floating point moves.
26236 (movtf_internal): Likewise.
26237 (mov<mode>_internal, TDmode/TFmode): Likewise.
26238 (movtf_softfloat): Likewise.
26239 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
26241 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
26242 movdi_internal64, using wg constraint for move direct operations.
26243 (movdi_internal64): Likewise.
26245 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
26246 MODES_TIEABLE_P for selected modes. Print the numerical value of
26247 the various virtual registers. Use GPR/FPR first/last values,
26248 instead of hard coding the register numbers. Print which modes
26249 have reload functions registered.
26250 (rs6000_option_override_internal): If -mdebug=reg, trace the options
26251 settings before/after setting cpu, target and subtarget settings.
26252 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
26253 and for secondary reload failures in rs6000_secondary_reload_inner.
26254 (rs6000_secondary_reload_fail): Likewise.
26255 (rs6000_secondary_reload_inner): Likewise.
26257 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
26258 macros for first/last GPR and FPR registers.
26259 (LAST_GPR_REGNO): Likewise.
26260 (FIRST_FPR_REGNO): Likewise.
26261 (LAST_FPR_REGNO): Likewise.
26263 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
26264 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
26265 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
26266 (vcond<mode><mode>): Likewise.
26267 (vcondu<mode><mode>): Likewise.
26268 (vector_gtu<mode>): Likewise.
26269 (vector_gte<mode>): Likewise.
26270 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
26271 to prevent the compiler from converting DImode operations to TImode.
26272 (ior<mode>3): Likewise.
26273 (and<mode>3): Likewise.
26274 (one_cmpl<mode>2): Likewise.
26275 (nor<mode>3): Likewise.
26276 (andc<mode>3): Likewise.
26278 * config/rs6000/constraints.md (wt constraint): New constraint
26279 that returns VSX_REGS if TImode is allowed in VSX registers.
26281 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
26282 constant under VSX.
26284 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
26285 similar to TImode, but it is restricted to being in the GPRs.
26287 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
26288 TImode to occupy a single VSX register.
26290 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
26291 -mvsx-timode for power7/power8.
26292 (power7 cpu): Likewise.
26293 (power8 cpu): Likewise.
26295 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
26296 sure that TFmode/TDmode take up two registers if they are ever
26297 allowed in the upper VSX registers.
26298 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
26300 (rs6000_init_hard_regno_mode_ok): Likewise.
26301 (rs6000_debug_reg_global): Add debugging for PTImode and wt
26302 constraint. Print if LRA is turned on.
26303 (rs6000_option_override_internal): Give an error if -mvsx-timode
26304 and VSX is not enabled.
26305 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
26306 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
26307 to reg+offset addressing. Use PTImode when checking offset
26308 addresses for validity.
26309 (reg_offset_addressing_ok_p): Likewise.
26310 (rs6000_legitimate_offset_address_p): Likewise.
26311 (rs6000_legitimize_address): Likewise.
26312 (rs6000_legitimize_reload_address): Likewise.
26313 (rs6000_legitimate_address_p): Likewise.
26314 (rs6000_eliminate_indexed_memrefs): Likewise.
26315 (rs6000_emit_move): Likewise.
26316 (rs6000_secondary_reload): Likewise.
26317 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
26318 reloads to fpr registers to continue to use reg+offset addressing,
26319 but 64-bit reloads to altivec registers need reg+reg addressing.
26320 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
26321 it. Treat LO_SUM like a PLUS operation.
26322 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
26323 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
26324 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
26325 registers to share a register with a smaller sized type, since VSX
26326 puts scalars in the upper 64-bits.
26327 (print_operand): Add support for PTImode.
26328 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
26329 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
26330 registers, but don't have arithmetic support.
26331 (rs6000_memory_move_cost): Add test for VSX.
26332 (rs6000_opt_masks): Add -mvsx-timode.
26334 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
26337 (VSr): Use wt constraint for TImode.
26338 (VSv): Drop TImode support.
26339 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
26340 (vsx_movti_64bit): Likewise.
26341 (vsx_movti_32bit): Likewise.
26342 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
26343 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
26344 one '?' on the appropriate output constraint. Do not allow TImode
26345 logical operations on 32-bit systems.
26346 (vsx_ior<mode>3): Likewise.
26347 (vsx_xor<mode>3): Likewise.
26348 (vsx_one_cmpl<mode>2): Likewise.
26349 (vsx_nor<mode>3): Likewise.
26350 (vsx_andc<mode>3): Likewise.
26351 (vsx_concat_<mode>): Likewise.
26352 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
26354 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
26355 OPTION_MASK_VSX_TIMODE.
26356 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
26357 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
26359 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
26360 (TI2 iterator): New iterator for TImode, PTImode.
26361 (wd mode attribute): Add values for vector types.
26362 (movti_string): Replace TI move operations with operations for TImode
26363 and PTImode. Add support for TImode being allowed in VSX registers.
26364 (mov<mode>_string, TImode/PTImode): Likewise.
26365 (movti_ppc64): Likewise.
26366 (mov<mode>_ppc64, TImode/PTImode): Likewise.
26367 (TI mode splitters): Likewise.
26369 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
26372 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
26374 PR tree-optimization/56355
26375 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
26376 Also handle integers with undefined overflow.
26378 2013-03-20 Catherine Moore <clm@codesourcery.com>
26379 Maciej W. Rozycki <macro@codesourcery.com>
26380 Tom de Vries <tom@codesourcery.com>
26381 Nathan Sidwell <nathan@codesourcery.com>
26382 Iain Sandoe <iain@codesourcery.com>
26383 Nathan Froyd <froydnj@codesourcery.com>
26384 Chao-ying Fu <fu@mips.com>
26386 * doc/extend.texi: (micromips, nomicromips, nocompression):
26387 Document new function attributes.
26388 * doc/invoke.texi (minterlink-compressed, mmicromips,
26389 m14k, m14ke, m14kec): Document new options.
26390 (minterlink-mips16): Update documentation.
26391 * doc/md.texi (ZC, ZD): Document new constraints.
26392 * configure.ac (gcc_cv_as_micromips): Check if linker
26393 supports the .set micromips directive.
26394 * configure: Regenerate.
26395 * config.in: Regenerate.
26396 * config/mips/mips-tables.opt: Regenerate.
26397 * config/mips/micromips.md: New file.
26398 * constraints.md (ZC, ZD): New constraints.
26399 * config/mips/predicates.md (movep_src_register): New predicate.
26400 (movep_src_operand): New predicate.
26401 (non_volatile_mem_operand): New predicate.
26402 * config/mips/mips.md (multimem): New type.
26403 (length): Differentiate between 17-bit and 18-bit branch offsets.
26404 (MOVEP1, MOVEP2): New mode iterator.
26405 (mov_<load>l): Use ZC constraint.
26406 (mov_<load>r): Likewise.
26407 (mov_<store>l): Likewise.
26408 (mov_<store>r): Likewise.
26409 (*branch_equality<mode>_inverted): Add microMIPS support.
26410 (*branch_equality<mode>): Likewise.
26411 (*jump_absolute): Likewise.
26412 (indirect_jump_<mode>): Likewise.
26413 (tablejump_<mode>): Likewise.
26414 (<optab>_internal): Likewise.
26415 (sibcall_internal): Likewise.
26416 (sibcall_value_internal): Likewise.
26417 (prefetch): Use constraint ZD.
26418 * config/mips/mips.opt (minterlink-compressed): New option.
26419 (minterlink-mips16): Now an alias for minterlink-compressed.
26420 (mmicromips): New option.
26421 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26422 (compare_and_swap_12): Likewise.
26423 (sync_add<mode>): Likewise.
26424 (sync_<optab>_12): Likewise.
26425 (sync_old_<optab>_12): Likewise.
26426 (sync_new_<optab>_12): Likewise.
26427 (sync_nand_12): Likewise.
26428 (sync_old_nand_12): Likewise.
26429 (sync_new_nand_12): Likewise.
26430 (sync_sub<mode>): Likewise.
26431 (sync_old_add<mode>): Likewise.
26432 (sync_old_sub<mode>): Likewise.
26433 (sync_new_add<mode>): Likewise.
26434 (sync_new_sub<mode>): Likewise.
26435 (sync_<optab><mode>): Likewise.
26436 (sync_old_<optab><mode>): Likewise.
26437 (sync_new_<optab><mode>): Likewise.
26438 (sync_nand<mode>): Likewise.
26439 (sync_old_nand<mode>): Likewise.
26440 (sync_new_nand<mode>): Likewise.
26441 (sync_lock_test_and_set<mode>): Likewise.
26442 (test_and_set_12): Likewise.
26443 (atomic_compare_and_swap<mode>): Likewise.
26444 (atomic_exchange<mode>_llsc): Likewise.
26445 (atomic_fetch_add<mode>_llsc): Likewise.
26446 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26447 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26448 (umips_save_restore_pattern_p): Likewise.
26449 (umips_load_store_pair_p): Likewise.
26450 (umips_output_load_store_pair): Likewise.
26451 (umips_movep_target_p): Likewise.
26452 (umips_12bit_offset_address_p): Likewise.
26453 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26454 (mips_base_mips16): Rename this...
26455 (mips_base_compression_flags): ...to this. Update all uses.
26456 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26457 (mips_mips16_decl_p): Delete.
26458 (mips_nomips16_decl_p): Delete.
26459 (mips_get_compress_on_flags): New function.
26460 (mips_get_compress_off_flags): New function.
26461 (mips_get_compress_mode): New function.
26462 (mips_get_compress_on_name): New function.
26463 (mips_get_compress_off_name): New function.
26464 (mips_insert_attributes): Support multiple compression types.
26465 (mips_merge_decl_attributes): Likewise.
26466 (umips_12bit_offset_address_p): New function.
26467 (mips_start_function_definition): Emit .set micromips directive.
26468 (mips_call_may_need_jalx_p): New function.
26469 (mips_function_ok_for_sibcall): Add microMIPS support.
26470 (mips_print_operand_punctuation): Support short delay slots and
26472 (umips_swm_mask, umips_swm_encoding): New.
26473 (umips_build_save_restore): New function.
26474 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26475 (was_mips16_p): Remove.
26476 (old_compression_mode): New.
26477 (mips_set_compression_mode): New function.
26478 (mips_set_current_function): Add microMIPS support.
26479 (mips_option_override): Likewise.
26480 (umips_save_restore_pattern_p): New function.
26481 (umips_output_save_restore): New function.
26482 (umips_load_store_pair_p_1): New function.
26483 (umips_load_store_pair_p): New function.
26484 (umips_output_load_store_pair_1): New function.
26485 (umips_output_load_store_pair): New function.
26486 (umips_movep_target_p) New function.
26487 (mips_prepare_pch_save): Add microMIPS support.
26488 * config/mips/mips.h (TARGET_COMPRESSION): New.
26489 (TARGET_CPU_CPP_BUILTINS): Update macro
26490 to use new compression flags and to support microMIPS.
26491 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26492 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26493 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26494 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26495 (ASM_SPEC): Support mmicromips and mno-micromips.
26496 (M16STORE_REG_P): New macro.
26497 (MIPS_CALL): Support TARGET_MICROMIPS.
26498 (MICROMIPS_J): New macro.
26499 (mips_base_mips16): Rename this...
26500 (mips_base_compression_flags): ...to this.
26501 (UMIPS_12BIT_OFFSET_P): New macro.
26502 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26503 (MULTILIB_DIRNAMES): Likewise.
26504 2013-03-20 Richard Biener <rguenther@suse.de>
26506 PR tree-optimization/56661
26507 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
26508 the result does not have to be distinct.
26510 2013-03-20 Richard Biener <rguenther@suse.de>
26512 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
26515 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
26516 Steven Bosscher <steven@gcc.gnu.org>
26518 PR rtl-optimization/56605
26519 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
26521 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26524 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
26525 that require movd instead of movq.
26527 2013-03-20 Richard Biener <rguenther@suse.de>
26529 * tree-ssa-structalias.c (struct variable_info): Add pointer
26530 to the first field of an aggregate with sub-vars. Make
26531 this and the pointer to the next subfield its ID.
26532 (vi_next): New function.
26533 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
26534 storedanything_id, integer_id): Increment by one.
26535 (new_var_info, get_call_vi, lookup_call_clobber_vi,
26536 get_call_clobber_vi): Adjust.
26537 (solution_set_expand): Simplify and speedup.
26538 (solution_set_add): Inline into ...
26539 (set_union_with_increment): ... this. Adjust accordingly.
26540 (do_sd_constraint): Likewise.
26541 (do_ds_constraint): Likewise.
26542 (do_complex_constraint): Simplify.
26543 (build_pred_graph): Adjust.
26544 (solve_graph): Likewise. Simplify and speedup.
26545 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
26546 get_constraint_for_component_ref, get_constraint_for_1,
26547 first_vi_for_offset, first_or_preceding_vi_for_offset,
26548 create_function_info_for, create_variable_info_for_1,
26549 create_variable_info_for, intra_create_variable_infos): Adjust.
26550 (init_base_vars): Push NULL for ID zero.
26551 (compute_points_to_sets): Adjust.
26553 2013-03-20 Richard Biener <rguenther@suse.de>
26555 * cfgloop.c (verify_loop_structure): Streamline and avoid
26556 ICEing on corrupt loop tree.
26557 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
26560 2013-03-20 Richard Biener <rguenther@suse.de>
26562 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
26563 check whether an SSA update is needed.
26565 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
26567 * config/mips/constraints.md (T): Rename to...
26571 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
26572 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
26574 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26576 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
26577 (*subsi3_carryin_uxtw): Likewise.
26579 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26581 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
26582 (*rorsi3_insn_uxtw): Likewise.
26584 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26586 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
26587 (*extrsi5_insn_uxtw): Likewise.
26589 2013-03-19 Richard Biener <rguenther@suse.de>
26591 PR tree-optimization/56273
26592 * passes.c (init_optimization_passes): Move second VRP after DOM.
26594 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
26596 * config/i386/i386.md (*movti_internal): Merge from
26597 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
26598 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
26599 nox64 isa attributes.
26601 2013-03-18 Richard Biener <rguenther@suse.de>
26603 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
26605 (merge_node_constraints): Likewise.
26606 (build_succ_graph): Likewise.
26607 (valid_graph_edge): Inline into single caller.
26608 (unify_nodes): Likewise. Use bitmap_set_bit return value
26610 (scc_visit): Fix formatting and variable use.
26611 (do_sd_constraint): Use gcc_checking_assert.
26612 (do_ds_constraint): Likewise.
26613 (do_complex_constraint): Likewise.
26614 (condense_visit): Likewise. Cleanup.
26615 (dump_pred_graph): New function.
26616 (perform_var_substitution): Dump the pred-graph before
26617 variable substitution.
26618 (find_equivalent_node): Use gcc_checking_assert.
26619 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
26621 2013-03-18 Richard Biener <rguenther@suse.de>
26623 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
26624 Remove cond_expr_stmt_list argument and do not gimplify the
26626 (vect_loop_versioning): Adjust.
26627 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
26628 Cleanup to use less temporaries.
26629 (vect_create_data_ref_ptr): Cleanup.
26631 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26633 PR tree-optimization/56635
26634 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
26635 require types_compatible_p types.
26637 2013-03-18 Nick Clifton <nickc@redhat.com>
26639 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
26640 spurious backslash.
26642 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
26643 Add missing line to comment describing function.
26645 2013-03-18 Richard Biener <rguenther@suse.de>
26647 PR tree-optimization/56210
26648 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26649 Handle string / character search functions.
26650 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
26652 2013-03-18 Richard Biener <rguenther@suse.de>
26654 PR middle-end/56483
26655 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
26656 and implement properly.
26657 * gimple.h (gimple_cond_single_var_p): Remove.
26659 2013-03-18 Richard Biener <rguenther@suse.de>
26661 * tree-data-ref.h (find_data_references_in_loop): Declare.
26662 * tree-data-ref.c (get_references_in_stmt): Use a stack
26663 vector pre-allocated in the callers.
26664 (find_data_references_in_stmt): Adjust.
26665 (graphite_find_data_references_in_stmt): Likewise.
26666 (create_rdg_vertices): Likewise.
26667 (find_data_references_in_loop): Export.
26668 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
26669 Compute dependences here...
26670 (vect_analyze_data_refs): ...not here. When we encounter
26671 a non-vectorizable data reference in basic-block vectorization
26672 truncate the data reference vector. Do not bother to
26673 fixup data-dependence information for gather loads.
26674 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
26675 of data references, as reported.
26677 2013-03-18 Richard Biener <rguenther@suse.de>
26679 PR tree-optimization/3713
26680 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
26681 has_constants and expr.
26682 (stmt_has_constants): Properly valueize SSA names when deciding
26683 whether the stmt has constants.
26685 2013-03-18 Richard Biener <rguenther@suse.de>
26687 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
26688 whole function when there is nothing to do.
26689 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
26690 * tree-vectorizer.c (vectorize_loops): Update virtual and
26691 loop-closed SSA once.
26692 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
26694 2013-03-18 Richard Biener <rguenther@suse.de>
26696 PR middle-end/56113
26697 * domwalk.c (bb_postorder): New global static.
26698 (cmp_bb_postorder): New function.
26699 (walk_dominator_tree): Replace scheme imposing an order for
26700 visiting dominator sons by one sorting them at the time they
26701 are pushed on the stack.
26703 2013-03-18 Richard Biener <rguenther@suse.de>
26705 PR tree-optimization/39326
26706 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
26707 (struct mem_ref): Replace mem member with ao_ref typed member.
26708 (MEM_ANALYZABLE): Adjust.
26709 (memref_eq): Likewise.
26710 (mem_ref_alloc): Likewise.
26711 (gather_mem_refs_stmt): Likewise.
26712 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
26713 (execute_sm_if_changed_flag_set): Adjust.
26714 (execute_sm): Likewise.
26715 (ref_always_accessed_p): Likewise.
26716 (refs_independent_p): Likewise.
26717 (can_sm_ref_p): Likewise.
26719 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26722 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
26723 return 1 even for !unsignedp.
26725 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26727 * config/i386/i386.md (isa): Add x64 and nox64.
26728 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
26729 (*pushtf): Enable *roF alternative for x64 isa only.
26730 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
26731 mode attribute of integer alternatives to DImode for TARGET_64BIT.
26732 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
26733 (*movtf_internal): Merge from *movtf_internal_rex64 and
26734 *movtf_internal_sse. Use x64 and nox64 isa attributes.
26735 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
26736 nox64 isa attributes.
26737 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
26738 nox64 isa attributes.
26739 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
26741 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26743 * config/alpha/alpha.c (TARGET_LRA_P): New define.
26745 2013-03-17 Jakub Jelinek <jakub@redhat.com>
26748 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
26749 class names. Remove trailing comma after "ALL_REGS".
26751 2013-03-16 Jan Hubicka <jh@suse.cz>
26753 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
26754 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
26755 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
26756 of cgraph_get_create_node.
26757 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
26759 2013-03-16 Jason Merrill <jason@redhat.com>
26762 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
26763 with DW_AT_default_value.
26765 2013-03-16 Jakub Jelinek <jakub@redhat.com>
26767 * BASE-VER: Set to 4.9.0.
26769 2013-03-14 Andi Kleen <ak@linux.intel.com>
26772 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
26773 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
26774 Document _x* TSX intrinsics.
26776 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
26777 David Holsgrove <david.holsgrove@xilinx.com>
26779 * configure.ac: Add MicroBlaze TLS support detection.
26780 * configure: Regenerate.
26781 * config/microblaze/microblaze-protos.h
26782 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
26783 symbol_mentioned_p, label_mentioned_p): Add prototypes.
26784 * config/microblaze/microblaze.c (microblaze_address_type): Add
26785 ADDRESS_TLS and tls_reloc address types.
26786 (microblaze_address_info): Add tls_reloc.
26787 (TARGET_HAVE_TLS): Define.
26788 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
26789 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
26790 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
26791 load_tls_operand, microblaze_call_tls_get_addr,
26792 microblaze_legitimize_tls_address): New functions.
26793 (microblaze_classify_unspec): Handle UNSPEC_TLS.
26794 (get_base_reg): Use microblaze_tls_symbol_p.
26795 (microblaze_classify_address): Handle TLS.
26796 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
26797 label_mentioned_p and microblaze_tls_referenced_p.
26798 (microblaze_legitimize_address): Handle TLS.
26799 (microblaze_address_insns): Handle ADDRESS_TLS.
26800 (pic_address_needs_scratch): Handle TLS.
26801 (print_operand_address): Handle TLS.
26802 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
26803 (microblaze_expand_move): Handle TLS.
26804 (microblaze_legitimate_constant_p): Check
26805 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
26806 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
26807 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
26808 (PIC_OFFSET_TABLE_REGNUM): Set.
26809 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
26810 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
26811 (addsi3, movsi_internal2, movdf_internal): Update constraints
26812 * config/microblaze/predicates.md (arith_plus_operand): Define
26813 (move_operand): Redefine as move_src_operand,
26814 check microblaze_tls_referenced_p.
26816 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26818 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
26819 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
26821 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26823 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
26826 2013-03-14 Jakub Jelinek <jakub@redhat.com>
26828 PR tree-optimization/53265
26829 * common.opt (Waggressive-loop-optimizations): New option.
26830 * tree-ssa-loop-niter.c: Include tree-pass.h.
26831 (do_warn_aggressive_loop_optimizations): New function.
26832 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
26833 if number_of_latch_executions returned constant.
26834 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
26835 early. If number_of_latch_executions returned constant, set
26836 nb_iterations_upper_bound back to it.
26837 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
26839 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
26840 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
26842 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
26843 (MULTILIB_OSDIRNAMES): Set.
26844 * genmultilib: If defaultosdirname doesn't start with :: , set
26845 defaultosdirname2 instead, clear it and emit two . multilib_raw
26846 entries instead of just one.
26848 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26850 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26851 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26852 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26853 (SUBTARGET_OVERRIDE_OPTIONS): New.
26855 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26858 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26859 (musermode): Convert to Var(TARGET_USERMODE).
26860 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26861 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26862 * config/sh/sh.c (sh_option_override): Use
26863 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26864 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26866 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26868 (udivsi3_i4_single, divsi3_i4_single): Use
26869 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26871 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26873 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26876 2013-03-13 Richard Biener <rguenther@suse.de>
26878 PR tree-optimization/56608
26879 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26880 calls when vectorizing basic-blocks.
26882 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26885 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26888 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26890 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26892 2013-03-11 Jan Hubicka <jh@suse.cz>
26895 * lto-streamer-out.c (output_symbol_p): Skip references from
26896 constructors of external variables.
26898 2013-03-11 Jan Hubicka <jh@suse.cz>
26900 PR middle-end/56571
26901 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26903 * emit-rtl.c (verify_rtx_sharing): Likewise.
26904 (copy_insn_1): Likewise.
26905 * rtl.c (copy_rtx): Likewise.
26907 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26910 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26911 output_operand_lossage message.
26913 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26916 * arm.c (shift_op): Validate RTL pattern on the fly.
26917 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26920 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26923 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26924 new call patterns that clobber %fr12.
26925 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26926 split and postreload patterns.
26927 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26928 registers %fr12 and %fr12R as call used.
26930 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26932 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26933 (canon_address, record_store, replace_read, check_mem_read_rtx,
26934 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26935 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26936 rest_of_handle_dse): Likewise.
26938 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26940 PR middle-end/56524
26941 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26943 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26944 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26945 (save_optabs_if_changed): Replace with...
26946 (init_tree_optimization_optabs): ...this.
26947 * optabs.c (save_optabs_if_changed): Rename to...
26948 (init_tree_optimization_optabs): ...this. Take the optimization node
26949 as argument. Do nothing if the base optabs are already correct.
26950 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26951 to recompute optabs.
26952 * function.h (function): Remove optabs field.
26953 * function.c (invoke_set_current_function_hook): Call
26954 init_tree_optimization_optabs. Use the result to initialize
26957 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26959 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26960 if GTMA_HAS_NO_INSTRUMENTATION.
26961 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26962 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26963 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26964 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26965 GTMA_HAS_NO_INSTRUMENTATION.
26967 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26969 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26972 2013-03-08 Marek Polacek <polacek@redhat.com>
26973 Jakub Jelinek <jakub@redhat.com>
26975 PR tree-optimization/56478
26976 * predict.c (is_comparison_with_loop_invariant_p): Change the
26977 type of loop_step to tree.
26978 (predict_loops): Adjust.
26979 (predict_iv_comparison): Perform the computations on double_ints.
26981 2013-03-08 Richard Biener <rguenther@suse.de>
26983 PR tree-optimization/56570
26984 * tree-cfg.c (verify_expr_location_1): Verify locations for
26986 * tree-sra.c (create_access_replacement): Strip locations
26987 from DECL_DEBUG_EXPRs.
26989 2013-03-08 Richard Biener <rguenther@suse.de>
26991 * tree-inline.c (expand_call_inline): Do not associate
26992 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26993 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
26995 2013-03-08 Richard Biener <rguenther@suse.de>
26997 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
26998 or block changes with -Og. Fix for location / block encoding
26999 changes and PHI arguments with locations.
27001 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
27003 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
27005 (struct output_info): Likewise.
27006 (register_overhead): Remove bad gcc_assert.
27007 (bitmap_find_bit): If there is only a single bitmap element, do not
27008 count a miss as a search.
27009 (print_statistics): Update for counter type changes.
27010 (dump_bitmap_statistics): Likewise. Print headers such that they
27011 are properly lined up with the printed counters.
27013 2013-03-07 Jakub Jelinek <jakub@redhat.com>
27015 PR tree-optimization/56559
27016 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
27017 check that it has only a single use.
27019 2013-03-07 Richard Biener <rguenther@suse.de>
27021 * doc/invoke.texi (fwhole-program): Discourage use in combination
27024 2013-03-06 Jakub Jelinek <jakub@redhat.com>
27026 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
27028 PR tree-optimization/56539
27029 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
27030 instead of GSI_CONTINUE_LINKING as last argument to
27031 force_gimple_operand_gsi. Adjust function comment.
27033 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
27036 PR middle-end/56548
27037 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
27038 promoted mode, convert the result back to the original mode.
27040 2013-03-06 Richard Biener <rguenther@suse.de>
27042 PR middle-end/56294
27043 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
27044 (insert_updated_phi_nodes_compare_uids): New function.
27045 (update_ssa): Sort symbols_to_rename after UID before
27046 traversing it to insert PHI nodes.
27048 2013-03-06 Richard Biener <rguenther@suse.de>
27050 PR middle-end/50494
27051 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
27052 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
27055 2013-02-13 Richard Biener <rguenther@suse.de>
27058 * varasm.c (output_constant_def_1): Get the decl representing
27059 the constant as argument.
27060 (output_constant_def): Wrap output_constant_def_1.
27061 (make_decl_rtl): Use output_constant_def_1 with the decl
27062 representing the constant.
27063 (build_constant_desc): Optionally re-use a decl already
27064 representing the constant.
27065 (tree_output_constant_def): Adjust.
27067 2013-03-06 Joey Ye <joey.ye@arm.com>
27070 * gcc.c (convert_white_space): New function.
27071 (main): Handles white space in function name.
27073 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
27076 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
27077 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
27078 to SH_DIV_CALL_TABLE for TARGET_SH2.
27079 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
27081 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
27082 call-table options.
27084 2013-03-05 Sterling Augustine <saugustine@google.com>
27085 Cary Coutant <ccoutant@google.com>
27088 * dwarf2out.c (resolve_addr): Don't call
27089 remove_loc_list_addr_table_entries a second time for the same
27092 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27095 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
27096 (avoid_complex_debug_insns): New function.
27097 (expand_debug_locations): Call it.
27099 PR rtl-optimization/56484
27100 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
27101 lifetimes of hard registers on small register class machines.
27103 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
27105 * config/microblaze/microblaze-protos.h: Rename
27106 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
27107 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
27109 (microblaze_fast_interrupt_function_p): New function.
27110 (microblaze_is_interrupt_handler): Rename to
27111 microblaze_is_interrupt_variant and add fast_interrupt check.
27112 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
27113 (save_restore_insns): Likewise.
27114 (compute_frame_size): Likewise.
27115 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
27116 (microblaze_globalize_label): Likewise.
27117 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
27118 * config/microblaze/microblaze.md: Use wrapper
27119 microblaze_is_interrupt_variant.
27121 2013-03-05 Kai Tietz <ktietz@redhat.com>
27123 * sdbout.c (sdbout_one_type): Switch to current function's section
27124 supporting cold/hot.
27126 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
27128 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
27131 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27133 PR middle-end/56461
27134 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
27135 if VALGRIND_GET_VBITS is defined, temporarily make object
27136 memory all defined, and restore previous valgrind addressability
27137 and definability afterwards. Free this_object at the end.
27139 PR middle-end/56461
27140 * lra.c (lra): Call lra_clear_live_ranges if live_p,
27141 right before calling lra_create_live_ranges, also call it
27142 when clearing live_p. Only call lra_clear_live_ranges
27143 at the end if live_p.
27145 PR middle-end/56461
27146 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
27148 2013-03-05 Richard Biener <rguenther@suse.de>
27150 PR tree-optimization/56521
27151 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
27154 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
27157 * except.h (remove_unreachable_eh_regions): New prototype.
27158 * except.c (remove_eh_handler_splicer): New function, split out
27159 of remove_eh_handler.
27160 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
27161 warning about running it on many EH regions one at a time.
27162 (remove_unreachable_eh_regions_worker): New function, walk the
27163 EH tree in depth-first order and remove non-marked regions.
27164 (remove_unreachable_eh_regions): New function.
27165 * tree-eh.c (mark_reachable_handlers): New function, split out
27166 from remove_unreachable_handlers.
27167 (remove_unreachable_handlers): Use mark_reachable_handlers and
27168 remove_unreachable_eh_regions.
27169 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
27170 and remove_unreachable_eh_regions.
27172 2013-03-05 Richard Biener <rguenther@suse.de>
27174 PR middle-end/56525
27175 * loop-init.c (fix_loop_structure): Remove loops in two stages,
27176 not freeing them until the end.
27178 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27180 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
27182 2013-03-05 Richard Biener <rguenther@suse.de>
27184 PR tree-optimization/56270
27185 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
27186 of loads after scheduling an SLP instance.
27188 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27190 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
27192 (check_gcc_parallelize): Run guality.exp as a separate job from
27193 vect.exp with unsorted.exp and $(dg_target_exps) separately from
27194 struct-layout-1.exp with stackalign.exp.
27196 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
27198 PR middle-end/56461
27199 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
27200 load_index sbitmap even if some bit in it isn't set.
27202 PR middle-end/56461
27203 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
27204 (discover_iteration_bound_by_body_walk): Change queues to
27205 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
27206 spelling in comment. Call safe_push on queues[bound_index] directly.
27207 Release queues[queue_index] in every iteration unconditionally.
27208 Release bounds vector.
27210 PR middle-end/56461
27211 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
27212 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
27213 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
27216 2013-03-05 Richard Biener <rguenther@suse.de>
27219 * tree-inline.c (remap_blocks_to_null): New function.
27220 (expand_call_inline): When expanding a call stmt without
27221 an associated block inline remap all callee blocks to NULL.
27223 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27225 PR rtl-optimization/56494
27226 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
27227 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
27228 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
27230 PR middle-end/56461
27231 * sel-sched-ir.c (free_sched_pools): Release
27232 succs_info_pool.stack[succs_info_pool.max_top] vectors too
27233 if succs_info_pool.max_top isn't -1.
27236 * opts.c (opts_obstack, opts_concat): Moved to...
27237 * opts-common.c (opts_obstack, opts_concat): ... here.
27239 2013-03-04 Jakub Jelinek <jakub@redhat.com>
27241 PR middle-end/56461
27242 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
27244 2013-03-04 Martin Jambor <mjambor@suse.cz>
27246 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
27247 all appropriate places.
27249 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
27251 PR tree-optimization/56424
27252 * ipa-split.c (split_function): Do not set the RSO flag if result is
27253 not by reference and its type is a register type.
27255 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
27257 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
27258 (microblaze_legitimate_pic_operand): Likewise
27259 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
27260 new function microblaze_legitimate_pic_operand
27261 * config/microblaze/microblaze-protos.h
27262 (microblaze_legitimate_pic_operand): Declare.
27264 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27266 * config/microblaze/predicates.md (call_insn_simple_operand):
27267 New predicate for supported rtx code types.
27268 * config/microblaze/microblaze.md (call_internal1): Use
27269 call_insn_simple_operand predicate.
27271 2013-03-04 Jakub Jelinek <jakub@redhat.com>
27273 PR middle-end/56461
27274 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
27275 partitions.ordered_remove.
27277 PR middle-end/56461
27278 * tree-vect-stmts.c (vectorizable_conversion): Don't call
27279 vec_oprnds0.create (1) for modifier == NONE.
27281 PR middle-end/56461
27282 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
27283 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
27284 vec_oprnds1 right before pushing anything to it for
27287 PR middle-end/56461
27288 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
27289 set nbbs to 0 instead of having separate code path.
27290 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
27291 instead of false as last argument if returning NULL.
27293 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
27295 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
27296 the attribute is now called "target" instead of "option".
27297 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
27298 * doc/tm.texi.in (Target Attributes): Likewise document the correct
27299 attribute/pragma name for TARGET_OPTION_VALID_P and
27300 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
27301 * doc/tm.texi: Regenerated.
27303 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
27305 * config/microblaze/microblaze.c:
27306 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
27307 * config/microblaze/microblaze.h: Add -mxl-reorder to
27309 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
27310 instructions emitted if TARGET_REORDER.
27311 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
27312 or 0 for -m/-mno case, but initialises as 2 to detect default use case
27315 2013-03-01 Xinliang David Li <davidxl@google.com>
27317 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
27320 2013-03-01 Jakub Jelinek <jakub@redhat.com>
27322 PR middle-end/56461
27323 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
27324 vector even when returning true. Fix up function comment formatting.
27326 PR middle-end/56461
27327 * ira-build.c (ira_loop_nodes_count): New variable.
27328 (create_loop_tree_nodes): Initialize it.
27329 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
27331 PR middle-end/56461
27332 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
27333 method on dr_chain and result_chain.
27334 * tree-vect-stmts.c (vectorizable_store): Only call
27335 result_chain.create if j == 0.
27337 PR middle-end/56461
27338 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
27339 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
27340 before overwriting it.
27342 2013-03-01 Tobias Burnus <burnus@net-b.de>
27344 * doc/extended.texi (C Extensions): Change order in @menu
27346 (Other MIPS Built-in Functions): Move last MIPS entry before
27347 "picoChip Built-in Functions".
27348 (SH Built-in Functions): Move after RX Built-in Functions.
27349 * doc/gcc.texi (Introduction): Change order in @menu
27351 * doc/md.texi (Constraints): Ditto.
27352 * gty.texi (Type Information): Ditto.
27353 (User-provided marking routines for template types): Make
27355 * doc/invoke.texi (AArch64 Options): Move before
27356 "Adapteva Epiphany Options".
27358 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
27359 Jakub Jelinek <jakub@redhat.com>
27362 * asan.c (gate_asan): Lookup no_sanitize_address instead of
27363 no_address_safety_analysis attribute.
27364 * doc/extend.texi (no_address_safety_attribute): Rename to
27365 no_sanitize_address attribute, mention no_address_safety_analysis
27366 attribute as deprecated alias.
27368 2013-02-28 Jakub Jelinek <jakub@redhat.com>
27370 PR middle-end/56461
27371 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
27372 type to vec<vec<tree> > *.
27373 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
27374 to be vec<tree> instead of vec<tree> *, set vec_defs
27375 to vNULL and call vec_defs.create (number_of_vects), adjust other
27377 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
27378 vectorizable_condition): Adjust vect_get_slp_defs callers.
27380 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27382 * config/aarch64/aarch64.c
27383 (aarch64_float_const_representable): Remove unused variable.
27385 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27387 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
27389 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27391 * config/aarch64/aarch64-builtins.c
27392 (aarch64_init_simd_builtins): Make static.
27394 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27396 * config/aarch64/aarch64.c
27397 (aarch64_simd_make_constant): Make static.
27399 2013-02-28 Martin Jambor <mjambor@suse.cz>
27401 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
27402 with no initialization to the RHS of debug statements.
27404 2013-02-28 Martin Jambor <mjambor@suse.cz>
27406 PR tree-optimization/56294
27407 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
27409 (get_access_replacement): Do not call create_access_replacement.
27410 Assert a replacement exists.
27411 (get_repl_default_def_ssa_name): Create the replacement declaration
27414 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
27416 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
27417 final_end_function.
27419 2013-02-28 Marek Polacek <polacek@redhat.com>
27421 PR rtl-optimization/56466
27422 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
27423 if we're changing a loop.
27424 (peel_loops_completely): Likewise.
27426 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
27429 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
27431 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27434 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
27435 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
27436 INTX_FTYPE_FX, FX_FTYPE_INTX.
27437 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
27439 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27441 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
27442 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
27443 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
27444 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
27445 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
27446 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
27447 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
27448 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
27449 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
27450 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
27451 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
27452 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
27453 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
27454 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
27455 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
27456 (avrxmega6): Increase max flash segments from 5 to 6.
27457 * config/avr/t-multilib: Regenerate.
27458 * config/avr/avr-tables.opt: Regenerate.
27459 * doc/avr-mmcu.texi: Regenerate.
27461 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27463 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
27464 (avr_device_to_arch): Rename to avr_device_to_ld.
27465 (avr_device_to_as): New prototype.
27466 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
27467 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
27468 * config/avr/driver-avr.c (avr_device_to_as): New.
27469 (avr_device_to_arch): Rename to avr_device_to_ld.
27471 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27473 PR middle-end/56461
27474 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
27475 method on dr_chain and result_chain.
27477 PR middle-end/56461
27478 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
27479 pointer_set_destroy on not_executed_last_iteration.
27481 PR middle-end/56461
27482 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
27484 PR middle-end/56461
27485 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
27486 FOR_EACH_DEFINED_FUNCTION when freeing state.
27488 PR middle-end/56461
27489 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
27491 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
27494 PR middle-end/56461
27495 * ipa-cp.c (decide_whether_version_node): Call vec_free on
27496 known_aggs[i].items and release known_aggs vector.
27498 PR middle-end/56461
27499 * ipa-reference.c (propagate): Free node_info even for alias nodes.
27501 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27503 * config/microblaze/microblaze.c (microblaze_emit_compare):
27504 Use xor for EQ/NE comparisions.
27505 * config/microblaze/microblaze.md (cstoresf4): Add constraints
27506 (cbranchsf4): Adjust operator to comparison_operator.
27508 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27510 PR middle-end/56461
27511 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
27513 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
27514 vec_safe_push, always update *slot.
27515 (redirect_edge_var_map_clear): Use vec_free.
27516 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
27517 (free_var_map_entry): Use vec_free.
27518 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
27519 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
27521 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
27523 PR middle-end/45472
27524 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
27525 when the may_trap_p bit of the exprs being merged differs.
27526 Reorder tests for speculativeness in the logical and operator.
27528 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27530 * incpath.c (add_standard_paths): Use reconcat instead of concat
27531 where appropriate and avoid leaking memory.
27533 * opts.h: Include obstack.h.
27534 (opts_concat): New prototype.
27535 (opts_obstack): New declaration.
27536 * opts.c (opts_concat): New function.
27537 (opts_obstack): New variable.
27538 (init_options_struct): Call gcc_init_obstack on opts_obstack.
27539 (finish_options): Use opts_concat instead of concat
27540 and XOBNEWVEC instead of XNEWVEC.
27541 * opts-common.c (generate_canonical_option, decode_cmdline_option,
27542 generate_option): Likewise.
27543 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
27544 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
27547 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
27548 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
27550 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27552 PR middle-end/56461
27553 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
27555 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27557 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
27558 (arm_block_move_unaligned_straight): Likewise.
27559 (arm_adjust_block_mem): Likewise.
27561 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27564 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
27565 temp, cond and label.
27566 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
27569 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
27570 * config/c6x/c6x.h (dbx_register_map): Update declaration.
27573 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
27574 of prologue/epilogue functions.
27577 * config/tilegx/tilegx.c (tilegx_expand_prologue):
27578 Remove unused variable cfa_offset.
27579 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
27582 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
27583 type promotion to unsigned.
27586 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
27587 for HOST_WIDE_INT of 32 bit / same size as int.
27588 (arm_block_move_unaligned_straight): Likewise.
27589 (arm_adjust_block_mem): Likewise.
27592 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
27595 2013-02-26 Marek Polacek <polacek@redhat.com>
27597 PR tree-optimization/56426
27598 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
27600 2013-02-26 Richard Biener <rguenther@suse.de>
27603 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
27604 unused variable loops.
27606 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27608 PR tree-optimization/56448
27609 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
27610 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
27611 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
27612 later operands of the references, or even first operand for
27613 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
27615 PR tree-optimization/56443
27616 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
27617 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
27618 to type_for_mode langhook.
27620 2013-02-25 Matt Turner <mattst88@gmail.com>
27622 * doc/invoke.texi: Document r4700.
27624 2013-02-25 Richard Biener <rguenther@suse.de>
27626 PR tree-optimization/56175
27627 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
27629 (simplify_bitwise_binary): ... here. Also guard the conversion
27630 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
27632 2013-02-25 Catherine Moore <clm@codesourcery.com>
27635 2013-02-24 Catherine Moore <clm@codesourcery.com>
27636 Maciej W. Rozycki <macro@codesourcery.com>
27637 Tom de Vries <tom@codesourcery.com>
27638 Nathan Sidwell <nathan@codesourcery.com>
27639 Iain Sandoe <iain@codesourcery.com>
27640 Nathan Froyd <froydnj@codesourcery.com>
27641 Chao-ying Fu <fu@mips.com>
27643 * doc/extend.texi: (micromips, nomicromips, nocompression):
27644 Document new function attributes.
27645 * doc/invoke.texi (minterlink-compressed, mmicromips,
27646 m14k, m14ke, m14kec): Document new options.
27647 (minterlink-mips16): Update documentation.
27648 * doc/md.texi (ZC, ZD): Document new constraints.
27649 * configure.ac (gcc_cv_as_micromips): Check if linker
27650 supports the .set micromips directive.
27651 * configure: Regenerate.
27652 * config.in: Regenerate.
27653 * config/mips/mips-tables.opt: Regenerate.
27654 * config/mips/micromips.md: New file.
27655 * constraints.md (ZC, AD): New constraints.
27656 * config/mips/predicates.md (movep_src_register): New predicate.
27657 (movep_src_operand): New predicate.
27658 (non_volatile_mem_operand): New predicate.
27659 * config/mips/mips.md (multimem): New type.
27660 (length): Differentiate between 17-bit and 18-bit branch offsets.
27661 (MOVEP1, MOVEP2): New mode iterator.
27662 (mov_<load>l): Use ZC constraint.
27663 (mov_<load>r): Likewise.
27664 (mov_<store>l): Likewise.
27665 (mov_<store>r): Likewise.
27666 (*branch_equality<mode>_inverted): Add microMIPS support.
27667 (*branch_equality<mode>): Likewise.
27668 (*jump_absolute): Likewise.
27669 (indirect_jump_<mode>): Likewise.
27670 (tablejump_<mode>): Likewise.
27671 (<optab>_internal): Likewise.
27672 (sibcall_internal): Likewise.
27673 (sibcall_value_internal): Likewise.
27674 (prefetch): Use constraint ZD.
27675 * config/mips/mips.opt (minterlink-compressed): New option.
27676 (minterlink-mips16): Now an alias for minterlink-compressed.
27677 (mmicromips): New option.
27678 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27679 (compare_and_swap_12): Likewise.
27680 (sync_add<mode>): Likewise.
27681 (sync_<optab>_12): Likewise.
27682 (sync_old_<optab>_12): Likewise.
27683 (sync_new_<optab>_12): Likewise.
27684 (sync_nand_12): Likewise.
27685 (sync_old_nand_12): Likewise.
27686 (sync_new_nand_12): Likewise.
27687 (sync_sub<mode>): Likewise.
27688 (sync_old_add<mode>): Likewise.
27689 (sync_old_sub<mode>): Likewise.
27690 (sync_new_add<mode>): Likewise.
27691 (sync_new_sub<mode>): Likewise.
27692 (sync_<optab><mode>): Likewise.
27693 (sync_old_<optab><mode>): Likewise.
27694 (sync_new_<optab><mode>): Likewise.
27695 (sync_nand<mode>): Likewise.
27696 (sync_old_nand<mode>): Likewise.
27697 (sync_new_nand<mode>): Likewise.
27698 (sync_lock_test_and_set<mode>): Likewise.
27699 (test_and_set_12): Likewise.
27700 (atomic_compare_and_swap<mode>): Likewise.
27701 (atomic_exchange<mode>_llsc): Likewise.
27702 (atomic_fetch_add<mode>_llsc): Likewise.
27703 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27704 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27705 (umips_save_restore_pattern_p): Likewise.
27706 (umips_load_store_pair_p): Likewise.
27707 (umips_output_load_store_pair): Likewise.
27708 (umips_movep_target_p): Likewise.
27709 (umips_12bit_offset_address_p): Likewise.
27710 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27711 (mips_base_mips16): Rename this...
27712 (mips_base_compression_flags): ...to this. Update all uses.
27713 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27714 (mips_mips16_decl_p): Delete.
27715 (mips_nomips16_decl_p): Delete.
27716 (mips_get_compress_on_flags): New function.
27717 (mips_get_compress_off_flags): New function.
27718 (mips_get_compress_mode): New function.
27719 (mips_get_compress_on_name): New function.
27720 (mips_get_compress_off_name): New function.
27721 (mips_insert_attributes): Support multiple compression types.
27722 (mips_merge_decl_attributes): Likewise.
27723 (umips_12bit_offset_address_p): New function.
27724 (mips_start_function_definition): Emit .set micromips directive.
27725 (mips_call_may_need_jalx_p): New function.
27726 (mips_function_ok_for_sibcall): Add microMIPS support.
27727 (mips_print_operand_punctuation): Support short delay slots and
27729 (umips_swm_mask, umips_swm_encoding): New.
27730 (umips_build_save_restore): New function.
27731 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27732 (was_mips16_p): Remove.
27733 (old_compression_mode): New.
27734 (mips_set_compression_mode): New function.
27735 (mips_set_current_function): Add microMIPS support.
27736 (mips_option_override): Likewise.
27737 (umips_save_restore_pattern_p): New function.
27738 (umips_output_save_restore): New function.
27739 (umips_load_store_pair_p_1): New function.
27740 (umips_load_store_pair_p): New function.
27741 (umips_output_load_store_pair_1): New function.
27742 (umips_output_load_store_pair): New function.
27743 (umips_movep_target_p) New function.
27744 (mips_prepare_pch_save): Add microMIPS support.
27745 * config/mips/mips.h (TARGET_COMPRESSION): New.
27746 (TARGET_CPU_CPP_BUILTINS): Update macro
27747 to use new compression flags and to support microMIPS.
27748 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27749 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27750 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27751 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27752 (ASM_SPEC): Support mmicromips and mno-micromips.
27753 (M16STORE_REG_P): New macro.
27754 (MIPS_CALL): Support TARGET_MICROMIPS.
27755 (MICROMIPS_J): New macro.
27756 (mips_base_mips16): Rename this...
27757 (mips_base_compression_flags): ...to this.
27758 (UMIPS_12BIT_OFFSET_P): New macro.
27759 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27760 (MULTILIB_DIRNAMES): Likewise.
27762 2013-02-25 Tom de Vries <tom@codesourcery.com>
27764 PR rtl-optimization/56131
27765 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
27766 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
27767 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
27769 2013-02-25 Tobias Burnus <burnus@net-b.de>
27771 * doc/invoke.texi (-fsanitize=): Move from optimization
27772 to debugging options.
27774 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27776 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
27778 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27779 Alexander Monakov <amonakov@ispras.ru>
27781 PR middle-end/56077
27782 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
27783 flush pending lists also on non-jumps. Adjust comment.
27785 2013-02-24 Catherine Moore <clm@codesourcery.com>
27786 Maciej W. Rozycki <macro@codesourcery.com>
27787 Tom de Vries <tom@codesourcery.com>
27788 Nathan Sidwell <nathan@codesourcery.com>
27789 Iain Sandoe <iain@codesourcery.com>
27790 Nathan Froyd <froydnj@codesourcery.com>
27791 Chao-ying Fu <fu@mips.com>
27793 * doc/extend.texi: (micromips, nomicromips, nocompression):
27794 Document new function attributes.
27795 * doc/invoke.texi (minterlink-compressed, mmicromips,
27796 m14k, m14ke, m14kec): Document new options.
27797 (minterlink-mips16): Update documentation.
27798 * doc/md.texi (ZC, ZD): Document new constraints.
27799 * configure.ac (gcc_cv_as_micromips): Check if linker
27800 supports the .set micromips directive.
27801 * configure: Regenerate.
27802 * config.in: Regenerate.
27803 * config/mips/mips-tables.opt: Regenerate.
27804 * config/mips/micromips.md: New file.
27805 * constraints.md (ZC, AD): New constraints.
27806 * config/mips/predicates.md (movep_src_register): New predicate.
27807 (movep_src_operand): New predicate.
27808 (non_volatile_mem_operand): New predicate.
27809 * config/mips/mips.md (multimem): New type.
27810 (length): Differentiate between 17-bit and 18-bit branch offsets.
27811 (MOVEP1, MOVEP2): New mode iterator.
27812 (mov_<load>l): Use ZC constraint.
27813 (mov_<load>r): Likewise.
27814 (mov_<store>l): Likewise.
27815 (mov_<store>r): Likewise.
27816 (*branch_equality<mode>_inverted): Add microMIPS support.
27817 (*branch_equality<mode>): Likewise.
27818 (*jump_absolute): Likewise.
27819 (indirect_jump_<mode>): Likewise.
27820 (tablejump_<mode>): Likewise.
27821 (<optab>_internal): Likewise.
27822 (sibcall_internal): Likewise.
27823 (sibcall_value_internal): Likewise.
27824 (prefetch): Use constraint ZD.
27825 * config/mips/mips.opt (minterlink-compressed): New option.
27826 (minterlink-mips16): Now an alias for minterlink-compressed.
27827 (mmicromips): New option.
27828 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27829 (compare_and_swap_12): Likewise.
27830 (sync_add<mode>): Likewise.
27831 (sync_<optab>_12): Likewise.
27832 (sync_old_<optab>_12): Likewise.
27833 (sync_new_<optab>_12): Likewise.
27834 (sync_nand_12): Likewise.
27835 (sync_old_nand_12): Likewise.
27836 (sync_new_nand_12): Likewise.
27837 (sync_sub<mode>): Likewise.
27838 (sync_old_add<mode>): Likewise.
27839 (sync_old_sub<mode>): Likewise.
27840 (sync_new_add<mode>): Likewise.
27841 (sync_new_sub<mode>): Likewise.
27842 (sync_<optab><mode>): Likewise.
27843 (sync_old_<optab><mode>): Likewise.
27844 (sync_new_<optab><mode>): Likewise.
27845 (sync_nand<mode>): Likewise.
27846 (sync_old_nand<mode>): Likewise.
27847 (sync_new_nand<mode>): Likewise.
27848 (sync_lock_test_and_set<mode>): Likewise.
27849 (test_and_set_12): Likewise.
27850 (atomic_compare_and_swap<mode>): Likewise.
27851 (atomic_exchange<mode>_llsc): Likewise.
27852 (atomic_fetch_add<mode>_llsc): Likewise.
27853 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27854 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27855 (umips_save_restore_pattern_p): Likewise.
27856 (umips_load_store_pair_p): Likewise.
27857 (umips_output_load_store_pair): Likewise.
27858 (umips_movep_target_p): Likewise.
27859 (umips_12bit_offset_address_p): Likewise.
27860 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27861 (mips_base_mips16): Rename this...
27862 (mips_base_compression_flags): ...to this. Update all uses.
27863 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27864 (mips_mips16_decl_p): Delete.
27865 (mips_nomips16_decl_p): Delete.
27866 (mips_get_compress_on_flags): New function.
27867 (mips_get_compress_off_flags): New function.
27868 (mips_get_compress_mode): New function.
27869 (mips_get_compress_on_name): New function.
27870 (mips_get_compress_off_name): New function.
27871 (mips_insert_attributes): Support multiple compression types.
27872 (mips_merge_decl_attributes): Likewise.
27873 (umips_12bit_offset_address_p): New function.
27874 (mips_start_function_definition): Emit .set micromips directive.
27875 (mips_call_may_need_jalx_p): New function.
27876 (mips_function_ok_for_sibcall): Add microMIPS support.
27877 (mips_print_operand_punctuation): Support short delay slots and
27879 (umips_swm_mask, umips_swm_encoding): New.
27880 (umips_build_save_restore): New function.
27881 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27882 (was_mips16_p): Remove.
27883 (old_compression_mode): New.
27884 (mips_set_compression_mode): New function.
27885 (mips_set_current_function): Add microMIPS support.
27886 (mips_option_override): Likewise.
27887 (umips_save_restore_pattern_p): New function.
27888 (umips_output_save_restore): New function.
27889 (umips_load_store_pair_p_1): New function.
27890 (umips_load_store_pair_p): New function.
27891 (umips_output_load_store_pair_1): New function.
27892 (umips_output_load_store_pair): New function.
27893 (umips_movep_target_p) New function.
27894 (mips_prepare_pch_save): Add microMIPS support.
27895 * config/mips/mips.h (TARGET_COMPRESSION): New.
27896 (TARGET_CPU_CPP_BUILTINS): Update macro
27897 to use new compression flags and to support microMIPS.
27898 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27899 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27900 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27901 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27902 (ASM_SPEC): Support mmicromips and mno-micromips.
27903 (M16STORE_REG_P): New macro.
27904 (MIPS_CALL): Support TARGET_MICROMIPS.
27905 (MICROMIPS_J): New macro.
27906 (mips_base_mips16): Rename this...
27907 (mips_base_compression_flags): ...to this.
27908 (UMIPS_12BIT_OFFSET_P): New macro.
27909 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27910 (MULTILIB_DIRNAMES): Likewise.
27912 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27915 * target-globals.c (save_target_globals): For init_reg_sets and
27916 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27918 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27920 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27921 * config/aarch64/t-aarch64
27922 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27924 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27926 PR inline-asm/56148
27927 * lra-constraints.c (process_alt_operands): Reload operand
27928 conflicting with earlier clobber only if no more other conflicting
27931 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27934 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27935 if not linking a shared library.
27937 2013-02-22 Seth LaForge <sethml@google.com>
27939 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27941 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27943 * config/arm/arm.md (split for extendsidi): Update condition.
27944 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27945 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27946 (qhs_zextenddi_cstr): Likewise.
27948 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27950 PR middle-end/56420
27951 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27952 avoid signed wrapping.
27953 (expand_mult): Handle properly multiplication by
27954 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27955 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27956 in the compiler if coeff is HOST_WIDE_INT_MIN.
27957 (expand_divmod): Don't make ext_op1 static, change it's type to
27958 uhwi. Avoid undefined behavior in -INTVAL (op1).
27960 PR rtl-optimization/50339
27961 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27963 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27964 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27965 into splitting_ashiftrt field.
27966 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27968 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27971 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27973 PR middle-end/56108
27974 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27975 are sure to go irrevocable.
27977 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27979 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27980 scalars are valid operands.
27982 2013-02-21 Martin Jambor <mjambor@suse.cz>
27984 PR tree-optimization/56310
27985 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27986 only matching indices and non-negative final offsets.
27987 (intersect_aggregates_with_edge): Pass src_idx to
27988 agg_replacements_to_vector. Pass src_idx insstead of index to
27989 intersect_with_agg_replacements.
27991 2013-02-21 Martin Jambor <mjambor@suse.cz>
27993 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
27994 instead of hard-wired defaults.
27996 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
27998 * doc/invoke.texi (MIPS Options): Update documentation of the
27999 floating-point multiply-accumulate instruction restrictions.
28001 2013-02-21 Kostya Serebryany <kcc@google.com>
28003 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
28004 asan_shadow_offset on x86_64 linux.
28006 2013-02-21 Richard Biener <rguenther@suse.de>
28008 PR tree-optimization/56415
28010 2013-02-11 Richard Biener <rguenther@suse.de>
28012 PR tree-optimization/56273
28013 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
28016 2013-02-21 Jakub Jelinek <jakub@redhat.com>
28019 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
28022 PR inline-asm/56405
28023 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
28024 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
28026 2013-02-20 Jan Hubicka <jh@suse.cz>
28028 PR tree-optimization/56265
28029 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
28030 when target is referenced for first time.
28032 2013-02-20 Richard Biener <rguenther@suse.de>
28034 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
28035 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
28036 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
28037 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
28038 not return anything.
28039 (rename_ssa_copies): Do not remove unused locals.
28040 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
28041 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
28042 * passes.c (execute_function_todo): Do not schedule unused locals
28043 removal if cleanup_tree_cfg did something.
28044 * tree-ssa-live.c (remove_unused_locals): Dump statistics
28045 about the number of removed locals.
28047 2013-02-20 Richard Biener <rguenther@suse.de>
28049 PR tree-optimization/56398
28050 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
28052 2013-02-20 Martin Jambor <mjambor@suse.cz>
28054 PR tree-optimization/55334
28055 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
28056 restricted pointers to arrays.
28058 2013-02-20 Richard Biener <rguenther@suse.de>
28059 Jakub Jelinek <jakub@redhat.com>
28061 PR tree-optimization/56396
28062 * tree-ssa-ccp.c (n_const_val): New static variable.
28063 (get_value): Return NULL for SSA names we don't have a lattice
28065 (ccp_initialize): Initialize n_const_val.
28066 * tree-ssa-copy.c (n_copy_of): New static variable.
28067 (init_copy_prop): Initialize n_copy_of.
28068 (get_value): Return NULL_TREE for SSA names we don't have a
28071 2013-02-20 Martin Jambor <mjambor@suse.cz>
28073 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
28075 2013-02-20 Richard Biener <rguenther@suse.de>
28077 * genpreds.c (write_lookup_constraint): Do not compare first
28078 letter of the constraint again.
28080 2013-02-20 Richard Biener <rguenther@suse.de>
28082 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
28084 (get_use_iv_cost): Terminate hashtable walk when coming across
28087 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
28089 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
28090 reassociation for avx2 targets.
28092 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28094 * config/microblaze/microblaze.c: microblaze_has_clz = 0
28095 Add version check for v8.10.a to enable microblaze_has_clz
28096 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
28097 version and TARGET_PATTERN_COMPARE check
28098 * config/microblaze/microblaze.md: New clzsi2 instruction
28100 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28102 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
28103 function before branching.
28105 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
28107 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
28109 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
28111 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
28113 PR middle-end/55889
28114 * sel-sched.c: Include ira.h.
28115 (implicit_clobber_conflict_p): New function.
28116 (moveup_expr): Use it.
28117 * Makefile.in (sel-sched.o): Depend on ira.h.
28119 2013-02-19 Richard Biener <rguenther@suse.de>
28121 PR tree-optimization/56384
28122 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
28123 (vn_hash_type): Split out from ...
28124 (vn_hash_constant_with_type): ... here.
28125 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
28126 (vn_phi_eq): Compare types from vn_phi_s structure.
28127 (vn_phi_lookup): Populate vn_phi_s type.
28128 (vn_phi_insert): Likewise.
28130 2013-02-19 Jakub Jelinek <jakub@redhat.com>
28132 PR tree-optimization/56350
28133 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
28134 if haven't found reduction or nested cycle operand, rather than
28135 asserting we must find it.
28137 PR tree-optimization/56381
28138 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
28141 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
28142 Jakub Jelinek <jakub@redhat.com>
28145 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
28146 (swap_optab_enable): Same.
28147 (init_all_optabs): Use argument instead of global.
28148 * tree.h (struct tree_optimization_option): New field target_optabs.
28149 * expr.h (init_all_optabs): Add argument to prototype.
28150 (TREE_OPTIMIZATION_OPTABS): New.
28151 (save_optabs_if_changed): Protoize.
28152 * optabs.h: Declare this_fn_optabs.
28153 * optabs.c (save_optabs_if_changed): New.
28154 Declare this_fn_optabs.
28155 (init_optabs): Add argument to init_all_optabs() call.
28156 * function.c (invoke_set_current_function_hook): Handle per
28158 * function.h (struct function): New field optabs.
28159 * config/mips/mips.c (mips_set_mips16_mode): Handle when
28160 optimization_current_node has changed.
28161 * target-globals.h (save_target_globals_default_opts): Protoize.
28162 * target-globals.c (save_target_globals_default_opts): New.
28164 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28167 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
28168 registers %fr12 and %fr12R as call used.
28171 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
28172 and HImode, require all displacements to be an integer multiple of
28174 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
28175 only allow QImode and HImode when reload is in progress and strict is
28176 true. Likewise for symbolic addresses. Use base14_operand to check
28177 displacements in REG+BASE addresses.
28179 2013-02-18 Richard Biener <rguenther@suse.de>
28181 PR tree-optimization/56366
28182 * tree-vect-loop.c (get_initial_def_for_induction): Properly
28183 handle sign-conversion of outer-loop initial induction value.
28185 2013-02-18 Richard Biener <rguenther@suse.de>
28187 PR middle-end/56349
28188 * cfghooks.c (merge_blocks): If we merge a latch into another
28189 block adjust references to it.
28190 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
28191 (verify_loop_structure): Verify that a recorded latch is in fact
28194 2013-02-18 Richard Biener <rguenther@suse.de>
28196 PR tree-optimization/56321
28197 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
28198 order SSA name release and virtual operand unlinking.
28200 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28202 * config/microblaze/microblaze.md (save_stack_block): Define.
28203 (restore_stack_block): Likewise.
28205 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28207 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
28208 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
28209 * config/microblaze/microblaze.c (microblaze_option_override):
28210 Bail out early for PIC modes when target does not support PIC.
28212 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28214 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
28215 Replace with a microblaze version.
28216 (microblaze_trampoline_init): Adapt for microblaze.
28217 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
28220 2013-02-16 Jakub Jelinek <jakub@redhat.com>
28221 Dodji Seketeli <dodji@redhat.com>
28224 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
28225 (instrument_mem_region_access): Do not forget to always put
28226 instrumentation of the of 'base' and 'base + len' in a "if (len !=
28227 0) statement, even for cases where either 'base' or 'base + len'
28228 are not instrumented -- because they have been previously
28229 instrumented. Simplify the logic by putting all the statements
28230 instrument 'base + len' inside a sequence, and then insert that
28231 sequence right before the current insertion point. Then, to
28232 instrument 'base + len', just get an iterator on that statement.
28233 And do not forget to update the pointer to iterator the function
28234 received as argument.
28236 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
28238 PR rtl-optimization/56348
28239 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
28241 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
28243 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
28244 (clean_graph_dump_file): Pass base to start_graph_dump.
28246 2013-02-14 Richard Henderson <rth@redhat.com>
28249 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
28251 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
28253 * collect2-aix.h: Define F_LOADONLY.
28255 2013-02-14 Richard Biener <rguenther@suse.de>
28258 * varasm.c (output_constant_def_1): Get the decl representing
28259 the constant as argument.
28260 (output_constant_def): Wrap output_constant_def_1.
28261 (make_decl_rtl): Use output_constant_def_1 with the decl
28262 representing the constant.
28263 (build_constant_desc): Optionally re-use a decl already
28264 representing the constant.
28265 (tree_output_constant_def): Adjust.
28267 2013-02-14 Dodji Seketeli <dodji@redhat.com>
28270 * asan.c (instrument_builtin_call): Really put the length of the
28271 second source argument into src1_len.
28273 2013-02-13 Jakub Jelinek <jakub@redhat.com>
28275 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
28276 argument. If it is false, don't create edge from then_bb to
28278 (insert_if_then_before_iter): Pass true to it.
28279 (build_check_stmt): Pass false to it.
28280 (transform_statements): Flush hash table only on extended basic
28281 block boundaries, rather than at the beginning of every bb.
28282 Don't flush hash table on nonfreeing_call_p calls.
28283 * tree-flow.h (nonfreeing_call_p): New prototype.
28284 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
28286 2013-02-13 David S. Miller <davem@davemloft.net>
28288 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
28290 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
28293 * ira.c (max_regno_before_ira): Move from ...
28295 (fix_reg_equiv_init): Use max_regno_before_ira instead of
28298 2013-02-13 Jakub Jelinek <jakub@redhat.com>
28300 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
28302 2013-02-13 Richard Biener <rguenther@suse.de>
28305 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
28306 globals in MEM_REFs.
28308 2013-02-13 Richard Biener <rguenther@suse.de>
28310 * loop-init.c (loop_optimizer_init): Clear loop state when
28311 re-initializing preserved loops.
28312 * loop-unswitch.c (unswitch_single_loop): Return whether
28313 we unswitched the loop. Do not verify loop state here.
28314 (unswitch_loops): When we unswitched a loop discover new loops.
28316 2013-02-13 Kostya Serebryany <kcc@google.com>
28318 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
28320 * sanitizer.def: Rename __asan_init to __asan_init_v1.
28322 2013-02-12 Dodji Seketeli <dodji@redhat.com>
28324 Avoid instrumenting duplicated memory access in the same basic block
28325 * Makefile.in (asan.o): Add new dependency on hash-table.h
28326 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
28327 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
28328 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
28329 (free_mem_ref_resources, has_mem_ref_been_instrumented)
28330 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
28331 (get_mem_ref_of_assignment): New functions.
28332 (get_mem_refs_of_builtin_call): Extract from
28333 instrument_builtin_call and tweak a little bit to make it fit with
28335 (instrument_builtin_call): Use the new
28336 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
28337 of is_gimple_builtin_call.
28338 (instrument_derefs, instrument_mem_region_access): Insert the
28339 instrumented memory reference into the hash table.
28340 (maybe_instrument_assignment): Renamed instrument_assignment into
28341 this, and change it to advance the iterator when instrumentation
28342 actually happened and return true in that case. This makes it
28343 homogeneous with maybe_instrument_assignment, and thus give a
28344 chance to callers to be more 'regular'.
28345 (transform_statements): Clear the memory reference hash table
28346 whenever we enter a new BB, when we cross a function call, or when
28347 we are done transforming statements. Use
28348 maybe_instrument_assignment instead of instrumentation. No more
28349 need to special case maybe_instrument_assignment and advance the
28350 iterator after calling it; it's now handled just like
28351 maybe_instrument_call. Update comment.
28353 2013-02-13 Richard Biener <rguenther@suse.de>
28355 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
28356 Fix loop discovery code.
28358 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
28360 PR inline-asm/56148
28361 * lra-constraints.c (process_alt_operands): Match early clobber
28362 operand with itself. Check conflicts with earlyclobber only if
28363 the operand is not reloaded. Prefer to reload conflicting operand
28364 if earlyclobber and matching operands are the same.
28366 2013-02-12 Richard Biener <rguenther@suse.de>
28369 * lto-streamer-out.c (write_symbol): Do not output symbols
28370 for hard register variables.
28372 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
28375 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
28376 (umulsidi3_insn, mulsidi3_insn): New insns.
28378 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
28380 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
28381 (struct tune_params): Add vec_costs field.
28382 * config/arm/arm.c (arm_builtin_vectorization_cost)
28383 (arm_add_stmt_cost): New functions.
28384 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
28385 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
28386 (arm_default_vec_cost): New struct of type cpu_vec_costs.
28387 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
28388 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
28389 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
28390 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
28392 2013-02-12 Richard Biener <rguenther@suse.de>
28395 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
28396 decls again if possible.
28398 2013-02-12 Richard Biener <rguenther@suse.de>
28400 PR middle-end/56288
28401 * tree-ssa.c (verify_ssa_name): Fix check, move
28402 SSA_NAME_IN_FREE_LIST check up.
28404 2013-02-12 Jakub Jelinek <jakub@redhat.com>
28405 Steven Bosscher <steven@gcc.gnu.org>
28407 PR rtl-optimization/56151
28408 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
28409 equal to op0 or op1, and last_insn pattern is CODE operation
28410 with MEM dest and one of the operands matches that MEM.
28412 2013-02-11 Sriraman Tallam <tmsriram@google.com>
28414 * doc/extend.texi: Document Function Multiversioning and "default"
28415 parameter string to target attribute.
28416 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
28417 target attribute parameter is "default".
28418 (ix86_compare_version_priority): Remove checks for target attribute.
28419 (ix86_mangle_function_version_assembler_name): Change error to sorry.
28420 Remove check for target attribute equal to NULL. Add assert.
28421 (ix86_generate_version_dispatcher_body): Change error to sorry.
28423 2013-02-11 Iain Sandoe <iain@codesourcery.com>
28424 Jack Howarth <howarth@bromo.med.uc.edu>
28425 Patrick Marlier <patrick.marlier@gmail.com>
28428 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
28429 define ENDFILE_SPEC as TM_DESTRUCTOR.
28430 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
28432 2013-02-11 Alexander Potapenko <glider@google.com>
28433 Jack Howarth <howarth@bromo.med.uc.edu>
28434 Jakub Jelinek <jakub@redhat.com>
28437 * config/darwin.c (cdtor_record): Rename ctor_record.
28438 (sort_cdtor_records): Rename sort_ctor_records.
28439 (finalize_dtors): New routine to sort destructors by
28440 priority before use in assemble_integer.
28441 (machopic_asm_out_destructor): Use finalize_dtors if needed.
28443 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
28445 PR rtl-optimization/56275
28446 * simplify-rtx.c (avoid_constant_pool_reference): Check that
28447 offset is non-negative and less than cmode size before
28448 calling simplify_subreg.
28450 2013-02-11 Richard Biener <rguenther@suse.de>
28452 PR tree-optimization/56264
28453 * cfgloop.h (fix_loop_structure): Adjust prototype.
28454 * loop-init.c (fix_loop_structure): Return the number of
28455 newly discovered loops.
28456 * tree-cfgcleanup.c (repair_loop_structures): When new loops
28457 are discovered, do a full loop-closed SSA rewrite.
28459 2013-02-11 Richard Biener <rguenther@suse.de>
28461 PR tree-optimization/56273
28462 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
28464 (check_array_ref): Fix missing newline in dumps.
28465 (search_for_addr_array): Likewise.
28467 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
28469 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
28471 2013-02-09 Jakub Jelinek <jakub@redhat.com>
28474 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
28476 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
28478 PR rtl-optimization/56246
28479 * lra-constraints.c (simplify_operand_subreg): Try to reuse
28481 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
28482 constraints are satisfied.
28484 2013-02-08 Jeff Law <law@redhat.com>
28487 * emit-rtl.c (reg_is_parm_p): New function.
28488 * regs.h (reg_is_parm_p): New prototype.
28489 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
28490 callee-clobbered registers.
28492 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
28495 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
28496 If there is no implicit builtin declaration, just return NULL.
28498 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28500 * config/i386/sse.md (FMAMODEM): New mode iterator.
28501 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
28502 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
28504 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28506 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
28507 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
28508 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
28510 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28512 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
28513 (microblaze*-*-elf): Likewise.
28514 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
28516 * config/microblaze/microblaze-c.c: Add builtin defines for
28517 _LITTLE_ENDIAN and _BIG_ENDIAN.
28518 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
28519 add to TARGET_DEFAULT flags.
28520 Expand ASM_SPEC and LINK_SPEC.
28521 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
28522 * config/microblaze/microblaze.md: Update extendsidi2 and
28523 movdi_internal instructions to use low-order / high-order reg
28525 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
28526 options and inversemask / mask of LITTLE_ENDIAN.
28527 * config/microblaze/t-microblaze: Expand multilib options to
28528 include mlittle-endian (le) and update exceptions patterns.
28530 2013-02-08 Jakub Jelinek <jakub@redhat.com>
28532 PR rtl-optimization/56195
28533 * lra-constraints.c (get_reload_reg): Don't reuse regs
28534 if they have smaller mode than requested, if they have
28535 wider mode than requested, try to return a SUBREG.
28537 PR tree-optimization/56250
28538 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
28539 if type is unsigned and code isn't MULT_EXPR.
28541 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28543 PR tree-optimization/56064
28544 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
28545 bits according to mode.
28546 * fixed-value.h (fixed_from_double_int)
28547 (const_fixed_from_double_int): Adjust comments.
28549 2013-02-08 Richard Biener <rguenther@suse.de>
28552 * lto-streamer.h (struct data_in): Remove current_file, current_line
28553 and current_col members.
28554 * lto-streamer-out.c (lto_output_location): Stream changed bits
28555 en-block for efficiency.
28556 * lto-streamer-in.c (clear_line_info): Remove.
28557 (lto_input_location): Cache current file, line and column
28558 globally via local statics. Read changed bits en-block.
28559 (input_function): Do not call clear_line_info.
28560 (lto_read_body): Likewise.
28561 (lto_input_toplevel_asms): Likewise.
28563 2013-02-08 Michael Matz <matz@suse.de>
28565 PR tree-optimization/52448
28566 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
28567 (nt_call_phase): New static.
28568 (add_or_mark_expr): Only mark accesses with newer phase than any
28570 (nonfreeing_call_p): New.
28571 (nt_init_block): Update nt_call_phase, mark blocks as visited.
28572 (nt_fini_block): Keep blocks marked as visited.
28573 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
28575 2013-02-08 Richard Biener <rguenther@suse.de>
28577 * ira.c (ira): Free broken dominator information.
28579 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28581 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
28583 2013-02-08 Marek Polacek <polacek@redhat.com>
28585 * cfgloop.c (verify_loop_structure): Add more checking of headers.
28587 2013-02-08 Richard Biener <rguenther@suse.de>
28589 PR middle-end/56181
28590 * cfgloop.h (flow_loops_find): Adjust.
28591 (bb_loop_header_p): Declare.
28592 * cfgloop.c (bb_loop_header_p): New function split out from ...
28593 (flow_loops_find): ... here. Adjust function signature,
28594 support incremental loop structure update.
28595 (verify_loop_structure): Cleanup. Verify a loop is a loop.
28596 * cfgloopmanip.c (fix_loop_structure): Move ...
28597 * loop-init.c (fix_loop_structure): ... here.
28598 (apply_loop_flags): Split out from ...
28599 (loop_optimizer_init): ... here.
28600 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
28601 in incremental mode, only remove dead loops here.
28603 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28606 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
28607 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
28608 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
28609 (*round<mode>3.libgcc): New insns for fixed-modes.
28610 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
28611 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
28612 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
28613 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
28614 implementations. Define to __builtin_avr_absFX,
28615 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
28616 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
28617 __builtin_avr_countlsFX, respectively.
28618 * config/avr/avr-c.c (target.h): Include it.
28619 (enum avr_builtin_id): New enum.
28620 (avr_resolve_overloaded_builtin): New static function.
28621 (avr_register_target_pragmas): Use it to set
28622 targetm.resolve_overloaded_builtin.
28623 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
28624 tree nodes used by DEF_BUILTIN.
28625 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
28626 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
28627 <AVR_BUILTIN_xxBITS>: Same.
28629 2013-02-08 Richard Biener <rguenther@suse.de>
28631 * cfgloop.c (verify_loop_structure): Properly handle
28632 a loop exiting to another loop header.
28633 * ira-int.h (ira_loops): Remove.
28634 * ira.c (ira_loops): Remove.
28635 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
28636 (do_reload): Use loop_optimizer_finalize.
28637 * ira-build.c (create_loop_tree_nodes): Use get_loops and
28638 number_of_loops to access the loop tree.
28639 (more_one_region_p): Likewise.
28640 (finish_loop_tree_nodes): Likewise.
28641 (rebuild_regno_allocno_maps): Likewise.
28642 (mark_loops_for_removal): Likewise.
28643 (mark_all_loops_for_removal): Likewise.
28644 (remove_unnecessary_regions): Likewise.
28645 (ira_build): Likewise.
28646 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
28648 2013-02-08 Richard Biener <rguenther@suse.de>
28650 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
28651 * ipa-pure-const.c (analyze_function): Avoid calling
28652 mark_irreducible_loops twice.
28653 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
28655 2013-02-07 David S. Miller <davem@davemloft.net>
28657 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
28659 * var-tracking.c (vt_add_function_parameter): Test the presence of
28660 HAVE_window_save properly and do not remap argument registers when
28661 we have a leaf function.
28663 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
28666 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
28668 * config/i386/i386.c (ix86_print_operand): Ditto.
28670 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28672 * lra-constraints.c (process_alt_operands): Fix recently added comment.
28674 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28676 PR rtl-optimization/56225
28677 * lra-constraints.c (process_alt_operands): Check that reload hard
28678 reg can hold value for strict_low_part.
28680 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28683 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
28684 dwarf2out_end_function.
28685 (in_first_function_p, maybe_at_text_label_p,
28686 first_loclabel_num_not_at_text_label): New variables.
28687 (dwarf2out_var_location): In the first function find out
28688 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
28689 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
28692 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
28694 PR rtl-optimization/56178
28695 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
28696 SUBREG of a register. Tidy up related block of code.
28697 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
28698 note if the source is a register or a SUBREG of a register.
28700 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28703 * config/rs6000/rs6000.md (ptrm): New mode attr.
28704 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
28705 call_value_indirect_aix<pttrsize>,
28706 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
28709 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28711 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
28712 if -bnortl. Convert to strcmp and strncmp.
28714 2013-02-07 Alan Modra <amodra@gmail.com>
28717 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
28718 addresses won't wrap when offsetting.
28719 (rs6000_secondary_reload): Provide secondary reloads needed for
28720 wrapping LO_SUM addresses.
28722 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
28724 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
28725 MACH, just __MACH__.
28727 2013-02-06 Richard Biener <rguenther@suse.de>
28729 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
28730 instead of calling fix_loop_structure.
28732 2013-02-06 Jakub Jelinek <jakub@redhat.com>
28734 PR middle-end/56217
28735 * omp-low.c (use_pointer_for_field): Return false if
28736 lower_send_shared_vars doesn't generate any copy-out code.
28738 2013-02-06 Tom de Vries <tom@codesourcery.com>
28740 PR rtl-optimization/56131
28741 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
28742 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
28743 of the label is NULL. Add comment.
28745 2013-02-05 Jakub Jelinek <jakub@redhat.com>
28747 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
28750 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
28751 (STATIC_LIBTSAN_LIBS): Likewise.
28752 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
28753 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
28754 is defined, don't add anything else beyond that.
28755 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
28756 (LINK_COMMAND_SPEC): Use them.
28758 PR tree-optimization/56205
28759 * tree-stdarg.c (check_all_va_list_escapes): Return true if
28760 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
28761 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
28763 2013-02-05 Richard Biener <rguenther@suse.de>
28765 PR tree-optimization/53342
28766 PR tree-optimization/53185
28767 * tree-vectorizer.h (vect_check_strided_load): Remove.
28768 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
28769 not disallow peeling for vectorized strided loads.
28770 (vect_check_strided_load): Make static and simplify.
28771 (vect_analyze_data_refs): Adjust.
28772 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
28773 correctly when vectorizing strided loads.
28775 2013-02-05 Richard Biener <rguenther@suse.de>
28777 * doc/install.texi: Refer to ISL, not PPL.
28779 2013-02-05 Jan Hubicka <jh@suse.cz>
28781 PR tree-optimization/55789
28782 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
28784 2013-02-05 Jan Hubicka <jh@suse.cz>
28786 PR tree-optimization/55789
28787 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
28788 the dead call anyway.
28790 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
28793 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
28795 2013-02-04 Alexander Potapenko <glider@google.com>
28796 Jack Howarth <howarth@bromo.med.uc.edu>
28797 Jakub Jelinek <jakub@redhat.com>
28800 * config/darwin.c (sort_ctor_records): Stabilized qsort
28801 on constructor priority by using original position.
28802 (finalize_ctors): New routine to sort constructors by
28803 priority before use in assemble_integer.
28804 (machopic_asm_out_constructor): Use finalize_ctors if needed.
28806 2013-02-04 Jakub Jelinek <jakub@redhat.com>
28809 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
28810 about visibility on artificial decls.
28811 * config/sol2.c (solaris_assemble_visibility): Likewise.
28813 2013-02-04 Kai Tietz <ktietz@redhat.com>
28816 * config/i386/i386.c (function_value_ms_64): Add additional valtype
28817 argument and improve checking of return-argument types for 16-byte
28819 (ix86_function_value_1): Add additional valtype argument on call
28820 of function_value_64.
28821 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
28822 handling infunction_value_64 function.
28824 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
28826 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
28828 2013-02-04 Richard Biener <rguenther@suse.de>
28830 PR tree-optimization/56188
28831 * tree-ssa-structalias.c (label_visit): Consider case with
28832 initially non-empty points-to set.
28833 (perform_var_substitution): Dump node mapping and clean up.
28835 2013-02-04 Richard Guenther <rguenther@suse.de>
28838 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
28839 node prevail as last resort.
28840 (lto_symtab_merge_decls): Remove guard on LTRANS here.
28841 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
28843 2013-02-04 Richard Biener <rguenther@suse.de>
28845 PR tree-optimization/56113
28846 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
28848 (equiv_class_lookup_or_add): ... this.
28849 (label_visit): Adjust and fix error in previous patch.
28850 (perform_var_substitution): Adjust.
28852 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28854 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28855 * config/sh/sh4-300.md: Likewise.
28856 * config/sh/sh4a.md: Likewise.
28857 * config/sh/constraints.md: Likewise.
28858 * config/sh/sh.md: Likewise.
28859 * config/sh/netbsd-elf.h: Likewise.
28860 * config/sh/predicates.md: Likewise.
28861 * config/sh/sh-protos.h: Likewise.
28862 * config/sh/ushmedia.h: Likewise.
28863 * config/sh/linux.h: Likewise.
28864 * config/sh/sh.c: Likewise.
28865 * config/sh/superh.h: Likewise.
28866 * config/sh/elf.h: Likewise.
28867 * config/sh/sh4.md: Likewise.
28868 * config/sh/sh.h: Likewise.
28870 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28872 * config/pa/constraints.md: Adjust unused letters. Change "T"
28873 constraint to match_test floating_point_store_memory_operand().
28874 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28875 (base14_operand): New.
28876 (floating_point_store_memory_operand): New.
28877 (integer_store_memory_operand): Revise to use base14_operand and
28878 reg_plus_base_memory_operand.
28879 (move_dest_operand): Allow symbolic_memory_operands.
28880 (symbolic_memory_operand): Check for LO_SOM.
28881 (symbolic_operand): Change default case to break.
28882 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28883 CONST_DOUBLE values to be reloaded by putting them into memory when
28884 the destination is a floating point register.
28885 (movdf): Remove code to handle CONST_DOUBLE.
28887 (reload_indf_r1): New.
28888 (reload_insf_r1): New.
28889 Consistently use "Q" and "T" constraints with integer and floating
28890 point move instructions, respectively.
28891 (movdi): Remove FAIL.
28892 Change predicate for source operand unamed DImode move from
28893 general_operand to move_src_operand.
28894 (umulsidi3): Change predicate for destination operand to
28896 Likewise for similar unamed patterns.
28897 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28898 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28899 (hppa_legitimize_address): Simplify mask calculation.
28900 (pa_emit_move_sequence): Revised handling of secondary reloads from
28901 REG+D addresses for floating point loads and stores. Directly handle
28902 loading CONST0_RTX (mode) to a floating point register.
28903 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28904 to floating point registers. Don't restrict secondary reloads to
28905 floating point registers to integer modes. Revise some comments and
28907 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28908 (pa_legitimate_address_p): New.
28909 (pa_legitimize_reload_address): New.
28910 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28911 (STRICT_REG_OK_FOR_BASE_P): New.
28912 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28913 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28915 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28916 Andrew Dixie <andrewd@gentrack.com>
28918 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28921 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28923 * expmed.c (extract_bit_field_1): Pass the full width of the
28924 structure to get_best_reg_extraction_insn.
28926 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28929 * configure.ac (use_cxa_atexit): Add AIX.
28930 * configure: Regenerate.
28932 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28934 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28937 * final.c (need_profile_function): New variable.
28938 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28939 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28940 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28941 notes, targetm.asm_out.function_prologue doesn't emit anything,
28942 HAVE_prologue and profiler should be emitted before prologue,
28943 set need_profile_function instead of emitting it.
28944 (final_scan_insn): If need_profile_function, emit
28945 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28946 NOTE_INSN_FUNCTION_BEG note.
28948 2013-02-01 Richard Henderson <rth@redhat.com>
28950 * config/rs6000/rs6000.md (smulditi3): New.
28953 * config/alpha/alpha.md (umulditi3): New.
28955 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28957 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28958 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28960 2013-02-01 Richard Biener <rguenther@suse.de>
28962 PR tree-optimization/56113
28963 * tree-ssa-structalias.c (label_visit): Reduce work for
28964 single-predecessor nodes.
28966 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28968 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28969 range isn't testing for zero.
28971 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28973 PR middle-end/56113
28974 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28976 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28977 Nick Clifton <nickc@redhat.com>
28979 * config/v850/constraints.md (Q): Define as a memory constraint.
28980 * config/v850/predicates.md (label_ref_operand): New predicate.
28981 (e3v5_shift_operand): New predicate.
28982 (ior_operator): New predicate.
28983 * config/v850/t-v850: Add e3v5 multilib.
28984 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28985 (v850_gen_movdi): Prototype.
28986 * config/v850/v850.c: Add support for e3v5 architecture.
28987 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28989 (construct_save_jarl): Add e3v5 long JARL support.
28990 (v850_adjust_insn_length): New function. Adjust length of call
28991 insns when using e3v5 instructions.
28992 (v850_gen_movdi): New function: Generate instructions to move a
28994 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
28995 (CPP_SPEC): Define __v850e3v5__ as appropriate.
28996 (TARGET_USE_FPU): Enable for e3v5.
28997 (CONST_OK_FOR_W): New macro.
28998 (ADJUST_INSN_LENGTH): Define.
28999 * config/v850/v850.md (UNSPEC_LOOP): Define.
29000 (attr cpu): Add v850e3v5.
29001 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
29002 (movdi): New pattern.
29003 (movdi_internal): New pattern.
29004 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
29005 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
29006 (cstoresf4): Likewise.
29007 (cstoredf4): Likewise.
29008 (insv): New pattern.
29009 (rotlso3_a): New pattern.
29010 (rotlsi3_b): New pattern
29011 (rotlsi3_v850e3v5): New pattern.
29012 (doloop_begin): New pattern.
29013 (fix_loop_counter): New pattern.
29014 (doloop_end): New pattern.
29015 (branch_normal): Add e3v5 long branch support.
29016 (branch_invert): Likewise.
29017 (branch_z_normal): Likewise.
29018 (branch_z_invert): Likewise.
29019 (branch_nz_normal): Likewise.
29020 (branch_nz_invert): Likewise.
29021 (call_internal_short): Add e3v5 register-indirect JARL support.
29022 (call_internal_long): Likewise.
29023 (call_value_internal_short): Likewise.
29024 (call_value_internal_long): Likewise.
29025 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
29026 (mloop): New option.
29027 * config.gcc: Add support for configuring v840e3v5 target.
29028 * doc/invoke.texi: Document new v850 specific command line options.
29030 2013-01-31 Paul Koning <ni1d@arrl.net>
29034 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
29035 children if parent is a class.
29036 (prune_unused_types_prune): Don't add DW_AT_declaration.
29038 2013-01-31 Richard Biener <rguenther@suse.de>
29040 PR tree-optimization/56157
29041 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
29042 match up operand with SLP child.
29044 2013-01-31 Jason Merrill <jason@redhat.com>
29047 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
29048 parameters the first time.
29049 (gen_scheduled_generic_parms_dies): Check completeness here.
29051 2013-01-31 Richard Biener <rguenther@suse.de>
29053 PR middle-end/53073
29054 * common.opt (faggressive-loop-optimizations): New flag,
29055 enabled by default.
29056 * doc/invoke.texi (faggressive-loop-optimizations): Document.
29057 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
29058 infer_loop_bounds_from_undefined by it.
29060 2013-01-31 Richard Biener <rguenther@suse.de>
29062 PR tree-optimization/56150
29063 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
29064 visit virtual operands.
29065 (find_uses_to_rename_bb): Likewise.
29067 2013-01-31 Richard Biener <rguenther@suse.de>
29069 PR tree-optimization/56150
29070 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
29071 mixed store non-store stmts.
29073 2013-01-30 Jakub Jelinek <jakub@redhat.com>
29076 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
29077 LIBASAN_EARLY_SPEC is defined.
29078 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
29079 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
29081 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
29084 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
29085 invalid args instead of ICEing on it.
29086 (ix86_valid_target_attribute_tree): Return error_mark_node if
29087 ix86_valid_target_attribute_inner_p failed.
29088 (ix86_valid_target_attribute_p): Return false only if
29089 ix86_valid_target_attribute_tree returned error_mark_node. Allow
29090 target("default") attribute.
29091 (sorted_attr_string): Change argument from const char * to tree,
29092 merge in all target attribute arguments rather than just one.
29093 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
29094 instead of free. Avoid using strcat.
29095 (ix86_mangle_function_version_assembler_name): Mangle
29096 target("default") as if no target attribute is present. Adjust
29097 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
29098 instead of xmalloc and XDELETEVEC instead of free.
29099 (ix86_function_versions): Don't return true if one of the decls
29100 doesn't have target attribute. If they don't and one of the decls
29101 is DECL_FUNCTION_VERSIONED, report an error. Adjust
29102 sorted_attr_string caller. Use XDELETEVEC instead of free.
29103 (ix86_supports_function_versions): Remove.
29104 (make_name): Fix up formatting.
29105 (make_dispatcher_decl): Remove resolver_name and its initialization.
29106 Avoid leaking memory.
29107 (is_function_default_version): Return true if there is
29108 target("default") attribute rather than no target attribute at all.
29109 (make_resolver_func): Avoid leaking memory.
29110 (ix86_generate_version_dispatcher_body): Likewise.
29111 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
29112 * target.def (supports_function_versions): Remove.
29113 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
29114 * doc/tm.texi: Regenerated.
29116 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
29118 PR rtl-optimization/56144
29119 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
29120 for values with side effects.
29122 2013-01-30 Richard Biener <rguenther@suse.de>
29124 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
29125 (sparseset_pop): Likewise.
29126 * cfganal.c (compute_idf): Likewise. Increase work-stack size
29127 to be able to use quick_push in the worker loop.
29129 2013-01-30 Marek Polacek <polacek@redhat.com>
29131 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
29133 2013-01-30 Richard Biener <rguenther@suse.de>
29136 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
29138 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
29140 PR tree-optimization/56064
29141 * fixed-value.c (fixed_from_double_int): New function.
29142 * fixed-value.h (fixed_from_double_int): New prototype.
29143 (const_fixed_from_double_int): New static inline function.
29144 * fold-const.c (native_interpret_fixed): New static function.
29145 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
29146 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
29147 (native_encode_fixed): New static function.
29148 (native_encode_expr) <FIXED_CST>: Use it.
29149 (native_interpret_int): Move double_int worker code to...
29150 * double-int.c (double_int::from_buffer): ...this new static method.
29151 * double-int.h (double_int::from_buffer): Prototype it.
29153 2013-01-30 Richard Biener <rguenther@suse.de>
29155 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
29156 New pointer-map and obstack.
29157 (init_alias_vars): Allocate pointer-map and obstack.
29158 (delete_points_to_sets): Free them.
29159 (find_what_var_points_to): Cache result.
29160 (find_what_p_points_to): Adjust for changed interface of
29161 find_what_var_points_to.
29162 (compute_points_to_sets): Likewise.
29163 (ipa_pta_execute): Likewise.
29165 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
29167 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
29168 * configure: Regenerate.
29169 * config.in: Regenerate.
29170 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
29171 #nobits/#progbits if supported.
29173 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
29176 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
29177 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
29178 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
29180 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29182 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
29183 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
29185 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29187 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
29188 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
29190 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29192 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
29194 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
29195 * config/arm/cortex-a7.md: New bypasses using
29196 arm_mac_accumulator_is_result.
29198 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29200 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
29201 (cortex_a7_neon_mla): Likewise.
29202 (cortex_a7_fpfmad): New reservation.
29203 (cortex_a7_fpmacs): Use ffmas and update required units.
29204 (cortex_a7_fpmuld): Update required units and latency.
29205 (cortex_a7_fpmacd): Likewise.
29206 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
29207 (cortex_a7_neon). Likewise.
29208 (bypass) Update participating units.
29210 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29212 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
29213 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
29215 * config/arm/vfp11.md (vfp_farith): Use ffmas.
29216 (vfp_fmul): Use ffmad.
29217 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
29218 (cortex_r4_fmacd): Use ffmad.
29219 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
29220 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
29221 (cortex_a9_fmacd): Use ffmad.
29222 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
29223 (cortex_a8_vfp_macd): Use ffmad.
29224 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
29225 (cortex_a5_fpmacd): Use ffmad.
29226 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
29227 (cortex_a15_vfp_macd): Use ffmad.
29228 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29230 2013-01-29 Jason Merrill <jason@redhat.com>
29233 * varasm.c (default_assemble_visibility): Don't warn about
29234 visibility on artificial decls.
29236 2013-01-29 Richard Biener <rguenther@suse.de>
29238 PR tree-optimization/56113
29239 * tree-ssa-structalias.c (equiv_class_lookup): Also return
29241 (label_visit): Free duplicate bitmaps and record the leader instead.
29242 (perform_var_substitution): Adjust.
29244 2013-01-29 Richard Biener <rguenther@suse.de>
29246 PR tree-optimization/55270
29247 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
29248 the CFG, schedule loops for fixup.
29250 2013-01-29 Nick Clifton <nickc@redhat.com>
29252 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
29255 2013-01-28 Leif Ekblad <leif@rdos.net>
29257 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
29258 * config/i386/i386.h (TARGET_RDOS): New macro.
29259 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
29260 * config/i386/i386.c (ix86_option_override_internal): For 64bit
29261 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
29262 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
29263 DEFAULT_LARGE_SECTION_THRESHOLD.
29264 * config/i386/i386.md (R14_REG, R15_REG): New constants.
29265 * config/i386/rdos.h: New file.
29266 * config/i386/rdos64.h: New file.
29268 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
29271 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
29274 2013-01-28 Jakub Jelinek <jakub@redhat.com>
29276 PR rtl-optimization/56117
29277 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
29278 call cselib_lookup_from_insn on the MEM before calling
29279 add_insn_mem_dependence.
29281 2013-01-28 Richard Biener <rguenther@suse.de>
29283 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
29284 to a stmt that didn't have one.
29285 (copy_phis_for_bb): Likewise for PHI arguments.
29286 (copy_debug_stmt): Likewise for debug stmts.
29288 2013-01-28 Richard Biener <rguenther@suse.de>
29290 PR tree-optimization/56034
29291 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
29292 (partition_builtin_p): Adjust.
29293 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
29294 it is the last partition.
29295 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
29296 up the vertex for the definition.
29297 (classify_partition): Classify whether a partition is a
29298 PKIND_REDUCTION, thus has uses outside of the loop.
29299 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
29300 Merge all PKIND_REDUCTION partitions into the last partition.
29301 (tree_loop_distribution): Seed partitions from reductions as well.
29303 2013-01-28 Jakub Jelinek <jakub@redhat.com>
29305 PR tree-optimization/56125
29306 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
29307 pow(x,c) into sqrt(x) * powi(x, n/2) or
29308 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
29309 optimizing for size.
29310 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
29311 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
29314 PR tree-optimization/56094
29315 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
29316 to UNKNOWN_LOCATION while gimplifying expr.
29318 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
29321 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
29322 operand 0 in movabs insn template for -masm=intel asm alternative.
29323 (*movabs<mode>_2): Ditto for operand 1.
29325 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
29328 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
29331 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29333 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
29336 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29338 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
29339 Undef to avoid warning.
29341 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
29343 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
29344 * configure: Regenerate.
29346 2013-01-25 Jakub Jelinek <jakub@redhat.com>
29348 PR tree-optimization/56098
29349 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
29350 for stmts with volatile ops.
29351 (cond_store_replacement): Don't optimize if assign has volatile ops.
29352 (cond_if_else_store_replacement_1): Don't optimize if either
29353 then_assign or else_assign have volatile ops.
29354 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
29357 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29359 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
29361 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29363 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
29364 missing ':' in asm example.
29366 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
29368 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
29369 entries into lane and laneq entries.
29370 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
29371 Remove AdvSIMD scalar modes.
29372 (aarch64_sq<r>dmulh_laneq<mode>): New.
29373 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
29375 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
29376 builtin implementations to relfect changes in RTL in aarch64-simd.md.
29377 * config/aarch64/iterators.md (VCOND): New.
29380 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29383 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
29384 Add NULL LIBNAME argument to existing definitions.
29385 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
29386 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
29387 * config/avr/avr.c (DEF_BUILTIN): Same.
29388 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
29389 (avr_expand_builtin): Expand to a vanilla call if a libgcc
29390 implementation is available (DECL_ASSEMBLER_NAME is set).
29391 (avr_fold_absfx): New static function.
29392 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
29393 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
29394 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
29395 AVR_BUILTIN_ABSLLK.
29396 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
29397 (abshk, absk, abslk, absllk): Provide as static inline functions.
29399 2013-01-25 Marek Polacek <polacek@redhat.com>
29401 PR tree-optimization/56035
29402 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
29404 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
29406 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
29407 (*movtf_internal_rex64): Add (!o,C) alternative
29408 (*movxf_internal_rex64): Ditto.
29409 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
29411 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
29413 * doc/invoke.texi: fix typo.
29414 * doc/objc.texi: fix typo.
29416 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
29418 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
29419 for the first two alternatives.
29421 2013-01-24 Diego Novillo <dnovillo@google.com>
29423 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
29424 (ggc-zone.o): Remove.
29425 * configure.ac: Remove option --with-gc.
29426 * configure: Re-generate.
29427 * doc/install.texi: Remove documentation for --with-gc.
29428 * gengtype.c (write_enum_defn): Remove. Update all users.
29429 (write_Types_process_field): Remove generation of gt_e_* argument.
29430 (output_type_enum): Remove. Update all users.
29431 (write_enum_defn): Remove. Update all users.
29432 (enum alloc_zone): Remove. Update all users.
29433 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
29434 * ggc-common.c (ggc_splay_alloc): Remove first argument.
29435 Update all callers.
29436 (struct ptr_data): Remove field TYPE. Update all users.
29437 (gt_pch_note_object): Remove argument TYPE. Update all users.
29438 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
29440 * ggc-none.c (ggc_alloc_typed_stat): Remove.
29441 (struct alloc_zone): Remove.
29442 (ggc_internal_alloc_zone_stat): Remove.
29443 (ggc_internal_cleared_alloc_zone_stat): Remove.
29444 * ggc-page.c (ggc_alloc_typed_stat): Remove.
29445 (ggc_pch_count_object): Remove last argument. Update all users.
29446 (ggc_pch_alloc_object): Remove last argument. Update all users.
29447 (struct alloc_zone): Remove.
29448 * ggc-zone.c: Remove.
29449 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
29450 (struct alloc_zone): Remove.
29451 (ggc_alloc_typed_stat): Remove.
29452 (ggc_alloc_typed): Remove.
29453 (ggc_splay_alloc): Remove first argument.
29454 (rtl_zone): Remove. Update all users.
29455 (tree_zone): Remove. Update all users.
29456 (tree_id_zone): Remove. Update all users.
29457 (ggc_internal_zone_alloc_stat): Remove. Update all users.
29458 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
29459 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
29460 * tree-ssanames.c: Remove references to zone allocator in comments.
29462 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
29464 * config/avr/avr.c (avr_out_fract): Make register numbers that
29465 might be outside of source operand signed.
29467 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
29469 * config/i386/constraints.md (Yf): New constraint.
29470 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
29471 of f constraint to conditionaly disable x87 register preferences.
29472 (*movdf_internal): Ditto.
29473 (*movsf_internal): Ditto.
29475 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
29477 PR inline-asm/55934
29478 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
29479 that have operands with impossible constraints.
29480 Add a FIXME for a speed-up opportunity.
29481 * lra-constraints.c (process_alt_operands): Verify that a class
29482 selected from constraints on asms is valid for the operand mode.
29483 (curr_insn_transform): Remove incorrect comment.
29485 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
29487 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
29488 TOC operand is a valid symbol ref in the constant pool.
29490 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29492 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
29494 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
29497 * config/avr/stdfix.h: New file.
29498 * t-avr (stdfix-gcc.h): New rule to build it.
29499 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
29501 2013-01-23 Kostya Serebryany <kcc@google.com>
29503 * config/darwin.h: remove dependency on
29504 CoreFoundation (asan on Mac OS).
29506 2013-01-23 Jakub Jelinek <jakub@redhat.com>
29509 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
29510 instead of cmpdi_operand for first comparison operand.
29511 Don't assert that comparison operands aren't both constants.
29513 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29515 * doc/install.texi (Downloading the Source): Update references to
29516 downloading separate components.
29518 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29520 * doc/extend.texi (__int128): Improve grammar.
29522 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
29525 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
29526 alternative to (o,r).
29527 (*movdi_internal_rex64): Remove (!o,n) alternative.
29528 (DImode immediate->memory splitter): Remove.
29529 (DImode immediate->memory peephole2): Remove.
29530 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
29531 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
29532 alternative to (!o,*r).
29533 (*movtf_internal_sse): New pattern.
29534 (*movxf_internal_rex64): New pattern.
29535 (*movxf_internal): Disable for TARGET_64BIT.
29536 (*movdf_internal_rex64): Remove (!o,F) alternative.
29538 2013-01-22 Jakub Jelinek <jakub@redhat.com>
29540 PR middle-end/56074
29541 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
29542 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
29543 * tree-vect-loop-manip.c (find_loop_location): Also ignore
29544 stmt locations where LOCATION_LOCUS of the stmt location is
29545 UNKNOWN_LOCATION or BUILTINS_LOCATION.
29548 * config/i386/i386.md (UNSPEC_STOS): New.
29549 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
29550 *strsetqi_1): Add UNSPEC_STOS.
29552 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
29555 * doc/invoke.texi: Remove left over -Wsynth example.
29557 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29559 PR tree-optimization/56051
29560 * fold-const.c (fold_binary_loc): Don't fold
29561 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
29562 a narrowing conversion, or widening conversion from signed
29565 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29567 PR rtl-optimization/56023
29568 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
29569 dependent on debug instruction.
29571 2013-01-21 Martin Jambor <mjambor@suse.cz>
29573 PR middle-end/56022
29574 * function.c (allocate_struct_function): Call
29575 invoke_set_current_function_hook earlier.
29577 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29579 * reload1.c (init_reload): Only initialize reload_obstack
29580 during the first call.
29582 2013-01-21 Marek Polacek <polacek@redhat.com>
29584 * cfgloop.c (verify_loop_structure): Fix up grammar.
29586 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
29588 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
29589 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
29591 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29594 * config/arm/marvell-pj4.md: Update copyright year.
29595 Fix up use of alu to alu_reg and simple_alu_imm.
29597 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29599 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
29601 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
29604 * lra-constraints.c (curr_insn_transform): Don't reuse original
29605 insn for secondary memory move when memory mode should be different.
29607 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29609 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
29610 atomic_storedi_1): New patterns.
29612 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29614 btver2 pipeline descriptions.
29615 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
29617 * config/i386/i386.md (btver2_decode): New type attributes.
29618 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
29620 * config/i386/btver2.md: New file describing btver2 pipelines.
29622 2013-01-19 Andrew Pinski <apinski@cavium.com>
29624 PR tree-optimization/52631
29625 * tree-ssa-sccvn (visit_use): Before looking up the original
29626 statement, try looking up the simplified expression.
29628 2013-01-19 Anthony Green <green@moxielogic.com>
29630 * config/moxie/moxie.c (moxie_expand_prologue): Set
29631 current_function_static_stack_size.
29633 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29635 PR tree-optimization/56029
29636 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
29637 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
29639 2013-01-18 Sharad Singhai <singhai@google.com>
29641 PR tree-optimization/55995
29642 * dumpfile.c (dump_loc): Print location only if available.
29643 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
29645 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
29648 * lra-constraints.c (curr_insn_transform): Reuse original insn for
29649 secondary memory move.
29650 (inherit_reload_reg): Use rclass instead of cl for
29651 check_secondary_memory_needed_p.
29653 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29655 PR middle-end/56015
29656 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
29657 the case where writing real complex part of target modifies op1.
29659 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
29661 * config/aarch64/aarch64-simd.md
29662 (aarch64_vcond_internal<mode>): Handle unordered cases.
29663 * config/aarch64/iterators.md (v_cmp_result): New.
29665 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
29666 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29668 * config/arm/marvell-pj4.md: New file.
29669 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
29670 * config/arm/arm.md (generic_sched): Add marvell_pj4.
29671 (generic_vfp): Likewise.
29672 * config/arm/arm-cores.def: Add marvell-pj4.
29673 * config/arm/arm-tune.md: Regenerate.
29674 * config/arm/arm-tables.opt: Regenerate.
29675 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
29676 * doc/invoke.texi: Document marvell-pj4.
29678 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
29680 * config/aarch64/arm_neon.h: Map scalar types to standard types.
29682 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29687 * var-tracking.c (negative_power_of_two_p): New.
29688 (global_get_addr_cache, local_get_addr_cache): New.
29689 (get_addr_from_global_cache, get_addr_from_local_cache): New.
29690 (vt_canonicalize_addr): Rewrite using the above. Adjust the
29692 (vt_stack_offset_p): Remove.
29693 (vt_canon_true_dep): Always canonicalize loc's address.
29694 (clobber_overlapping_mems): Make sure we have a MEM.
29695 (local_get_addr_clear_given_value): New.
29696 (val_reset): Clear local cached entries.
29697 (compute_bb_dataflow): Create and release the local cache.
29698 Disable duplicate MEMs clobbering.
29699 (emit_notes_in_bb): Clobber MEMs likewise.
29700 (vt_emit_notes): Create and release the local cache.
29701 (vt_initialize, vt_finalize): Create and release the global
29702 cache, respectively.
29703 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
29705 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29707 PR libmudflap/53359
29708 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
29709 not found in the symtab.
29711 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29714 PR rtl-optimization/55547
29715 PR rtl-optimization/53827
29718 * alias.c (offset_overlap_p): New, factored out of...
29719 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
29720 the conservative special case for symbolic constants. Don't
29721 adjust zero sizes on alignment.
29723 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
29725 PR rtl-optimization/52573
29726 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
29727 REG_UNUSED for the same register.
29729 2013-01-17 Richard Biener <rguenther@suse.de>
29730 Marek Polacek <polacek@redhat.com>
29732 PR rtl-optimization/55833
29733 * loop-unswitch.c (unswitch_loops): Move loop verification...
29734 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
29735 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
29736 Set it to true when we're removing a loop from hierarchy tree in
29737 an irreducible region.
29738 (fix_bb_placements): Adjust caller.
29739 (fix_loop_placements): Likewise.
29741 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
29743 * config/avr/builtins.def (DEF_BUILTIN): Factor out
29744 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
29745 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
29746 Remove ID. Adjust comments.
29747 * config/avr/avr-c.c (avr_builtin_name): Remove.
29748 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
29749 * config/avr/avr.c (avr_tolower): New static function.
29750 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
29751 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
29752 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
29755 2013-01-17 Jan Hubicka <jh@suse.cz>
29757 PR tree-optimization/55273
29758 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
29760 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
29763 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
29764 store through atomic_store<mode>_1.
29765 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
29767 2013-01-17 Martin Jambor <mjambor@suse.cz>
29769 PR tree-optimizations/55264
29770 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
29771 for virtual methods.
29772 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
29773 virtual methods before inlining is over.
29774 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
29776 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
29779 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
29781 PR rtl-optimization/56005
29782 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
29783 pending reads for prefetch.
29785 2013-01-16 Ian Bolton <ian.bolton@arm.com>
29787 * config/aarch64/aarch64.md
29788 (*cstoresi_neg_uxtw): New pattern.
29789 (*cmovsi_insn_uxtw): New pattern.
29790 (*<optab>si3_uxtw): New pattern.
29791 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
29792 (*<optab>si3_insn_uxtw): New pattern.
29793 (*bswapsi2_uxtw): New pattern.
29795 2013-01-16 Richard Biener <rguenther@suse.de>
29797 * tree-inline.c (tree_function_versioning): Remove set but
29798 never used variable.
29800 2013-01-16 Richard Biener <rguenther@suse.de>
29802 PR tree-optimization/55964
29803 * tree-flow.h (rename_variables_in_loop): Remove.
29804 (rename_variables_in_bb): Likewise.
29805 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
29806 (copy_loop_before): Adjust and delete update-ssa status.
29807 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
29808 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
29809 (rename_variables_in_loop): Remove.
29810 (slpeel_update_phis_for_duplicate_loop): Likewise.
29811 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
29812 use available cfg machinery instead of duplicating it.
29813 Update PHI nodes and perform poor-mans SSA update here.
29814 (slpeel_tree_peel_loop_to_edge): Adjust.
29816 2013-01-16 Richard Biener <rguenther@suse.de>
29818 PR tree-optimization/54767
29819 PR tree-optimization/53465
29820 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
29821 (vrp_visit_phi_node): For PHI arguments coming via backedges
29822 drop all symbolical range information.
29823 (execute_vrp): Compute backedges.
29825 2013-01-16 Richard Biener <rguenther@suse.de>
29827 * doc/install.texi: Update CLooG and ISL requirements to
29830 2013-01-16 Christian Bruel <christian.bruel@st.com>
29833 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
29834 (broken_move): Handle UNSPECV_SP_SWITCH_B.
29835 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
29837 2013-01-16 DJ Delorie <dj@redhat.com>
29839 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
29840 (UNSPECV_SP_SWITCH_E): New.
29841 (sp_switch_1): Change to an unspec.
29842 (sp_switch_2): Change to an unspec. Don't use post-inc when we
29845 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
29847 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29848 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29849 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29850 (expand_mem_thread_fence): Ditto.
29851 (expand_mem_signal_fence): Ditto.
29852 (expand_atomic_load): Ditto.
29853 (expand_atomic_store): Ditto.
29855 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29857 PR rtl-optimization/55547
29858 PR rtl-optimization/53827
29861 * alias.c (memrefs_conflict_p): Set sizes to negative after
29864 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29867 * function.c (thread_prologue_and_epilogue_insns): Always
29868 add crtl->drap_reg to set_up_by_prologue.set, even if
29869 stack_realign_drap is false.
29871 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29873 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29874 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29875 *call): Fix indention.
29877 2013-01-15 Tom de Vries <tom@codesourcery.com>
29880 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29883 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29885 PR rtl-optimization/55153
29886 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29888 2013-01-15 Martin Jambor <mjambor@suse.cz>
29890 PR tree-optimization/55920
29891 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29892 accesses as grp_to_be_debug_replaced.
29894 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29896 PR tree-optimization/55920
29897 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29898 there is non-useless type conversion needed from debug rhs to lhs,
29899 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29901 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29902 Mikael Pettersson <mikpe@it.uu.se>
29905 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29907 (ASM_OUTPUT_CASE_LABEL): Remove.
29908 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29909 * final.c (shorten_branches): Update alignment of labels before
29910 jump tables if CASE_VECTOR_SHORTEN_MODE.
29912 2013-01-15 Richard Biener <rguenther@suse.de>
29915 * system.h: Do not include gmp.h for building host tools.
29917 2013-01-15 Richard Biener <rguenther@suse.de>
29919 PR middle-end/55882
29920 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29921 account for bitpos when computing alignment.
29923 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29925 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29926 (ix86_target_macros_internal): Likewise.
29928 * config/i386/i386.c (m_CORE2I7): Removed.
29929 (m_CORE_HASWELL): New macro.
29930 (m_CORE_ALL): Likewise.
29931 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29932 (initial_ix86_arch_features): Likewise.
29933 (processor_target_table): Initializations for Core avx2.
29934 (cpu_names): New names "core-avx2".
29935 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29936 PROCESSOR_CORE_HASWELL.
29937 (ix86_issue_rate): New case.
29938 (ia32_multipass_dfa_lookahead): Likewise.
29939 (ix86_sched_init_global): Likewise.
29941 * config/i386/i386.h (TARGET_HASWELL): New macro.
29942 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29943 (processor_type): New PROCESSOR_HASWELL.
29945 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29947 PR tree-optimization/55955
29948 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29949 *SHIFT_EXPR and *ROTATE_EXPR codes.
29951 PR tree-optimization/48766
29952 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29953 -ftrapv disable -fwrapv.
29955 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29958 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29959 etc. to 1 and not to __flash.
29960 Use LL suffix for __INT24_MAX__ with -mint8.
29961 Use ULL suffix for __UINT24_MAX__ with -mint8.
29963 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29965 * config/avr/avr-arch.h
29966 (struct base_arch_s): Use typedef avr_arch_t instead.
29967 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29968 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29969 * config/avr/avr.c: Same.
29970 * config/avr/avr-devices.c: Same.
29971 * config/avr/driver-avr.c: Same.
29972 * config/avr/gen-avr-mmcu-texi.c: Same.
29973 * config/avr/avr-mcus.def: Adjust comment.
29975 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29977 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29978 * config/aarch64/iterators.md (VALLDI): New.
29980 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29981 Andi Kleen <ak@linux.intel.com>
29984 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29985 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29988 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29990 * config/avr/avr-stdint.h: Remove trailing blanks.
29991 * config/avr/avr-log.h: Same.
29992 * config/avr/avr-arch.h: Same.
29993 * config/avr/avr-devices.c: Same.
29994 * config/avr/avr-dimode.md: Same.
29995 * config/avr/predicates.md: Same.
29996 * config/avr/avr-c.c: Same. And fix typo.
29998 * config/avr/avr-protos.h: Same. And:
29999 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
30000 (init_cumulative_args): Rename to avr_init_cumulative_args.
30001 (expand_prologue): Rename to avr_expand_prologue.
30002 (expand_epilogue): Rename to avr_expand_epilogue.
30003 (adjust_insn_length): Rename to avr_adjust_insn_length.
30004 (notice_update_cc): Rename to avr_notice_update_cc.
30005 (final_prescan_insn): Rename to avr_final_prescan_insn.
30006 * config/avr/avr.c: Same.
30007 * config/avr/avr.h: Same.
30008 * config/avr/avr.md: Remove trailing blanks.
30009 (prologue): Use avr_expand_prologue.
30010 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
30012 2013-01-14 Richard Biener <rguenther@suse.de>
30014 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
30015 verify_location, collect_subblocks): New functions.
30016 (verify_gimple_in_cfg): Verify that locations only reference
30017 BLOCKs in the functions BLOCK tree.
30019 2013-01-14 Richard Biener <rguenther@suse.de>
30021 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
30023 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
30025 (insert_out_of_ssa_copy_on_edge): Likewise.
30026 (rewrite_close_phi_out_of_ssa): Likewise.
30027 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
30029 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
30030 propagated constants.
30031 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
30034 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
30036 * config/avr/avr-modes.def: Add GPL copyright notice.
30038 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
30040 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
30041 MEMMODEL_MASK to determine memory model.
30042 (atomic_store<mode>): Ditto from operands[2].
30043 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
30045 2013-01-13 Jakub Jelinek <jakub@redhat.com>
30048 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
30049 (fold_gimple_assign): Don't call unshare_expr here.
30050 (fold_ctor_reference): Call unshare_expr.
30052 2013-01-13 Terry Guo <terry.guo@arm.com>
30054 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
30055 * doc/fragments.texi: Document MULTILIB_REUSE.
30056 * gcc.c (multilib_reuse): New internal spec.
30057 (set_multilib_dir): Also search multilib from multilib_reuse.
30058 * genmultilib (tmpmultilib3): Refactor code.
30059 (tmpmultilib4): Ditto.
30060 (multilib_reuse): New multilib argument.
30062 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
30064 * Makefile.in: Update copyright.
30066 2013-01-12 Tom de Vries <tom@codesourcery.com>
30068 PR middle-end/55890
30069 * calls.c (expand_call): Check if arg_nr is valid.
30071 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
30073 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
30074 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
30075 documentation. Add missing '__' in front of
30076 __builtin_ia32_packssdw256.
30078 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
30081 * config/s390/s390.c (s390_preferred_reload_class): Do not return
30082 NO_REGS for larl operands.
30083 (s390_reload_larl_operand): Use s390_load_address instead of
30086 2013-01-11 Richard Biener <rguenther@suse.de>
30088 * tree-cfg.c (verify_node_sharing_1): Split out from ...
30089 (verify_node_sharing): ... here.
30090 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
30092 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
30094 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
30095 Substitute TREECHECKING.
30096 * configure: Regenerate.
30097 * Makefile.in (TREECHECKING): New.
30099 2013-01-11 Richard Guenther <rguenther@suse.de>
30101 PR tree-optimization/44061
30102 * tree-vrp.c (extract_range_basic): Compute zero as
30103 value-range for __builtin_constant_p of function parameters.
30105 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
30107 Update copyright years.
30109 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
30111 PR rtl-optimization/55672
30112 * lra-eliminations.c (mark_not_eliminable): Permit addition with
30113 const to be eliminable.
30115 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
30117 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
30118 * configure: Regenerate.
30120 2013-01-10 Richard Biener <rguenther@suse.de>
30122 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
30124 2013-01-10 Richard Biener <rguenther@suse.de>
30127 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
30128 locations for virtual PHI arguments.
30129 (rewrite_update_phi_arguments): Likewise.
30131 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
30133 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
30136 2013-01-10 Jakub Jelinek <jakub@redhat.com>
30138 PR tree-optimization/55921
30139 * tree-complex.c (expand_complex_asm): New function.
30140 (expand_complex_operations_1): Call it for GIMPLE_ASM.
30142 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
30145 * config/s390/s390.c (s390_symref_operand_p)
30146 (s390_loadrelative_operand_p): Merge the two functions.
30147 (s390_check_qrst_address, print_operand_address): Add parameters
30148 to s390_loadrelative_operand_p invokation.
30149 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
30150 (s390_reload_larl_operand, s390_secondary_reload): Use
30151 s390_loadrelative_operand_p instead of s390_symref_operand_p.
30152 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
30154 2013-01-09 Mike Stump <mikestump@comcast.net>
30156 * dse.c (record_store): Remove unnecessary assert.
30158 2013-01-09 Jan Hubicka <jh@suse.cz>
30160 PR tree-optimization/55569
30161 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
30162 * cfgloop.h (scale_loop_profile): Likewise.
30164 2013-01-09 Jan Hubicka <jh@suse.cz>
30167 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
30169 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
30171 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
30173 PR middle-end/55114
30174 * expr.h (maybe_emit_group_store): Declare.
30175 * expr.c (maybe_emit_group_store): New function.
30176 * builtins.c (expand_builtin_int_roundingfn): Call it.
30177 (expand_builtin_int_roundingfn_2): Likewise.
30179 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
30181 PR rtl-optimization/55829
30182 * lra-constraints.c (match_reload): Add code for absent output.
30183 (curr_insn_transform): Add code for reloads of matched inputs
30186 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
30188 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
30189 attribute of movddup insn to DF.
30190 (*vec_interleave_lowv2df): Ditto.
30191 (vec_dupv2df): Ditto.
30193 2013-01-09 Jan Hubicka <jh@suse.cz>
30195 PR tree-optimiation/55875
30196 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
30197 EVERY_ITERATION parameter.
30198 (number_of_iterations_exit): Check if exit is executed every iteration.
30199 (idx_infer_loop_bounds): Similarly here.
30200 (n_of_executions_at_most): Simplify
30201 to only test for cases where statement is dominated by the
30202 particular bound; handle correctly the "postdominance" test.
30203 (scev_probably_wraps_p): Use max loop iterations info
30204 as a global bound first.
30206 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
30207 Nick Clifton <nickc@redhat.com>
30209 * config/v850/v850.md (cbranchsf4): New pattern.
30210 (cstoresf4): New pattern.
30211 (cbranchdf4): New pattern.
30212 (cstoredf4): New pattern.
30213 (movsicc): Disallow floating point comparisons.
30214 (cmpsf_le_insn): Fix order of operators.
30215 (cmpsf_lt_insn): Likewise.
30216 (cmpsf_eq_insn): Likewise.
30217 (cmpdf_le_insn): Likewise.
30218 (cmpdf_lt_insn): Likewise.
30219 (cmpdf_eq_insn): Likewise.
30220 (cmpsf_ge_insn): Use LE comparison.
30221 (cmpdf_ge_insn): Likewise.
30222 (cmpsf_gt_insn): Use LT comparison.
30223 (cmpdf_gt_insn): Likewise.
30224 (cmpsf_ne_insn): Delete pattern.
30225 (cmpdf_ne_insn): Delete pattern.
30226 * config/v850/v850.c (v850_gen_float_compare): Use
30227 gen_cmpdf_eq_insn for NE comparison.
30228 (v850_float_z_comparison_operator)
30229 (v850_float_nz_comparison_operator): Move from here ...
30230 * config/v850/predicates.md: ... to here. Move GT and GE
30231 comparisons into v850_float_z_comparison_operator.
30232 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
30234 (v850_float_nz_comparison_operator): Likewise.
30236 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30238 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
30239 with calls to gen_insvsi/gen_insvdi.
30241 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
30243 * config/i386/i386.c (initial_ix86_tune_features): Set up
30244 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
30246 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
30247 Jakub Jelinek <jakub@redhat.com>
30249 PR tree-optimization/48189
30250 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
30251 If nitercst is 0, don't predict the exit edge.
30253 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
30255 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
30256 in asm_fprintf with reg_names.
30257 (aarch64_print_operand_address): Likewise.
30258 (aarch64_return_addr): Likewise.
30259 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
30261 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30263 * config/pa/pa.h (VAL_U6_BITS_P): Define.
30264 (INT_U6_BITS): Likewise.
30265 * config/pa/predicates.md (uint6_operand): New predicate.
30266 (shift5_operand, shift6_operand): Likewise.
30267 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
30269 (lshrdi3): Use shift6_operand.
30270 (shrpsi4, shrpdi4): New insn patterns.
30271 (extzv): Delete expander.
30272 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
30273 predicates in unamed zero extract patterns. Tighten common constraint.
30274 (extv): Delete expander.
30275 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
30276 predicates in unamed sign extract patterns. Tighten common constraint.
30277 (insv): Delete expander.
30278 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
30279 predicates in unamed insert patterns. Tighten common constraint.
30280 Change uint32_operand predicate to uint6_operand predicate in unamed
30281 DImode pattern to insert constant values of type 1...1xxxx.
30283 2013-01-04 Jan Hubicka <jh@suse.cz>
30285 PR tree-optimization/55823
30286 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
30289 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30290 Uros Bizjak <ubizjak@gmail.com>
30292 PR rtl-optimization/55845
30293 * df-problems.c (can_move_insns_across): Stop scanning at
30294 volatile_insn_p source instruction or give up if
30295 across_from .. across_to range contains any volatile_insn_p
30298 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
30300 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
30301 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
30303 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
30304 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
30306 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30309 * asan.c (asan_clear_shadow): New function.
30310 (asan_emit_stack_protection): Use it.
30312 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
30314 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
30315 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
30316 with tab instead of space.
30318 2013-01-08 Nick Clifton <nickc@redhat.com>
30320 * config/rl78/rl78.c (rl78_expand_prologue): Always select
30321 register bank 0 at the start of an interrupt handler.
30322 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
30325 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
30327 * config/aarch64/aarch64-simd.md
30328 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
30329 (aarch64_simd_bsl): Likewise.
30330 (aarch64_vcond_internal<mode>): Likewise.
30331 (vcond<mode><mode>): Likewise.
30332 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
30333 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
30335 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
30337 * config/aarch64/aarch64-builtins.c
30338 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
30340 2013-01-08 Martin Jambor <mjambor@suse.cz>
30343 * tree-sra.c (analyze_access_subtree): Return true also after
30344 potentially creating a debug-only replacement.
30346 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30348 PR middle-end/55890
30349 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
30351 PR tree-optimization/54120
30352 * tree-vrp.c (range_fits_type_p): Don't allow
30353 src_precision < precision from signed vr to unsigned_p
30354 if vr->min or vr->max is negative.
30355 (simplify_float_conversion_using_ranges): Test can_float_p
30356 against CODE_FOR_nothing.
30358 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30359 Richard Biener <rguenther@suse.de>
30361 PR middle-end/55851
30362 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
30363 types instead of just INTEGER_TYPE types.
30365 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
30367 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
30370 2013-01-07 Steve Ellcey <sellcey@mips.com>
30373 * config/mips/mips.opt: Change mad to mmad to match documentation.
30375 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30378 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
30379 .progmemx.data now.
30381 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30384 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
30385 (avr_addrspace_t): Add .section_name field.
30386 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
30388 (avr_addrspace): Same. Initialize .section_name. Remove last
30389 NULL entry. Put __memx into .progmemx.data.
30390 (progmem_section_prefix): Remove.
30391 (avr_asm_init_sections): No need to initialize progmem_section.
30392 (avr_asm_named_section): Use avr_addrspace[].section_name to get
30393 section name prefix.
30394 (avr_asm_select_section): Ditto. And use get_unnamed_section to
30395 retrieve the progmem section.
30396 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
30397 boundary to run over avr_addrspace[].
30398 (avr_register_target_pragmas): Ditto.
30400 2013-01-06 Jakub Jelinek <jakub@redhat.com>
30402 * varasm.c (output_constant_def_contents): For asan_protect_global
30403 protected strings, adjust DECL_ALIGN if needed, before testing for
30405 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
30406 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
30408 (output_object_block): For asan protected decls, emit asan padding
30409 after their contents.
30410 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
30411 (asan_finish_file): Test it here instead.
30413 2013-01-07 Nick Clifton <nickc@redhat.com>
30414 Matthias Klose <doko@debian.org>
30415 Doug Kwan <dougkwan@google.com>
30416 H.J. Lu <hongjiu.lu@intel.com>
30419 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
30421 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
30423 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
30425 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
30427 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
30429 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30432 * doc/install.texi (Cross-Compiler-Specific Options): Document
30435 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
30437 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
30438 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
30439 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
30440 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
30441 vqmovun_high_s64): Fix source operand number and update copyright.
30443 2013-01-07 Richard Biener <rguenther@suse.de>
30445 PR middle-end/55890
30446 * gimple.h (gimple_call_builtin_p): New overload.
30447 * gimple.c (validate_call): New function.
30448 (gimple_call_builtin_p): Likewise.
30449 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
30450 Use gimple_call_builtin_p.
30451 (find_func_clobbers): Likewise.
30452 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
30453 (strlen_optimize_stmt): Likewise.
30455 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30457 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
30458 (vld1q_dup_*): Likewise.
30459 (vld1_*): Likewise.
30460 (vld1q_*): Likewise.
30461 (vld1_lane_*): Likewise.
30462 (vld1q_lane_*): Likewise.
30464 2013-01-07 Richard Biener <rguenther@suse.de>
30466 * lto-streamer.h (LTO_minor_version): Bump to 2.
30468 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30470 * config/aarch64/aarch64-protos.h
30471 (aarch64_const_double_zero_rtx_p): Rename to...
30472 (aarch64_float_const_zero_rtx_p): ...this.
30473 (aarch64_float_const_representable_p): New.
30474 (aarch64_output_simd_mov_immediate): Likewise.
30475 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
30476 move immediate case.
30477 * config/aarch64/aarch64.c
30478 (aarch64_const_double_zero_rtx_p): Rename to...
30479 (aarch64_float_const_zero_rtx_p): ...this.
30480 (aarch64_print_operand): Allow printing of new constants.
30481 (aarch64_valid_floating_const): New.
30482 (aarch64_legitimate_constant_p): Check for valid floating-point
30484 (aarch64_simd_valid_immediate): Likewise.
30485 (aarch64_vect_float_const_representable_p): New.
30486 (aarch64_float_const_representable_p): Likewise.
30487 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
30488 (aarch64_output_simd_mov_immediate): New.
30489 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
30490 (*movdf_aarch64): Likewise.
30491 * config/aarch64/constraints.md (Ufc): New.
30492 (Y): call aarch64_float_const_zero_rtx.
30493 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
30495 2013-01-07 Richard Biener <rguenther@suse.de>
30497 PR tree-optimization/55888
30498 PR tree-optimization/55862
30499 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
30500 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
30501 not if it is contained therein.
30503 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30505 * config/avr/t-avr: Typo.
30507 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30510 * config/avr/t-avr: Don't automatically rebuild
30511 $(srcdir)/config/avr/t-multilib
30512 $(srcdir)/config/avr/avr-tables.opt
30513 $(srcdir)/doc/avr-mmcu.texi
30514 (avr-mcus): New phony target to build them on request.
30515 (s-avr-mlib, s-avr-mmcu-texi): Remove.
30516 * avr/avr-mcus.def: Adjust comments.
30518 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
30520 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
30522 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30524 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
30526 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30528 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
30530 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
30532 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
30533 to generate profiling.
30534 * config/rs6000/aix64.h (LIB_SPEC): Same.
30536 2013-01-04 Andrew Pinski <apinski@cavium.com>
30538 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
30540 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
30542 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
30544 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
30546 (ix86_expand_move): Ditto.
30547 (ix86_zero_extend_to_Pmode): Ditto.
30548 (ix86_expand_call): Ditto.
30549 (ix86_expand_special_args_builtin): Ditto.
30550 (ix86_expand_builtin): Ditto.
30552 2013-01-04 Richard Biener <rguenther@suse.de>
30554 PR tree-optimization/55862
30555 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
30556 translating them through PHI nodes.
30558 2013-01-04 Martin Jambor <mjambor@suse.cz>
30560 PR tree-optimization/55755
30561 * tree-sra.c (sra_modify_assign): Do not check that an access has no
30562 children when trying to avoid producing a VIEW_CONVERT_EXPR.
30564 2013-01-04 Marek Polacek <polacek@redhat.com>
30566 PR middle-end/55859
30567 * opts.c (default_options_optimization): Clarify error message.
30569 2013-01-04 Richard Biener <rguenther@suse.de>
30571 PR middle-end/55863
30572 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
30575 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30578 * config/pa/pa.md (movsi): Revert previous change.
30579 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
30582 2013-01-03 Richard Henderson <rth@redhat.com>
30584 * config/i386/i386.c (ix86_expand_move): Always assign to op1
30585 after eliminating TLS symbols.
30587 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
30590 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
30591 * graphite-poly.c (debug_gmp_value): Likewise.
30593 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
30596 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
30597 selected code model, define __code_mode_small__, __code_model_medium__,
30598 __code_model_large__, __code_model_32__ or __code_model_kernel__.
30599 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
30600 xchg temporary register with %k. Declare temporary register as
30602 [__x86_64__]: For medium and large code models, preserve %rbx register.
30604 2013-01-03 Richard Biener <rguenther@suse.de>
30606 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
30607 (dump_subscript): Adjust.
30608 (finalize_ddr_dependent): Do not dump redundant info.
30609 (analyze_siv_subscript): Adjust.
30610 (subscript_dependence_tester): Likewise.
30611 (compute_affine_dependence): Likewise.
30613 2013-01-03 Richard Biener <rguenther@suse.de>
30616 2013-01-03 Richard Biener <rguenther@suse.de>
30618 PR tree-optimization/55857
30619 * tree-vect-stmts.c (vectorizable_load): Do not setup
30620 re-alignment for invariant loads.
30622 2013-01-02 Richard Biener <rguenther@suse.de>
30624 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30625 invariant load do not generate a vector load from the scalar location.
30627 2013-01-03 Richard Biener <rguenther@suse.de>
30629 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
30630 for not vectorizing.
30631 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
30632 not build INDIRECT_REFs, call get_name once only.
30633 (vect_create_data_ref_ptr): Likewise. Dump base object kind
30634 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
30636 2013-01-03 Richard Biener <rguenther@suse.de>
30638 PR tree-optimization/55857
30639 * tree-vect-stmts.c (vectorizable_load): Do not setup
30640 re-alignment for invariant loads.
30642 2013-01-03 Richard Biener <rguenther@suse.de>
30645 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
30646 prefer a built-in decl.
30648 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30650 * gcc.c (process_command): Update copyright notice dates.
30651 * gcov.c (print_version): Likewise.
30652 * gcov-dump.c (print_version): Likewise.
30654 PR rtl-optimization/55838
30655 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
30656 iv0.step, iv1.step and step.
30658 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30659 Marc Glisse <marc.glisse@inria.fr>
30661 PR tree-optimization/55832
30662 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
30663 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
30664 integer_{one,zero}_node.
30666 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30669 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
30670 * var-tracking.c (reverse_op): Don't add reverse ops to
30671 VALUEs that have already
30672 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
30674 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
30676 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
30678 2013-01-02 Teresa Johnson <tejohnson@google.com>
30680 * dumpfile.c (dump_loc): Print filename with location.
30681 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
30682 new location_t parameter to emit complete unroll message with
30683 new dump framework.
30684 (canonicalize_loop_induction_variables): Compute loops location
30685 and pass to try_unroll_loop_completely.
30686 * loop-unroll.c (report_unroll_peel): New function.
30687 (peel_loops_completely): Use new dump format with location
30688 for main dumpfile message, and invoke report_unroll_peel on success.
30689 (decide_unrolling_and_peeling): Ditto.
30690 (decide_peel_once_rolling): Remove old dumpfile message subsumed
30691 by report_unroll_peel.
30692 (decide_peel_completely): Ditto.
30693 (decide_unroll_constant_iterations): Ditto.
30694 (decide_unroll_runtime_iterations): Ditto.
30695 (decide_peel_simple): Ditto.
30696 (decide_unroll_stupid): Ditto.
30697 * cfgloop.c (get_loop_location): New function.
30698 * cfgloop.h (get_loop_location): Declare.
30700 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30702 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
30705 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30707 PR middle-end/55198
30708 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
30709 BLKmode objects when EXPAND_MEMORY is specified.
30711 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30713 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
30715 (fold_builtin_cpu): Do not share cpu model decls across statements.
30717 2013-01-02 Jason Merrill <jason@redhat.com>
30720 * tree.c (build_array_type_1): Revert earlier change.
30722 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
30724 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
30726 * config/aarch64/aarch64-tune.md: Re-generate.
30728 2013-01-02 Richard Biener <rguenther@suse.de>
30730 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30731 invariant load do not generate a vector load from the scalar location.
30733 2013-01-02 Richard Biener <rguenther@suse.de>
30736 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
30737 * configure: Regenerate.
30739 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30741 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
30742 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
30743 (expand_builtin_int_roundingfn_2): Keep the original target around
30744 for the fallback case.
30746 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30748 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
30749 to be clear for sign changes.
30751 2013-01-01 Jan Hubicka <jh@suse.cz>
30753 * ipa-inline-analysis.c: Fix formatting.
30755 2013-01-01 Jakub Jelinek <jakub@redhat.com>
30757 PR tree-optimization/55831
30758 * tree-vect-loop.c (get_initial_def_for_induction): Use
30759 gsi_after_labels instead of gsi_start_bb.
30761 Copyright (C) 2013 Free Software Foundation, Inc.
30763 Copying and distribution of this file, with or without modification,
30764 are permitted in any medium without royalty provided the copyright
30765 notice and this notice are preserved.