1 2013-11-06 Christian Bruel <christian.bruel@st.com>
3 * gcc/config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
4 Factorize probabilities, Use adjust_address instead of
5 adjust_automodify_address when possible. Enable for optimize.
6 (sh_expand_strlen): New function.
7 * gcc/config/sh/sh-protos.h (sh_expand_strlen): Declare.
8 * gcc/config/sh/sh.md (strlensi): New pattern.
9 (UNSPEC_BUILTIN_STRLEN): Define.
11 2013-11-06 Jakub Jelinek <jakub@redhat.com>
14 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
15 (expand_assignment): If *bitpos is negative, set *offset
16 and adjust *bitpos, so that it is not negative.
18 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20 * config/i386/bdver3.md : Added two additional decoder units
21 to support issue rate of 4 and remodeled vector unit.
22 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
23 architectures is set to 4.
24 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
25 lookahead is set to 4 for BD architectures.
27 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
29 * config/rs6000/rs6000.c (rs6000_option_override_internal):
30 Remove restriction against use of VSX instructions when generating
31 code for little endian mode.
33 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
35 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
36 for both big and little endian.
37 (mulv8hi3): Swap input operands for merge high and merge low
38 instructions for little endian.
40 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
42 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
43 define_insn to define_expand that uses even patterns for big
44 endian and odd patterns for little endian.
45 (vec_widen_smult_even_v16qi): Likewise.
46 (vec_widen_umult_even_v8hi): Likewise.
47 (vec_widen_smult_even_v8hi): Likewise.
48 (vec_widen_umult_odd_v16qi): Likewise.
49 (vec_widen_smult_odd_v16qi): Likewise.
50 (vec_widen_umult_odd_v8hi): Likewise.
51 (vec_widen_smult_odd_v8hi): Likewise.
52 (altivec_vmuleub): New define_insn.
53 (altivec_vmuloub): Likewise.
54 (altivec_vmulesb): Likewise.
55 (altivec_vmulosb): Likewise.
56 (altivec_vmuleuh): Likewise.
57 (altivec_vmulouh): Likewise.
58 (altivec_vmulesh): Likewise.
59 (altivec_vmulosh): Likewise.
61 2013-11-05 Mike Stump <mikestump@comcast.net>
63 * Makefile.in (mostlyclean): Remove c-family objects.
65 2013-11-05 Ian Lance Taylor <iant@google.com>
67 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
68 If possible, add .cfi directives to record change to bx.
69 * config/i386/i386.c (ix86_emit_cfi): New function.
70 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
72 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
75 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
76 immediately after a label for a tablejump pattern.
78 * config/arm/arm.c (is_jump_table): Remove.
79 (create_fix_barrier): Use tablejump_p instead.
80 (arm_reorg): Likewise.
81 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
82 (thumb2_output_casesi): Likewise.
83 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
84 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
85 casesi_shift_media, casesi_load_media): Likewise.
86 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
87 * config/microblaze/microblaze.md: Likewise.
89 2013-11-05 Tobias Burnus <burnus@net-b.de>
91 * doc/invoke.texi (-Wdate-time): Document.
93 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
95 * double-int.c (lshift_double, rshift_double): Remove
96 SHIFT_COUNT_TRUNCATED handling.
98 2013-11-05 Jeff Law <law@redhat.com>
100 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
101 * common.opt (-fisolate-erroneous-paths): Add option and
103 * gimple-ssa-isolate-paths.c: New file.
104 * gimple.c (check_loadstore): New function.
105 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
106 Verify OP is in the argument list and the argument corresponding
107 to OP is a pointer type. Use operand_equal_p rather than
108 pointer equality when testing if OP is on the nonnull list.
109 Use check_loadstore rather than count_ptr_derefs. Handle
110 GIMPLE_RETURN statements.
111 * tree-vrp.c (infer_nonnull_range): Remove.
112 * gimple.h (infer_nonnull_range): Declare.
113 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
114 * passes.def: Add pass_isolate_erroneous_paths.
115 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
116 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
117 * tree-ssa.c (struct count_ptr_d): Remove.
118 (count_ptr_derefs, count_uses_and_derefs): Remove.
119 * tree-ssa.h (count_uses_and_derefs): Remove.
121 2013-11-05 Jakub Jelinek <jakub@redhat.com>
123 PR rtl-optimization/58997
124 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
125 get_iv_value to be in iv->mode rather than iv->extend_mode.
126 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
127 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
128 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
131 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
133 * gimple.h: Move some prototypes to gimple-expr.h and add to include
135 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
136 Move to gimple-expr.h.
137 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
138 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
139 is_gimple_id, virtual_operand_p, is_gimple_addressable,
140 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
142 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
143 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
144 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
145 is_gimple_address, is_gimple_invariant_address,
146 is_gimple_ip_invariant_address, is_gimple_min_invariant,
147 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
148 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
149 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
150 gimple_decl_printable_name, useless_type_conversion_p,
151 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
153 * gimple-expr.c: New File.
154 (useless_type_conversion_p, gimple_set_body, gimple_body,
155 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
156 gimple_can_coalesce_p, extract_ops_from_tree_1,
157 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
158 is_gimple_address, is_gimple_invariant_address,
159 is_gimple_ip_invariant_address, is_gimple_min_invariant,
160 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
161 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
162 is_gimple_mem_ref_addr): Relocate here.
163 * Makefile.in (OBJS): Add gimple-expr.o.
165 2013-11-05 David Malcolm <dmalcolm@redhat.com>
167 * gengtype-parse.c (struct_field_seq): Support empty structs.
169 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
171 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
173 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
175 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
177 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
179 (LONG_DOUBLE_TYPE_SIZE): New define.
180 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
182 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
185 * doc/extend.texi [visibility ("visibility_type")]: Add example
186 about visibility attribute on namespace declaration.
188 2013-11-05 Richard Biener <rguenther@suse.de>
191 * passes.def (all_passes): Start with pass_fixup_cfg again.
193 2013-11-05 Richard Biener <rguenther@suse.de>
195 PR tree-optimization/58955
196 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
199 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
201 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
203 (vec_pack_ufix_trunc_v2df): Likewise.
205 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
208 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
209 pattern, instead of word_mode.
211 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
213 (set_storage_via_setmem): Likewise.
215 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
217 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
218 gimple.h and the rest of the condition in eliminate_build.
219 (eliminate_build): Call new routine.
220 * gimple.h (phi_ssa_name_p): Delete.
222 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
224 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
225 case of no prefix and reserving zero slots, because when that's the
226 case we'll never get here.
227 * vec.h (va_heap::reserve): Don't try and handle
228 vec_prefix::calculate_allocation returning zero because that should
231 2013-11-05 Richard Biener <rguenther@suse.de>
234 * tree-dfa.c (get_ref_base_and_extent): Merge common code
235 in MEM_REF and TARGET_MEM_REF handling. Make sure to
236 process trailing array detection before diving into the
237 view-converted object (and possibly apply some extra offset).
239 2013-11-05 Joseph Myers <joseph@codesourcery.com>
241 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
243 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
245 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
247 PR tree-optimization/58958
248 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
249 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
251 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
253 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
254 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
256 2013-11-05 Jakub Jelinek <jakub@redhat.com>
258 PR tree-optimization/58984
259 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
260 set *SIZE_P if non-NULL on success.
261 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
263 (ipcp_transform_function): Likewise. Punt if size of access
264 is different from TYPE_SIZE on v->value's type.
266 2013-11-05 Tobias Burnus <burnus@net-b.de>
268 * doc/invoke.texi (-fopenmp-simd): Document new option.
269 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
270 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
271 * tree.c (attribute_value_equal): Ditto.
273 2013-11-04 Wei Mi <wmi@google.com>
275 * sched-rgn.c (add_branch_dependences): Keep insns in
276 a SCHED_GROUP at the end of BB to remain their location.
278 2013-11-04 Wei Mi <wmi@google.com>
280 * gcc/config/i386/i386.c (memory_address_length): Extract a part
281 of code to rip_relative_addr_p.
282 (rip_relative_addr_p): New Function.
283 (ix86_macro_fusion_p): Ditto.
284 (ix86_macro_fusion_pair_p): Ditto.
285 * gcc/config/i386/i386.h: Add new tune features about macro-fusion.
286 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
287 * gcc/doc/tm.texi: Generated.
288 * gcc/doc/tm.texi.in: Ditto.
289 * gcc/haifa-sched.c (try_group_insn): New Function.
290 (group_insns_for_macro_fusion): Ditto.
291 (sched_init): Call group_insns_for_macro_fusion.
292 * gcc/target.def: Add two hooks: macro_fusion_p and
295 2013-11-04 Kostya Serebryany <kcc@google.com>
297 Update to match the changed asan API.
298 * asan.c (asan_function_start): New function.
299 (asan_emit_stack_protection): Update the string stored in the
300 stack red zone to match new API. Store the PC of the current
301 function in the red zone.
302 (asan_global_struct): Update the __asan_global definition to match
304 (asan_add_global): Ditto.
305 * asan.h (asan_function_start): New prototype.
306 * final.c (final_start_function): Call asan_function_start.
307 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
310 2013-11-04 Wei Mi <wmi@google.com>
312 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): Separate
313 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
314 * gcc/config/i386/i386.c (ix86_option_override_internal): Ditto.
315 (ix86_issue_rate): Ditto.
316 (ix86_adjust_cost): Ditto.
317 (ia32_multipass_dfa_lookahead): Ditto.
318 (ix86_sched_init_global): Ditto.
319 (get_builtin_code_for_version): Ditto.
320 * gcc/config/i386/i386.h (enum target_cpu_default): Ditto.
321 (enum processor_type): Ditto.
322 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
324 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
326 PR rtl-optimization/58967
327 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
328 !lra_in_progress for mode sizes bigger word.
330 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
332 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
333 arguments to merge instruction for little endian.
334 (vec_widen_umult_lo_v16qi): Likewise.
335 (vec_widen_smult_hi_v16qi): Likewise.
336 (vec_widen_smult_lo_v16qi): Likewise.
337 (vec_widen_umult_hi_v8hi): Likewise.
338 (vec_widen_umult_lo_v8hi): Likewise.
339 (vec_widen_smult_hi_v8hi): Likewise.
340 (vec_widen_smult_lo_v8hi): Likewise.
342 2013-11-04 Ian Lance Taylor <iant@google.com>
344 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
345 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
346 builtins that take pointers.
347 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
348 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
349 (run_gcc): Pass -fnon-call-exceptions.
351 2013-11-04 Jakub Jelinek <jakub@redhat.com>
353 * optabs.c (expand_vec_perm): Revert one incorrect line from
356 PR tree-optimization/58978
357 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
358 use_stmt by single_imm_use directly. Only call single_imm_use
361 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
363 PR rtl-optimization/58968
364 * lra-spills.c (return_regno_p): New function.
365 (lra_final_code_change): Use it.
367 2013-11-04 Joseph Myers <joseph@codesourcery.com>
369 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
370 * target.def (float_exceptions_rounding_supported_p): New hook.
371 * targhooks.c (default_float_exceptions_rounding_supported_p): New
373 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
374 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
376 * doc/tm.texi: Regenerate.
377 * config.gcc (powerpc*-*-linux*): Set extra_objs.
378 * config/rs6000/rs6000-linux.c: New file.
379 * config/rs6000/rs6000-protos.h
380 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
381 * config/rs6000/linux.h
382 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
383 * config/rs6000/linux64.h
384 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
385 * config/rs6000/t-linux (rs6000-linux.o): New rule.
386 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
388 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
390 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
391 Replace the define_insn_and_split with a define_insn and two
392 define_splits, with the split after reload re-permuting the source
393 register to its original value.
394 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
395 (*vsx_le_perm_store_v8hi): Likewise.
396 (*vsx_le_perm_store_v16qi): Likewise.
398 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
400 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
403 2013-11-04 Jakub Jelinek <jakub@redhat.com>
405 PR tree-optimization/58946
406 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
407 bbs with bbinfo[idx].op != NULL before all blocks with
408 bbinfo[idx].op == NULL.
410 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
412 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
413 (avr_dump_double_int_hex): Likewise.
414 (avr_log_vadump): Remove %D and %X handling.
415 * config/avr/avr.c (avr_double_int_push_digit): Delete.
416 (avr_map_op_t): Change map from double_int to unsigned int.
417 (avr_map_op): Update accordingly.
418 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
419 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
420 unsigned ints rather than double_ints.
422 2013-11-03 Marek Polacek <polacek@redhat.com>
424 Implement -fsanitize=vla-bound.
425 * opts.c (common_handle_option): Handle vla-bound.
426 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
427 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
428 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
430 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
432 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
435 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
437 * config/i386/constraints.md (Ts, Tv): New address constrains.
438 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
439 constraint for address_no_seg_operand.
440 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
441 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
442 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
443 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
444 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
445 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
446 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
447 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
448 constraint for vsib_address_operand.
450 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
452 * gcse.c (pre_delete): Remove references to regmove from comments.
453 * recog.c: (validate_replace_rtx_1): Likewise.
454 * config/rl78/rl78.c: Likewise.
455 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
456 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
457 * common/config/mmix/mmix-common.c: Likewise.
459 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
461 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
462 * gimplify.c (gimplify_compound_lval): Likewise.
463 * graphite-clast-to-gimple.c (gloog): Likewise.
464 * graphite-dependences.c (loop_is_parallel_p): Likewise.
465 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
466 (limit_scop); Likewise.
467 (build_scops): Likewise.
468 (dot_scop): Likewise.
469 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
470 (build_scop_drs): Likewise.
471 (insert_stmts): Likewise.
472 (insert_out_of_ssa_copy): Likewise.
473 (remove_phi): Likewise.
474 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
475 * hw-doloop.c (discover_loop): Likewise.
476 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
477 * tree-dfa.c (dump_enumerated_decls): Likewise.
478 * tree-if-conv.c (if_convertable_loop_p): Likewise.
479 * tree-inline.c (tree_function_versioning): Likewise.
480 * tree-loop-distribution.c (build_rdg): Likewise.
481 (rdg_flag_vertex_and_dependent): Likewise.
482 (distribute_loop): Likewise.
483 * tree-parloops.c (loop_parallel_p): Likewise.
484 (eliminate_local_variables): Likewise.
485 (separate_decls_in_region): Likewise.
486 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
487 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
488 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
489 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
490 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
491 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
492 (vectorizable_condition): Likewise.
494 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
496 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
497 * configure: Regenerate.
498 * config/i386/i386.md (*movdi_internal): Change
499 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
500 (*movdf_internal): Ditto.
501 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
502 * config/i386/sse.md (vec_concatv2di): Output interunit movq
503 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
505 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
507 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
510 2013-10-31 David Malcolm <dmalcolm@redhat.com>
512 Automated part of renaming of symtab_node_base to symtab_node.
514 Patch autogenerated by rename_symtab.py from
515 https://github.com/davidmalcolm/gcc-refactoring-scripts
516 revision 58bb219cc090b2f4516a9297d868c245495ee622
517 with ChangeLog entry fixed up by hand.
519 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
521 (cgraph_node_for_asm): Likewise.
522 * cgraph.h (symtab_node_base): Likewise.
523 (cgraph_node): Likewise.
524 (varpool_node): Likewise.
525 (is_a_helper <cgraph_node>::test): Likewise.
526 (is_a_helper <varpool_node>::test): Likewise.
527 (symtab_nodes): Likewise.
528 (symtab_register_node): Likewise.
529 (symtab_unregister_node): Likewise.
530 (symtab_remove_node): Likewise.
531 (symtab_get_node): Likewise.
532 (symtab_node_for_asm): Likewise.
533 (symtab_node_asm_name): Likewise.
534 (symtab_node_name): Likewise.
535 (symtab_insert_node_to_hashtable): Likewise.
536 (symtab_add_to_same_comdat_group): Likewise.
537 (symtab_dissolve_same_comdat_group_list): Likewise.
538 (dump_symtab_node): Likewise.
539 (debug_symtab_node): Likewise.
540 (dump_symtab_base): Likewise.
541 (verify_symtab_node): Likewise.
542 (verify_symtab_base): Likewise.
543 (symtab_used_from_object_file_p): Likewise.
544 (symtab_alias_ultimate_target): Likewise.
545 (symtab_resolve_alias): Likewise.
546 (fixup_same_cpp_alias_visibility): Likewise.
547 (symtab_for_node_and_aliases): Likewise.
548 (symtab_nonoverwritable_alias): Likewise.
549 (availability symtab_node_availability): Likewise.
550 (symtab_semantically_equivalent_p): Likewise.
551 (fixup_same_cpp_alias_visibility): Likewise.
552 (symtab_prevail_in_asm_name_hash): Likewise.
555 (varpool_first_variable): Likewise.
556 (varpool_next_variable): Likewise.
557 (varpool_first_static_initializer): Likewise.
558 (varpool_next_static_initializer): Likewise.
559 (varpool_first_defined_variable): Likewise.
560 (varpool_next_defined_variable): Likewise.
561 (cgraph_first_defined_function): Likewise.
562 (cgraph_next_defined_function): Likewise.
563 (cgraph_first_function): Likewise.
564 (cgraph_next_function): Likewise.
565 (cgraph_first_function_with_gimple_body): Likewise.
566 (cgraph_next_function_with_gimple_body): Likewise.
567 (symtab_alias_target): Likewise.
568 (symtab_real_symbol_p): Likewise.
569 (symtab_can_be_discarded): Likewise.
570 * cgraphbuild.c (mark_address): Likewise.
571 (mark_load): Likewise.
572 (mark_store): Likewise.
573 * cgraphunit.c (decide_is_symbol_needed): Likewise.
575 (enqueue_node): Likewise.
576 (referred_to_p): Likewise.
577 (cgraph_process_same_body_aliases): Likewise.
578 (analyze_functions): Likewise.
579 (handle_alias_pairs): Likewise.
580 (output_weakrefs): Likewise.
582 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
583 * ipa-inline-analysis.c (inline_write_summary): Likewise.
584 * ipa-prop.c (remove_described_reference): Likewise.
585 (try_decrement_rdesc_refcount): Likewise.
586 (ipa_edge_duplication_hook): Likewise.
587 * ipa-ref.c (ipa_record_reference): Likewise.
588 (ipa_maybe_record_reference): Likewise.
589 (ipa_clone_ref): Likewise.
590 (ipa_clone_references): Likewise.
591 (ipa_clone_referring): Likewise.
592 (ipa_find_reference): Likewise.
593 (ipa_remove_stmt_references): Likewise.
594 (ipa_clear_stmts_in_references): Likewise.
595 * ipa-ref.h (symtab_node_base): Likewise.
597 (ipa_record_reference): Likewise.
598 (ipa_maybe_record_reference): Likewise.
599 (ipa_clone_references): Likewise.
600 (ipa_clone_referring): Likewise.
601 (ipa_clone_ref): Likewise.
602 (ipa_find_reference): Likewise.
603 (ipa_remove_stmt_references): Likewise.
604 (ipa_clear_stmts_in_references): Likewise.
605 * ipa-reference.c (ipa_reference_write_optimization_summary):
607 * ipa.c (enqueue_node): Likewise.
608 (process_references): Likewise.
609 (walk_polymorphic_call_targets): Likewise.
610 (symtab_remove_unreachable_nodes): Likewise.
611 (address_taken_from_non_vtable_p): Likewise.
612 (comdat_can_be_unshared_p_1): Likewise.
613 (comdat_can_be_unshared_p): Likewise.
614 (can_replace_by_local_alias): Likewise.
615 (function_and_variable_visibility): Likewise.
616 * is-a.h: Likewise (within example in comment).
617 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
618 (lto_symtab_encoder_encode): Likewise.
619 (lto_symtab_encoder_delete_node): Likewise.
620 (lto_symtab_encoder_in_partition_p): Likewise.
621 (lto_set_symtab_encoder_in_partition): Likewise.
622 (output_refs): Likewise.
623 (compute_ltrans_boundary): Likewise.
624 (output_symtab): Likewise.
625 (input_node): Likewise.
626 (input_ref): Likewise.
627 (input_edge): Likewise.
628 (input_cgraph_1): Likewise.
629 (input_refs): Likewise.
630 (output_cgraph_opt_summary): Likewise.
631 (input_node_opt_summary): Likewise.
632 (input_cgraph_opt_section): Likewise.
633 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
635 * lto-streamer-out.c (lto_output): Likewise.
636 (output_symbol_p): Likewise.
637 (produce_symtab): Likewise.
638 * lto-streamer.h (lto_encoder_entry): Likewise.
639 (lto_free_function_in_decl_state_for_node): Likewise.
640 (lto_symtab_encoder_encode): Likewise.
641 (lto_symtab_encoder_delete_node): Likewise.
642 (lto_symtab_encoder_in_partition_p): Likewise.
643 (lto_set_symtab_encoder_in_partition): Likewise.
644 (lto_symtab_encoder_lookup): Likewise.
645 (lsei_node): Likewise.
646 (lto_symtab_encoder_deref): Likewise.
647 * symtab.c (symtab_hash): Likewise.
648 (assembler_name_hash): Likewise.
649 (symtab_nodes): Likewise.
650 (hash_node): Likewise.
652 (hash_node_by_assembler_name): Likewise.
653 (eq_assembler_name): Likewise.
654 (insert_to_assembler_name_hash): Likewise.
655 (unlink_from_assembler_name_hash): Likewise.
656 (symtab_prevail_in_asm_name_hash): Likewise.
657 (symtab_register_node): Likewise.
658 (symtab_insert_node_to_hashtable): Likewise.
659 (symtab_unregister_node): Likewise.
660 (symtab_get_node): Likewise.
661 (symtab_remove_node): Likewise.
662 (symtab_initialize_asm_name_hash): Likewise.
663 (symtab_node_for_asm): Likewise.
664 (symtab_add_to_same_comdat_group): Likewise.
665 (symtab_dissolve_same_comdat_group_list): Likewise.
666 (symtab_node_asm_name): Likewise.
667 (symtab_node_name): Likewise.
668 (dump_symtab_base): Likewise.
669 (dump_symtab_node): Likewise.
670 (dump_symtab): Likewise.
671 (debug_symtab_node): Likewise.
672 (verify_symtab_base): Likewise.
673 (verify_symtab_node): Likewise.
674 (verify_symtab): Likewise.
675 (symtab_used_from_object_file_p): Likewise.
676 (symtab_node_availability): Likewise.
677 (symtab_alias_ultimate_target): Likewise.
678 (fixup_same_cpp_alias_visibility): Likewise.
679 (symtab_resolve_alias): Likewise.
680 (symtab_for_node_and_aliases): Likewise.
681 (symtab_for_node_and_aliases): Likewise.
682 (symtab_nonoverwritable_alias_1): Likewise.
683 (symtab_nonoverwritable_alias): Likewise.
684 (symtab_semantically_equivalent_p): Likewise.
685 * value-prof.c (init_node_map): Likewise.
686 * varasm.c (find_decl): Likewise.
687 * varpool.c (varpool_node_for_asm): Likewise.
688 (varpool_remove_unreferenced_decls): Likewise.
690 2013-10-31 David Malcolm <dmalcolm@redhat.com>
692 Manual part of renaming of symtab_node_base to symtab_node.
694 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
695 clashes with the preferred name for the base class.
696 (const_symtab_node): Remove redundant typedef.
698 2013-10-31 Jakub Jelinek <jakub@redhat.com>
700 * optabs.c (expand_vec_perm): Avoid vector mode punning
702 * expmed.c (store_bit_field_1): Likewise.
703 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
704 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
705 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
706 * config/i386/i386.c (ix86_expand_vector_move_misalign,
707 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
708 ix86_expand_sse_unpack, ix86_expand_args_builtin,
709 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
710 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
711 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
712 expand_vec_perm_vpshufb2_vpermq,
713 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
714 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
715 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
716 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
719 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
721 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
722 Use gcc_unreachable for unhandled modes. Do not check results of
723 expand_simple_binop. If not expanded to target, move the result.
725 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
726 Shiva Chen <shiva0217@gmail.com>
728 * config.gcc (nds32*-*-*): Add nds32 target.
729 * config/nds32/nds32.c: New file.
730 * config/nds32/nds32.h: New file.
731 * config/nds32/nds32.md: New file.
732 * config/nds32/constants.md: New file.
733 * config/nds32/constraints.md: New file.
734 * config/nds32/iterators.md: New file.
735 * config/nds32/nds32-doubleword.md: New file.
736 * config/nds32/nds32-intrinsic.md: New file.
737 * config/nds32/nds32_intrinsic.h: New file.
738 * config/nds32/nds32-modes.def: New file.
739 * config/nds32/nds32-multiple.md: New file.
740 * config/nds32/nds32.opt: New file.
741 * config/nds32/nds32-opts.h: New file.
742 * config/nds32/nds32-protos.h: New file.
743 * config/nds32/nds32-peephole2.md: New file.
744 * config/nds32/pipelines.md: New file.
745 * config/nds32/predicates.md: New file.
746 * config/nds32/t-mlibs: New file.
747 * common/config/nds32: New directory and files.
749 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
750 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
751 * doc/install.texi (Cross-Compiler-Specific Options): Document
752 --with-nds32-lib for nds32 target.
753 * doc/extend.texi (Function Attributes, Target Builtins): Document
754 nds32 specific attributes.
756 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
758 * lra-constraints (process_alt_operands): Use the result
759 elimination register for operand when matching constraints.
761 2013-10-31 Jakub Jelinek <jakub@redhat.com>
763 * tree-vrp.c (maybe_set_nonzero_bits): New function.
764 (remove_range_assertions): Call it.
766 * tree.c (tree_ctz): New function.
767 * tree.h (tree_ctz): New prototype.
768 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
769 first argument from tree to const_tree.
770 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
771 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
772 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
774 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
775 * tree-vect-loop.c (vect_analyze_loop_operations,
776 vect_transform_loop): Don't force scalar loop for bound just because
777 number of iterations is unknown, only do it if it is not known to be
778 a multiple of vectorization_factor.
779 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
781 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
782 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
783 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
784 (set_nonzero_bits, get_nonzero_bits): New prototypes.
785 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
786 a default def isn't partially constant.
787 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
788 is known to be partially zero.
789 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
790 to see if a default def isn't partially constant.
791 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
792 creation of a range, if VR_RANGE, try to improve nonzero_bits from
794 (set_nonzero_bits, get_nonzero_bits): New functions.
796 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
797 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
798 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
799 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
800 uses but in the condition and ASSERT_EXPR and the other successor of
801 the predecessor bb is __builtin_unreachable (), set_range_info of the
802 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
804 2013-10-31 Martin Jambor <mjambor@suse.cz>
806 PR rtl-optimization/58934
808 2013-10-30 Martin Jambor <mjambor@suse.cz>
809 PR rtl-optimization/10474
810 * ira.c (find_moveable_pseudos): Do not calculate dominance info
812 (interesting_dest_for_shprep): New function.
813 (split_live_ranges_for_shrink_wrap): Likewise.
814 (ira): Calculate dominance info and df analysis. Call
815 split_live_ranges_for_shrink_wrap.
817 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
818 Yury Gribov <y.gribov@samsung.com>
821 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
822 shadow pointer to avoid clobbering the main one.
824 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
826 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
828 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
831 * ira-color.c (update_costs_from_copies): Add new parameter. Use
832 it for calling update_costs_from_allocno.
833 (assign_hard_reg): Call restore_costs_from_copies only for
834 !retry_p. Pass new argument to update_costs_from_copies.
835 (color_pass): Pass new argument to update_costs_from_copies.
836 (ira_mark_allocation_change): Ditto.
838 2013-10-30 Sharad Singhai <singhai@google.com>
841 * opts.c (common_handle_option): Remove deprecated option
842 -ftree-vectorizer-verbose.
843 * doc/invoke.texi (Debugging Options): Ditto.
844 * opts-global.c (handle_common_deferred_options): Ditto.
845 (dump_remap_tree_vectorizer_verbose): Delete.
846 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
848 2013-10-30 DJ Delorie <dj@redhat.com>
850 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
851 that take no arguments.
853 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
856 * reload1.c (update_eliminables_and_spill): New function, broken
858 (reload): Use it. Check for frame size change after frame size
859 alignment, and call update_eliminables_and_spill first if continue-ing.
861 2013-10-30 Cong Hou <congh@google.com>
864 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
865 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
866 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
868 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
870 PR rtl-optimization/58369
871 * reload1.c (compute_reload_subreg_offset): New function.
872 (choose_reload_regs): Use it to pass endian-correct
873 offset to subreg_regno_offset.
875 2013-10-30 Tobias Burnus <burnus@net-b.de>
878 * gcc/tree-cfg.c (replace_loop_annotate): Replace warning by
881 2013-10-30 Jason Merrill <jason@redhat.com>
883 * configure.ac (loose_warn): Add -Wno-format if
884 --disable-build-format-warnings.
886 2013-10-30 David Malcolm <dmalcolm@redhat.com>
888 * cgraphunit.c (analyze_functions): Split symtab_node declarations
889 onto multiple lines to make things easier for rename_symtab.py.
891 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
892 (symtab_semantically_equivalent_p): Likewise.
894 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
897 * lra.c (check_rtl): Remove address check before LRA work.
899 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
901 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
902 POINTER_PLUS_EXPR in the defining statement.
904 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
907 * tree-pass.h (make_pass_regmove): Remove.
908 * timevar.def (TV_REGMOVE): Remove.
909 * passes.def (pass_regmove): Remove.
910 * opts.c (default_options_table): Remove entry for regmove.
911 * doc/passes.texi: Remove regmove pass description.
912 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
914 (-fdump-rtl-regmove): Ditto.
915 * common.opt (foptimize-register-move, fregmove): Ignore.
916 * Makefile.in (OBJS): Remove regmove.o.
918 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
920 (struct ira_allocno) New member allocno_prefs.
921 (ALLOCNO_PREFS): New macro.
922 (ira_prefs, ira_prefs_num): New external vars.
923 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
924 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
925 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
927 (ira_add_allocno_copy_to_list): Remove prototype.
928 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
929 (ira_pref_iterator): New type.
930 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
931 (FOR_EACH_PREF): New macro.
932 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
933 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
935 (ira_setup_alts): New function.
936 (decrease_live_ranges_number): New function.
937 (ira): Call the above function.
938 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
939 (ira_create_allocno): Initialize allocno prefs.
940 (pref_pool, pref_vec): New static vars.
941 (initiate_prefs, find_allocno_pref, ira_create_pref): New
943 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
944 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
945 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
946 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
947 (ira_add_allocno_copy_to_list): Make static. Rename to
948 add_allocno_copy_to_list.
949 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
950 swap_allocno_copy_ends_if_necessary.
951 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
952 ira_remove_allocno_prefs.
953 (ira_flattening): Ditto.
954 (ira_build): Call initiate_prefs, print_prefs.
955 (ira_destroy): Call finish_prefs.
956 * ira-color.c (struct update_cost_record): New.
957 (struct allocno_color_data): Add new member update_cost_records.
958 (update_cost_record_pool): New static var.
959 (init_update_cost_records, get_update_cost_record): New functions.
960 (free_update_cost_record_list, finish_update_cost_records): Ditto.
961 (struct update_cost_queue_elem): Add member from.
962 (initiate_cost_update): Call init_update_cost_records.
963 (finish_cost_update): Call finish_update_cost_records.
964 (queue_update_cost, get_next_update_cost): Add new param from.
965 (Update_allocno_cost, update_costs_from_allocno): New functions.
966 (update_costs_from_prefs): Ditto.
967 (update_copy_costs): Rename to update_costs_from_copies.
968 (restore_costs_from_copies): New function.
969 (update_conflict_hard_regno_costs): Don't go back.
970 (assign_hard_reg): Call restore_costs_from_copies. Add printing
972 (pop_allocnos): Add priniting more debug info.
973 (color_allocnos): Remove prefs for conflicting hard regs.
974 Call update_costs_from_prefs.
975 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
976 (get_dup_num): Rename, modify, and move to ira.c
977 (process_regs_for_copy): Add prefs.
978 (add_insn_allocno_copies): Put src as first arg of
979 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
980 * ira-costs.c (record_reg_classes): Modify and move code into
981 record_operands_costs.
982 (find_costs_and_classes): Create prefs for the hard reg of small
984 (process_bb_node_for_hard_reg_moves): Add prefs.
986 2013-10-30 Richard Biener <rguenther@suse.de>
989 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
990 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
992 (pre_and_rev_post_order_compute): ... which now wraps it.
993 * graph.c (draw_cfg_nodes_no_loops): Use
994 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
997 2013-10-30 Christian Bruel <christian.bruel@st.com>
999 * gcc/config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
1000 (sh_expand_cmpstr): Handle known align and schedule improvements.
1001 * gcc/config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
1002 * gcc/config/sh/sh.md (cmpstrnsi): New pattern.
1004 2013-10-30 Martin Jambor <mjambor@suse.cz>
1006 PR rtl-optimization/10474
1007 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1009 (interesting_dest_for_shprep): New function.
1010 (split_live_ranges_for_shrink_wrap): Likewise.
1011 (ira): Calculate dominance info and df analysis. Call
1012 split_live_ranges_for_shrink_wrap.
1014 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1017 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1019 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1021 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1022 * tree.h (POINTER_BOUNDS_P): New.
1023 (BOUNDED_TYPE_P): New.
1025 (pointer_bounds_type_node): New.
1026 * tree.c (build_common_tree_nodes): Initialize
1027 pointer_bounds_type_node.
1028 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1029 (gimple_call_num_nobnd_args): New.
1030 (gimple_call_nobnd_arg): New.
1031 (gimple_return_retbnd): New.
1032 (gimple_return_set_retbnd): New
1033 * gimple.c (gimple_build_return): Increase number of ops
1034 for return statement.
1035 (gimple_call_get_nobnd_arg_index): New.
1036 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1038 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1040 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1041 with "chkp ctor" and "bnd_legacy" attributes.
1042 * gimplify.c (gimplify_init_constructor): Avoid infinite
1043 loop during gimplification of bounds initializer.
1045 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1047 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1048 (handle_bnd_legacy): New.
1049 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1050 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1053 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1055 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1056 (BT_FN_PTR_CONST_PTR): New.
1057 (BT_FN_CONST_PTR_CONST_PTR): New.
1058 (BT_FN_PTR_CONST_PTR_SIZE): New.
1059 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1060 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1061 (BT_FN_VOID_CONST_PTR_SIZE): New.
1062 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1063 * chkp-builtins.def: New.
1064 * builtins.def: include chkp-builtins.def.
1065 (DEF_CHKP_BUILTIN): New.
1066 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1067 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1068 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1069 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1070 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1071 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1072 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1073 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1074 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1075 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1076 * common.opt (fcheck-pointer-bounds): New.
1077 * toplev.c (process_options): Check Pointer Bounds Checker is
1079 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1081 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1083 * target.def (builtin_chkp_function): New.
1084 (chkp_bound_type): New.
1085 (chkp_bound_mode): New.
1086 (fn_abi_va_list_bounds_size): New.
1087 (load_bounds_for_arg): New.
1088 (store_bounds_for_arg): New.
1089 * targhooks.h (default_load_bounds_for_arg): New.
1090 (default_store_bounds_for_arg): New.
1091 (default_fn_abi_va_list_bounds_size): New.
1092 (default_chkp_bound_type): New.
1093 (default_chkp_bound_mode): New.
1094 (default_builtin_chkp_function): New.
1095 * targhooks.c (default_load_bounds_for_arg): New.
1096 (default_store_bounds_for_arg): New.
1097 (default_fn_abi_va_list_bounds_size): New.
1098 (default_chkp_bound_type): New.
1099 (default_chkp_bound_mode); New.
1100 (default_builtin_chkp_function): New.
1101 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1102 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1103 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1104 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1105 (TARGET_CHKP_BOUND_TYPE): New.
1106 (TARGET_CHKP_BOUND_MODE): New.
1107 * doc/tm.texi: Regenerated.
1108 * langhooks.h (lang_hooks): Add chkp_supported field.
1109 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1110 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1112 2013-10-29 Andrew Pinski <apinski@cavium.com>
1114 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1115 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1116 fails, combining the branches anyways.
1117 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1118 increases the number of combinings.
1119 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1121 2013-10-29 Mike Stump <mikestump@comcast.net>
1123 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1124 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1125 (make_vector_mode): Increase namebuf to 16.
1126 (emit_insn_modes_h): When processing BImode, don't
1127 also match partial int modes.
1128 (emit_class_narrowest_mode): Likewise.
1130 * config/bfin/bfin-modes.def: Add precision to PDI.
1131 * config/m32c/m32c-modes.def: Add precision to PSI.
1132 * config/msp430/msp430-modes.def: Add precision to PSI.
1133 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1134 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1136 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1139 * config/sh/sh.md (*addc): Rename existing variations to ...
1140 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1141 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1142 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1143 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1145 2013-10-29 Teresa Johnson <tejohnson@google.com>
1148 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1149 insanities when updating probabilities.
1151 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1153 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1154 prettyprinter for cgraph_node to reflect the conversion of the
1155 symtable types to a C++ class hierarchy: it now *is* a
1156 symtab_node_base, rather than having one (named "symbol").
1158 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1160 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
1161 __cilkrts_pop_frame. If matched, then return true for built-in
1163 (expand_builtin): Added BUILT_IN_CILK_DETACH and
1164 BUILT_IN_CILK_POP_FRAME case.
1165 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
1166 (lhs_cilk_detect_spawn): Likewise.
1167 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
1168 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
1169 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
1170 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
1171 (LANG_HOOKS_CILKPLUS): Likewise.
1172 * tree.h (CILK_SPAWN_FN): Likewise.
1173 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
1174 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
1175 (OBJS): Added cilk-common.o.
1176 (BUILTINS_DEF): Added cilk-builtins.def.
1177 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
1178 (lhd_cilk_detect_spawn): Likewise.
1179 * langhooks.h (lang_hooks_for_cilkplus): New struct.
1180 (struct lang_hooks): Added new field called "cilkplus."
1181 * cilk-common.c: New file.
1183 * cilk-builtins.def: Likewise.
1184 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
1185 "__cilk" macro and set it to 200.
1186 * function.h (struct function::cilk_frame_decl): New field.
1187 (struct function::is_cilk_function): Likewise.
1188 (struct function::calls_cilk_spawn): Likewise.
1189 * gimplify.c (gimplify_call_expr): Added a check if the function call
1190 being gimplified is a spawn detach point. If so, then add pop_frame
1191 and detach function calls.
1192 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
1193 for gimplifying _Cilk_spawn and _Cilk_sync statements.
1194 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
1195 function. If so, added a _Cilk_sync and gimplified it.
1196 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
1197 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
1198 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
1200 (can_inline_edge_p): Prevent inling of spawnee function.
1201 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
1202 for functions that use Cilk keywords.
1203 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
1204 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
1205 CILK_SYNC_STMT cases.
1206 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
1208 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
1209 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1210 * passes.texi (Cilk Keywords): New section that describes the compiler
1211 code changes for handling Cilk Keywords.
1213 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1215 Patch autogenerated by refactor_symtab.py from
1216 https://github.com/davidmalcolm/gcc-refactoring-scripts
1217 revision 58bb219cc090b2f4516a9297d868c245495ee622
1219 * asan.c (asan_finish_file): Update for conversion of symtab types to
1220 a true class hierarchy.
1221 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1222 * cgraph.c (cgraph_get_body): Likewise.
1223 (cgraph_get_create_real_symbol_node): Likewise.
1224 (verify_cgraph_node): Likewise.
1225 (verify_edge_corresponds_to_fndecl): Likewise.
1226 (verify_edge_count_and_frequency): Likewise.
1227 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1228 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1229 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1230 (cgraph_node_cannot_return): Likewise.
1231 (cgraph_set_pure_flag_1): Likewise.
1232 (cgraph_set_const_flag_1): Likewise.
1233 (cgraph_set_nothrow_flag_1): Likewise.
1234 (cgraph_make_node_local_1): Likewise.
1235 (cgraph_for_node_and_aliases): Likewise.
1236 (cgraph_for_node_thunks_and_aliases): Likewise.
1237 (cgraph_node_can_be_local_p): Likewise.
1238 (cgraph_node_cannot_be_local_p_1): Likewise.
1239 (cgraph_function_body_availability): Likewise.
1240 (dump_cgraph_node): Likewise.
1241 (cgraph_rtl_info): Likewise.
1242 (cgraph_mark_address_taken_node): Likewise.
1243 (cgraph_remove_node): Likewise.
1244 (cgraph_release_function_body): Likewise.
1245 (cgraph_update_edges_for_call_stmt_node): Likewise.
1246 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1247 (cgraph_make_edge_direct): Likewise.
1248 (cgraph_resolve_speculation): Likewise.
1249 (cgraph_speculative_call_info): Likewise.
1250 (cgraph_turn_edge_to_speculative): Likewise.
1251 (cgraph_create_edge_1): Likewise.
1252 (cgraph_set_call_stmt): Likewise.
1253 (cgraph_node_for_asm): Likewise.
1254 (cgraph_add_thunk): Likewise.
1255 (cgraph_same_body_alias): Likewise.
1256 (cgraph_create_function_alias): Likewise.
1257 (cgraph_create_node): Likewise.
1258 (cgraph_create_empty_node): Likewise.
1259 (record_function_versions): Likewise.
1260 (used_from_object_file_p): Likewise.
1261 * cgraph.h (symtab_can_be_discarded): Likewise.
1262 (symtab_real_symbol_p): Likewise.
1263 (cgraph_mark_force_output_node): Likewise.
1264 (cgraph_edge_recursive_p): Likewise.
1265 (symtab_alias_target): Likewise.
1266 (varpool_all_refs_explicit_p): Likewise.
1267 (varpool_can_remove_if_no_refs): Likewise.
1268 (cgraph_only_called_directly_or_aliased_p): Likewise.
1269 (cgraph_next_function_with_gimple_body): Likewise.
1270 (cgraph_first_function_with_gimple_body): Likewise.
1271 (cgraph_function_with_gimple_body_p): Likewise.
1272 (cgraph_next_function): Likewise.
1273 (cgraph_first_function): Likewise.
1274 (cgraph_next_defined_function): Likewise.
1275 (cgraph_first_defined_function): Likewise.
1276 (varpool_next_defined_variable): Likewise.
1277 (varpool_first_defined_variable): Likewise.
1278 (varpool_next_static_initializer): Likewise.
1279 (varpool_first_static_initializer): Likewise.
1280 (varpool_next_variable): Likewise.
1281 (varpool_first_variable): Likewise.
1282 (varpool_node_name): Likewise.
1283 (varpool): Likewise.
1285 (is_a_helper <varpool_node>::test): Likewise.
1286 (is_a_helper <cgraph_node>::test): Likewise.
1287 (varpool_variable_node): Likewise.
1288 (cgraph_function_or_thunk_node): Likewise.
1289 (varpool_alias_target): Likewise.
1290 (cgraph_alias_target): Likewise.
1291 (cgraph_node_name): Likewise.
1292 (varpool_node_asm_name): Likewise.
1293 (cgraph_node_asm_name): Likewise.
1294 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1295 (cgraph_rebuild_references): Likewise.
1296 (rebuild_cgraph_edges): Likewise.
1297 (record_eh_tables): Likewise.
1298 (build_cgraph_edges): Likewise.
1299 (mark_store): Likewise.
1300 (mark_load): Likewise.
1301 (mark_address): Likewise.
1302 (record_type_list): Likewise.
1303 (record_reference): Likewise.
1304 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1305 (cgraph_materialize_clone): Likewise.
1306 (cgraph_function_versioning): Likewise.
1307 (cgraph_copy_node_for_versioning): Likewise.
1308 (update_call_expr): Likewise.
1309 (cgraph_find_replacement_node): Likewise.
1310 (cgraph_create_virtual_clone): Likewise.
1311 (cgraph_clone_node): Likewise.
1312 * cgraphunit.c (compile): Likewise.
1313 (output_weakrefs): Likewise.
1314 (output_in_order): Likewise.
1315 (expand_function): Likewise.
1316 (assemble_thunks_and_aliases): Likewise.
1317 (expand_thunk): Likewise.
1318 (mark_functions_to_output): Likewise.
1319 (handle_alias_pairs): Likewise.
1320 (analyze_functions): Likewise.
1321 (walk_polymorphic_call_targets): Likewise.
1322 (varpool_finalize_decl): Likewise.
1323 (process_function_and_variable_attributes): Likewise.
1324 (cgraph_process_same_body_aliases): Likewise.
1325 (analyze_function): Likewise.
1326 (cgraph_add_new_function): Likewise.
1327 (cgraph_finalize_function): Likewise.
1328 (referred_to_p): Likewise.
1329 (cgraph_reset_node): Likewise.
1330 (cgraph_process_new_functions): Likewise.
1331 (enqueue_node): Likewise.
1332 (decide_is_symbol_needed): Likewise.
1333 * coverage.c (coverage_compute_profile_id): Likewise.
1334 * dbxout.c (dbxout_expand_expr): Likewise.
1335 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1336 (reference_to_unused): Likewise.
1337 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1338 * gimplify.c (unvisit_body): Likewise.
1339 (unshare_body): Likewise.
1340 * ipa-cp.c (ipcp_generate_summary): Likewise.
1341 (ipcp_decision_stage): Likewise.
1342 (identify_dead_nodes): Likewise.
1343 (decide_whether_version_node): Likewise.
1344 (decide_about_value): Likewise.
1345 (perhaps_add_new_callers): Likewise.
1346 (create_specialized_node): Likewise.
1347 (update_profiling_info): Likewise.
1348 (ipcp_propagate_stage): Likewise.
1349 (estimate_local_effects): Likewise.
1350 (good_cloning_opportunity_p): Likewise.
1351 (devirtualization_time_bonus): Likewise.
1352 (propagate_constants_accross_call): Likewise.
1353 (initialize_node_lattices): Likewise.
1354 (ipcp_cloning_candidate_p): Likewise.
1355 (determine_versionability): Likewise.
1356 (print_all_lattices): Likewise.
1357 (print_lattice): Likewise.
1358 (ipcp_discover_new_direct_edges): Likewise.
1359 * ipa-devirt.c (ipa_devirt): Likewise.
1360 (likely_target_p): Likewise.
1361 (update_type_inheritance_graph): Likewise.
1362 (possible_polymorphic_call_target_p): Likewise.
1363 (dump_possible_polymorphic_call_targets): Likewise.
1364 (devirt_variable_node_removal_hook): Likewise.
1365 (record_binfo): Likewise.
1366 (maybe_record_node): Likewise.
1367 (build_type_inheritance_graph): Likewise.
1368 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1369 (inline_generate_summary): Likewise.
1370 (inline_analyze_function): Likewise.
1371 (do_estimate_growth): Likewise.
1372 (simple_edge_hints): Likewise.
1373 (estimate_node_size_and_time): Likewise.
1374 (estimate_edge_devirt_benefit): Likewise.
1375 (compute_inline_parameters): Likewise.
1376 (estimate_function_body_sizes): Likewise.
1377 (compute_bb_predicates): Likewise.
1378 (initialize_inline_failed): Likewise.
1379 (dump_inline_summary): Likewise.
1380 (dump_inline_edge_summary): Likewise.
1381 * ipa-inline-transform.c (inline_transform): Likewise.
1382 (preserve_function_body_p): Likewise.
1383 (save_inline_function_body): Likewise.
1384 (inline_call): Likewise.
1385 (clone_inlined_nodes): Likewise.
1386 (can_remove_node_now_p): Likewise.
1387 (can_remove_node_now_p_1): Likewise.
1388 * ipa-inline.c (early_inliner): Likewise.
1389 (early_inline_small_functions): Likewise.
1390 (inline_always_inline_functions): Likewise.
1391 (ipa_inline): Likewise.
1392 (flatten_function): Likewise.
1393 (inline_small_functions): Likewise.
1394 (speculation_useful_p): Likewise.
1395 (recursive_inlining): Likewise.
1396 (update_caller_keys): Likewise.
1397 (reset_edge_caches): Likewise.
1398 (update_edge_key): Likewise.
1399 (edge_badness): Likewise.
1400 (relative_time_benefit): Likewise.
1401 (want_inline_self_recursive_call_p): Likewise.
1402 (want_inline_small_function_p): Likewise.
1403 (want_early_inline_function_p): Likewise.
1404 (num_calls): Likewise.
1405 (can_early_inline_edge_p): Likewise.
1406 (can_inline_edge_p): Likewise.
1407 (report_inline_failed_reason): Likewise.
1408 * ipa-profile.c (ipa_profile): Likewise.
1409 (ipa_propagate_frequency): Likewise.
1410 (ipa_propagate_frequency_1): Likewise.
1411 (ipa_profile_generate_summary): Likewise.
1412 * ipa-prop.c (ipcp_transform_function): Likewise.
1413 (read_replacements_section): Likewise.
1414 (ipa_prop_read_section): Likewise.
1415 (ipa_modify_call_arguments): Likewise.
1416 (ipa_print_node_params): Likewise.
1417 (propagate_controlled_uses): Likewise.
1418 (update_indirect_edges_after_inlining): Likewise.
1419 (remove_described_reference): Likewise.
1420 (ipa_make_edge_direct_to_target): Likewise.
1421 (ipa_analyze_node): Likewise.
1422 (ipa_analyze_params_uses): Likewise.
1423 (ipa_compute_jump_functions): Likewise.
1424 (ipa_get_callee_param_type): Likewise.
1425 (ipa_print_node_jump_functions): Likewise.
1426 (ipa_initialize_node_params): Likewise.
1427 (ipa_populate_param_decls): Likewise.
1428 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1429 (write_agg_replacement_chain): Likewise.
1430 (ipa_write_node_info): Likewise.
1431 (ipa_edge_duplication_hook): Likewise.
1432 (try_decrement_rdesc_refcount): Likewise.
1433 * ipa-pure-const.c (propagate_nothrow): Likewise.
1434 (propagate_pure_const): Likewise.
1435 (pure_const_read_summary): Likewise.
1436 (pure_const_write_summary): Likewise.
1437 (analyze_function): Likewise.
1438 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1439 (ipa_ref_referring_ref_list): Likewise.
1440 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1441 (ipa_remove_stmt_references): Likewise.
1442 (ipa_find_reference): Likewise.
1443 (ipa_dump_referring): Likewise.
1444 (ipa_dump_references): Likewise.
1445 (ipa_record_reference): Likewise.
1446 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1447 (ipa_reference_write_optimization_summary): Likewise.
1448 (write_node_summary_p): Likewise.
1449 (propagate): Likewise.
1450 (read_write_all_from_decl): Likewise.
1451 (generate_summary): Likewise.
1452 (analyze_function): Likewise.
1453 (propagate_bits): Likewise.
1454 (ipa_reference_get_not_written_global): Likewise.
1455 (ipa_reference_get_not_read_global): Likewise.
1456 * ipa-split.c (execute_split_functions): Likewise.
1457 (split_function): Likewise.
1458 * ipa-utils.c (ipa_merge_profiles): Likewise.
1459 (dump_cgraph_node_set): Likewise.
1460 (ipa_reverse_postorder): Likewise.
1461 (ipa_edge_within_scc): Likewise.
1462 (ipa_get_nodes_in_cycle): Likewise.
1463 (ipa_free_postorder_info): Likewise.
1464 (ipa_reduced_postorder): Likewise.
1465 (searchc): Likewise.
1466 (recursive_call_p): Likewise.
1467 * ipa.c (ipa_cdtor_merge): Likewise.
1468 (record_cdtor_fn): Likewise.
1469 (function_and_variable_visibility): Likewise.
1470 (varpool_externally_visible_p): Likewise.
1471 (cgraph_externally_visible_p): Likewise.
1472 (comdat_can_be_unshared_p): Likewise.
1473 (comdat_can_be_unshared_p_1): Likewise.
1474 (address_taken_from_non_vtable_p): Likewise.
1475 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1476 (symtab_remove_unreachable_nodes): Likewise.
1477 (walk_polymorphic_call_targets): Likewise.
1478 (process_references): Likewise.
1479 (enqueue_node): Likewise.
1480 (has_addr_references_p): Likewise.
1481 (cgraph_non_local_node_p_1): Likewise.
1482 * is-a.h (varpool_analyze_node): Likewise.
1483 * lto-cgraph.c (input_symtab): Likewise.
1484 (merge_profile_summaries): Likewise.
1485 (input_cgraph_1): Likewise.
1486 (input_edge): Likewise.
1487 (input_varpool_node): Likewise.
1488 (input_node): Likewise.
1489 (input_overwrite_node): Likewise.
1490 (compute_ltrans_boundary): Likewise.
1491 (output_refs): Likewise.
1492 (lto_output_varpool_node): Likewise.
1493 (lto_output_node): Likewise.
1494 (reachable_from_other_partition_p): Likewise.
1495 (referenced_from_other_partition_p): Likewise.
1496 (lto_output_edge): Likewise.
1497 (output_node_opt_summary): Likewise.
1498 (add_node_to): Likewise.
1499 (reachable_from_this_partition_p): Likewise.
1500 (lto_set_symtab_encoder_in_partition): Likewise.
1501 (lto_symtab_encoder_in_partition_p): Likewise.
1502 (lto_set_symtab_encoder_encode_initializer): Likewise.
1503 (lto_symtab_encoder_encode_initializer_p): Likewise.
1504 (lto_set_symtab_encoder_encode_body): Likewise.
1505 (lto_symtab_encoder_encode_body_p): Likewise.
1506 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1508 * lto-streamer-in.c (lto_read_body): Likewise.
1509 (fixup_call_stmt_edges): Likewise.
1510 (fixup_call_stmt_edges_1): Likewise.
1511 * lto-streamer-out.c (produce_symtab): Likewise.
1512 (output_symbol_p): Likewise.
1513 (write_symbol): Likewise.
1514 (lto_output): Likewise.
1515 (copy_function): Likewise.
1516 (output_function): Likewise.
1517 * passes.c (function_called_by_processed_nodes_p): Likewise.
1518 (ipa_write_optimization_summaries): Likewise.
1519 (ipa_write_summaries): Likewise.
1520 (do_per_function_toporder): Likewise.
1521 (do_per_function): Likewise.
1522 (dump_passes): Likewise.
1523 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1524 (symtab_nonoverwritable_alias): Likewise.
1525 (symtab_nonoverwritable_alias_1): Likewise.
1526 (symtab_for_node_and_aliases): Likewise.
1527 (symtab_resolve_alias): Likewise.
1528 (fixup_same_cpp_alias_visibility): Likewise.
1529 (symtab_alias_ultimate_target): Likewise.
1530 (symtab_used_from_object_file_p): Likewise.
1531 (verify_symtab_base): Likewise.
1532 (dump_symtab_base): Likewise.
1533 (symtab_node_name): Likewise.
1534 (symtab_node_asm_name): Likewise.
1535 (symtab_dissolve_same_comdat_group_list): Likewise.
1536 (symtab_add_to_same_comdat_group): Likewise.
1537 (symtab_unregister_node): Likewise.
1538 (symtab_insert_node_to_hashtable): Likewise.
1539 (symtab_register_node): Likewise.
1540 (unlink_from_assembler_name_hash): Likewise.
1541 (insert_to_assembler_name_hash): Likewise.
1542 (eq_assembler_name): Likewise.
1543 (hash_node_by_assembler_name): Likewise.
1544 (eq_node): Likewise.
1545 (hash_node): Likewise.
1546 * toplev.c (wrapup_global_declaration_2): Likewise.
1547 * trans-mem.c (ipa_tm_execute): Likewise.
1548 (ipa_tm_transform_clone): Likewise.
1549 (ipa_tm_transform_transaction): Likewise.
1550 (ipa_tm_transform_calls_redirect): Likewise.
1551 (ipa_tm_insert_gettmclone_call): Likewise.
1552 (ipa_tm_insert_irr_call): Likewise.
1553 (ipa_tm_create_version): Likewise.
1554 (ipa_tm_create_version_alias): Likewise.
1555 (ipa_tm_mark_forced_by_abi_node): Likewise.
1556 (ipa_tm_mark_force_output_node): Likewise.
1557 (ipa_tm_diagnose_tm_safe): Likewise.
1558 (ipa_tm_mayenterirr_function): Likewise.
1559 (ipa_tm_scan_irr_function): Likewise.
1560 (ipa_tm_note_irrevocable): Likewise.
1561 (ipa_tm_scan_calls_clone): Likewise.
1562 (get_cg_data): Likewise.
1563 * tree-eh.c (tree_could_trap_p): Likewise.
1564 * tree-emutls.c (ipa_lower_emutls): Likewise.
1565 (create_emultls_var): Likewise.
1566 (lower_emutls_function_body): Likewise.
1567 (gen_emutls_addr): Likewise.
1568 (emutls_decl): Likewise.
1569 (new_emutls_decl): Likewise.
1570 * tree-inline.c (tree_function_versioning): Likewise.
1571 (optimize_inline_calls): Likewise.
1572 (expand_call_inline): Likewise.
1573 (estimate_num_insns): Likewise.
1574 (copy_bb): Likewise.
1575 (delete_unreachable_blocks_update_callgraph): Likewise.
1576 * tree-nested.c (gimplify_all_functions): Likewise.
1577 (create_nesting_tree): Likewise.
1578 (check_for_nested_with_variably_modified): Likewise.
1579 * tree-pretty-print.c (dump_function_header): Likewise.
1580 * tree-profile.c (tree_profiling): Likewise.
1581 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1582 (modify_function): Likewise.
1583 (convert_callers): Likewise.
1584 (convert_callers_for_node): Likewise.
1585 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1586 (associate_varinfo_to_alias): Likewise.
1587 (create_variable_info_for): Likewise.
1588 (get_constraint_for_ssa_var): Likewise.
1589 * tree-vectorizer.c (increase_alignment): Likewise.
1590 * tree.c (find_decls_types_in_var): Likewise.
1591 (find_decls_types_in_node): Likewise.
1592 (free_lang_data_in_decl): Likewise.
1593 * value-prof.c (gimple_ic_transform): Likewise.
1594 (gimple_ic): Likewise.
1595 (check_ic_target): Likewise.
1596 (init_node_map): Likewise.
1597 * varasm.c (decl_binds_to_current_def_p): Likewise.
1598 (default_binds_local_p_1): Likewise.
1599 (dump_tm_clone_pairs): Likewise.
1600 (assemble_alias): Likewise.
1601 (find_decl): Likewise.
1602 (mark_decl_referenced): Likewise.
1603 * varpool.c (varpool_for_node_and_aliases): Likewise.
1604 (varpool_extra_name_alias): Likewise.
1605 (varpool_create_variable_alias): Likewise.
1606 (add_new_static_var): Likewise.
1607 (varpool_finalize_named_section_flags): Likewise.
1608 (varpool_remove_unreferenced_decls): Likewise.
1609 (enqueue_node): Likewise.
1610 (varpool_assemble_decl): Likewise.
1611 (assemble_aliases): Likewise.
1612 (varpool_analyze_node): Likewise.
1613 (cgraph_variable_initializer_availability): Likewise.
1614 (varpool_add_new_variable): Likewise.
1615 (ctor_for_folding): Likewise.
1616 (dump_varpool_node): Likewise.
1617 (varpool_remove_initializer): Likewise.
1618 (varpool_remove_node): Likewise.
1619 (varpool_node_for_decl): Likewise.
1620 (varpool_create_empty_node): Likewise.
1621 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1622 (ix86_get_function_versions_dispatcher): Likewise.
1624 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1626 * cgraph.h (symtab_node_base): Convert to a class;
1627 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1628 chain_next/prev from symtab_node_def.
1629 (cgraph_node): Inherit from symtab_node; add GTY option
1630 tag ("SYMTAB_FUNCTION").
1631 (varpool_node): Inherit from symtab_node; add GTY option
1632 tag ("SYMTAB_VARIABLE").
1633 (symtab_node_def): Remove.
1634 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
1635 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
1636 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
1637 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
1639 * ipa-ref.h (symtab_node_def): Drop.
1640 (symtab_node): Change underlying type from symtab_node_def to
1642 (const_symtab_node): Likwise.
1644 * is-a.h: Update examples in comment.
1646 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
1647 (assembler_name_hash): Likewise.
1649 2013-10-29 Martin Liska <marxin.liska@gmail.com>
1651 * doc/tree-ssa.texi (gimple_phi_result): Document.
1652 (gimple_phi_num_args, gimple_phi_arg): Likewise.
1653 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
1654 (PHI_RESULT, PHI_NUM_ARGS): Remove.
1655 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
1657 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1659 * expr.h: Revert change and include tree-core.h.
1660 * rtl.h: Revert change and don't include tree-core.h.
1662 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1664 * config/darwin.c: Include gimple.h.
1665 * config/i386/winnt.c: Likewise.
1667 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
1669 PR tree-optimization/19831
1670 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
1672 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1674 * tree-outof-ssa.h: Remove include files.
1675 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
1677 * tree-ssa-coalesce.c: Likewise.
1678 * cfgexpand.c: Likewise.
1679 * tree-ssa-ter.c: Likewise.
1680 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
1681 * lto-streamer.h: Likewise.
1682 * cgraphbuild.c: Add gimple.h to include list.
1683 * data-streamer-in.c: Likewise.
1684 * ipa-cp.c: Likewise.
1685 * tree-streamer.c: Likewise.
1686 * lto-compress.c: Likewise.
1687 * ipa-reference.c: Likewise.
1688 * data-streamer-out.c: Likewise.
1689 * lto-cgraph.c: Likewise.
1690 * cgraphclones.c: Likewise.
1691 * ipa-utils.c: Likewise.
1692 * data-streamer.c: Likewise.
1693 * ipa-split.c: Likewise.
1694 * lto-section-in.c: Likewise.
1695 * tree-streamer-out.c: Likewise.
1696 * ipa-prop.c: Likewise.
1697 * tree-streamer-in.c: Likewise.
1698 * symtab.c: Likewise.
1699 * opts-global.c: Likewise.
1700 * lto-opts.c: Likewise.
1701 * lto-section-out.c: Likewise.
1702 * lto-streamer.c: Likewise.
1703 * rtl.h: Add tree-core.h to include list.
1704 * expr.h: Remove tree-core.h from include list.
1705 * gimple.h: Likewise.
1706 * ipa-utils.h: Likewise.
1707 * streamer-hooks.h: Likewise.
1708 * streamer-hooks.c: Include input.h.
1710 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1712 * config/arm/arm.c (cortexa7_extra_costs): New table.
1713 (arm_cortex_a7_tune): New.
1714 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
1716 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
1718 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
1720 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1722 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
1723 to use autogenerated markers for a class-hierarchy, every class
1724 must have a GTY marker.
1725 * gengtype.h (struct type): Add linked list of subclasses to
1726 the "s" member of the union.
1727 (add_subclass): New decl.
1728 * gengtype-state.c (read_state_struct_type): Set up subclass
1730 * gengtype.c (get_ultimate_base_class): New.
1731 (add_subclass): New.
1732 (new_structure): Set up subclass linked list.
1733 (set_gc_used_type): Propagate usage information to subclasses.
1734 (output_mangled_typename): Use get_ultimate_base_class.
1735 (walk_subclasses): Use the subclass linked list, avoiding an
1736 O(N^2) when writing out all types.
1737 (walk_type): Issue an error if the base class is missing a tag,
1738 rather than generating bogus C code. Add a gcc_unreachable
1739 default case, in case people omit tags from concrete subclasses,
1740 or get the values wrong.
1741 (write_func_for_structure): Issue an error for subclasses for
1742 which the base doesn't have a "desc", since otherwise the
1743 autogenerated routines for the base would silently fail to visit
1744 any subclass fields.
1745 (write_root): Use get_ultimate_base_class, tweaking constness of
1746 tp to match that function's signature.
1748 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1750 * doc/gty.texi (GTY Options): Add note about inheritance to
1751 description of desc and tag.
1752 (Inheritance and GTY): New.
1754 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1756 * gengtype-parse.c (opts_have): Drop "static" so that
1757 we can use this from gengtype.c.
1758 * gengtype.c (set_gc_used_type): Mark any base class as used;
1759 update field traversal to visit inherited fields.
1760 (output_mangled_typename): Convert references to classes within
1761 an inheritance hierarchy to reference the ultimate base class,
1762 since only it will have gt_ functions.
1763 (get_string_option): New.
1764 (walk_subclasses): New.
1765 (walk_type): Treat GTY structs that have a "desc" as being the
1766 root of an inheritance hierarchy. Generate a switch on it
1767 within the marking function which walks all subclasses, adding
1768 cases for them via walk_subclasses. For subclasses, visit all
1769 fields of the type (including inherited ones).
1770 (write_func_for_structure): Don't write fns for subclasses, only
1771 for the ultimate base class within an inheritance hierarchy.
1772 Subclasses-marking will be handled by the base class marking functions.
1773 (write_types): Likewise.
1774 (write_local_func_for_structure): Likewise.
1775 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
1776 a "tag" option (and are thus concrete subclasses).
1777 (write_root): Use the marker function for the ultimate base class.
1778 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
1779 (opts_have): Add declaration.
1781 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
1783 * lra-spills.c (lra_final_code_change): Remove useless move insns
1784 originated from moves of pseudos.
1786 2013-10-28 Jeff Law <law@redhat.com>
1788 * lower-subreg.c (resolve_simple_move): Fix comment typo.
1790 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
1792 * df-scan.c (df_collection_rec): Adjust.
1793 (copy_defs): New constant.
1794 (copy_uses): Likewise.
1795 (copy_eq_uses): Likewise.
1796 (copy_mw): Likewise.
1797 (copy_all): Likewise.
1798 (df_insn_rescan): Adjust.
1799 (df_notes_rescan): Likewise.
1800 (df_swap_refs): Likewise.
1801 (df_sort_and_compress_refs): Likewise.
1802 (df_sort_and_compress_mws): Likewise.
1803 (df_install_refs): Likewise.
1804 (df_install_mws): Likewise.
1805 (df_refs_add_to_chains): Add flags parameter controlling which vectors
1807 (df_bb_refs_record): Adjust.
1808 (df_record_entry_block_defs): Likewise.
1809 (df_record_exit_block_defs): Likewise.
1810 (df_refs_verify): Likewise.
1811 (df_mws_verify): Likewise.
1812 (df_insn_refs_verify): Likewise.
1813 (df_bb_verify): Likewise.
1814 * ipa-pure-const.c (finish_state): Remove.
1815 (propagate): Adjust.
1816 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
1817 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
1818 var-tracking.c: Adjust.
1819 * vec.c (stack_vecs): Remove.
1820 (register_stack_vec): Likewise.
1821 (stack_vec_register_index): Likewise.
1822 (unregister_stack_vec): Likewise.
1823 * vec.h (struct va_stack): Remove.
1824 (struct vec<T, A, vl_ptr>): Specialize as
1825 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1826 allocation strategy compatable with the vl_ptr layout.
1827 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1828 specialization anyway.
1829 (class stack_vec): New class.
1830 (vec_stack_alloc): Remove.
1831 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1833 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1834 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1835 Sergey Lega <sergey.s.lega@intel.com>
1836 Anna Tikhonova <anna.tikhonova@intel.com>
1837 Ilya Tocar <ilya.tocar@intel.com>
1838 Andrey Turetskiy <andrey.turetskiy@intel.com>
1839 Ilya Verbin <ilya.verbin@intel.com>
1840 Kirill Yukhin <kirill.yukhin@intel.com>
1841 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1843 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1844 (*prefetch_avx512pf_<mode>): New.
1845 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1846 (avx512f_maskcmp<mode>3): Ditto.
1847 (vashrv16si3): Ditto.
1849 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1850 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1851 Sergey Lega <sergey.s.lega@intel.com>
1852 Anna Tikhonova <anna.tikhonova@intel.com>
1853 Ilya Tocar <ilya.tocar@intel.com>
1854 Andrey Turetskiy <andrey.turetskiy@intel.com>
1855 Ilya Verbin <ilya.verbin@intel.com>
1856 Kirill Yukhin <kirill.yukhin@intel.com>
1857 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1859 * config/i386/i386.md (any_truncate): New.
1860 (trunsuffix): Ditto.
1861 * config/i386/predicates.md (const_8_to_9_operand): New.
1862 (const_10_to_11_operand): Ditto.
1863 (const_12_to_13_operand): Ditto.
1864 (const_14_to_15_operand): Ditto.
1865 (const_16_to_19_operand): Ditto.
1866 (const_20_to_23_operand): Ditto.
1867 (const_24_to_27_operand): Ditto.
1868 (const_28_to_31_operand): Ditto.
1869 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
1870 (cvtusi2<ssescalarmodesuffix>32): New.
1871 (cvtusi2<ssescalarmodesuffix>64): Ditto.
1872 (ufloatv16siv16sf2): Ditto.
1873 (avx512f_fix_notruncv16sfv16si): Ditto.
1874 (avx512f_ufix_notruncv16sfv16si): Ditto.
1875 (avx512f_vcvtss2usi): Ditto.
1876 (avx512f_vcvtss2usiq): Ditto.
1877 (avx512f_vcvttss2usi): Ditto.
1878 (avx512f_vcvttss2usiq): Ditto.
1879 (avx512f_vcvtsd2usi): Ditto.
1880 (avx512f_vcvtsd2usiq): Ditto.
1881 (avx512f_vcvttsd2usi): Ditto.
1882 (avx512f_vcvttsd2usiq): Ditto.
1883 (ufloatv8siv8df): Ditto.
1884 (avx512f_cvtdq2pd512_2): Ditto.
1885 (avx512f_cvtpd2dq512): Ditto.
1886 (avx512f_ufix_notruncv8dfv8si): Ditto.
1887 (avx512f_cvtpd2ps512): Ditto.
1888 (vec_unpacks_lo_v16sf): Ditto.
1889 (vec_unpacks_hi_v16sf): Ditto.
1890 (vec_unpacks_float_hi_v16si): Ditto.
1891 (vec_unpacks_float_lo_v16si): Ditto.
1892 (avx512f_unpckhps512): Ditto.
1893 (avx512f_unpcklps512): Ditto.
1894 (avx512f_movshdup512): Ditto.
1895 (avx512f_movsldup512): Ditto.
1896 (vec_extract_lo_v32hi): Ditto.
1897 (vec_extract_hi_v32hi): Ditto.
1898 (vec_extract_lo_v64qi): Ditto.
1899 (vec_extract_hi_v64qi): Ditto.
1900 (avx512f_unpckhpd512): Ditto.
1901 (avx512f_movddup512): Ditto.
1902 (avx512f_unpcklpd512): Ditto.
1903 (*avx512f_unpcklpd512): Ditto.
1904 (avx512f_shufps512_1): Ditto.
1905 (avx512f_shufpd512_1): Ditto.
1906 (avx512f_interleave_highv8di): Ditto.
1907 (avx512f_interleave_lowv8di): Ditto.
1908 (PMOV_DST_MODE): Ditto.
1909 (pmov_src_mode): Ditto.
1910 (pmov_src_lower): Ditto.
1912 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
1913 (*avx512f_<code>v8div16qi2): Ditto.
1914 (*avx512f_<code>v8div16qi2_store): Ditto.
1915 (vec_widen_umult_even_v16si): Ditto.
1916 (*vec_widen_umult_even_v16si): Ditto.
1917 (vec_widen_smult_even_v16si): Ditto.
1918 (*vec_widen_smult_even_v16si): Ditto.
1919 (avx512f_interleave_highv16si): Ditto.
1920 (avx512f_interleave_lowv16si): Ditto.
1921 (avx512f_<code>v16qiv16si2): Ditto.
1922 (avx512f_<code>v16hiv16si2): Ditto.
1923 (avx512f_<code>v8qiv8di2): Ditto.
1924 (avx512f_<code>v8hiv8di2): Ditto.
1925 (avx512f_<code>v8siv8di2): Ditto.
1926 (avx512cd_maskb_vec_dupv8di): Ditto.
1927 (avx512cd_maskw_vec_dupv16si): Ditto.
1928 (avx512f_vcvtph2ps512): Ditto.
1929 (avx512f_vcvtps2ph512): Ditto.
1930 (VEC_EXTRACT_MODE): Extened with wider modes.
1931 (VEC_PERM_AVX2): Ditto.
1932 (VEC_PERM_CONST): Ditto.
1934 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
1936 * config/arc/arc.c (arc_ccfsm_post_advance):
1937 Add comment about TYPE_RETURN.
1939 2013-10-28 Bin Cheng <bin.cheng@arm.com>
1941 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
1942 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
1943 (strip_offset): Convert offset to unsigned number.
1945 2013-10-27 Tom de Vries <tom@codesourcery.com>
1947 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
1948 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
1950 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
1952 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
1953 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
1954 (TARGET_BIG_ENDIAN): New macro.
1955 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
1956 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
1957 * config/sh/sh.c: Likewise.
1958 * config/sh/sh.h: Likewise.
1960 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
1962 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
1963 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
1965 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1967 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
1969 (cmpstr_t, cmpstrsi): Fix formatting.
1971 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1974 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
1975 addressing, do not use general_operand for memory operands.
1977 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
1980 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1981 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1983 2013-10-26 Jeff Law <law@redhat.com>
1985 * predict.c (PRED_MUDFLAP): Remove.
1986 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
1988 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
1989 (OBJS): Remove tree-nomudflap.o
1990 (GTFILES): Remove tree-mudflap.c
1991 * builtins.c (expand_builtin_alloc): Remove mudflap support.
1992 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1993 (mfwrap_spec, mflib_spec): Likewise.
1994 (cpp_unique_options, cc1_options, static_specs): Likewise.
1995 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
1996 * passes.def: Likewise.
1997 * toplev.c (compile_file, process_options): Likewise.
1998 * tree-inline.c (copy_tree_r): Likewise.
1999 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
2000 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
2001 (build_constant_desc, output_constant_def_contents): Likewise.
2002 (categorize_decl_for_section): Likewise.
2003 * tree-mudflap.c: Removed.
2004 * tree-mudflap.h: Removed.
2005 * tree-nomudflap.c: Removed.
2006 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
2007 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
2008 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2009 * config/sol2.h (MFLIB_SPEC): Likewise.
2010 * doc/install.texi: Remove mudflap references.
2011 * doc/passes.texi: Similarly.
2012 * doc/sourcebuild.texi: Similarly.
2013 * doc/invoke.texi: Remove mudlfap related options.
2015 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2017 PR rtl-optimization/58759
2018 * lra-constraints.c (lra_constraints): Remove wrong condition to
2019 remove insn setting up an equivalent pseudo.
2021 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2023 * config/rs6000/rs6000-protos.h
2024 (rs6000_secondary_memory_needed_mode): New prototype.
2025 * config/rs6000/rs6000.c: Include ira.h.
2026 (TARGET_LRA_P): Redefine.
2027 (rs6000_legitimate_offset_address_p): Call
2028 legitimate_constant_pool_address_p in strict mode for LRA.
2029 (rs6000_legitimate_address_p): Ditto.
2030 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2031 (rs6000_emit_move): Add LRA version of code to generate load/store
2033 (rs6000_secondary_memory_needed_mode): New.
2034 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2035 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2036 constants, memory, and FP registers.
2037 (rs6000_lra_p): New.
2038 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2039 * config/rs6000/rs6000.opt (mlra): New option.
2040 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2042 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2044 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2045 has_single_use () and not do the conversion if has_single_use ()
2046 returns false for the multiplication result.
2048 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2050 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2051 to "NODE", since this works on a "tree", not an
2053 (CONSTANT_CLASS_P): Likewise.
2056 (INDIRECT_REF_P): Likewise.
2057 (REFERENCE_CLASS_P): Likewise.
2058 (COMPARISON_CLASS_P): Likewise.
2059 (UNARY_CLASS_P): Likewise.
2060 (BINARY_CLASS_P): Likewise.
2061 (STATEMENT_CLASS_P): Likewise.
2062 (VL_EXP_CLASS_P): Likewise.
2063 (EXPRESSION_CLASS_P): Likewise.
2064 (IS_TYPE_OR_DECL_P): Likewise.
2066 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2068 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2069 ADDR_EXPR in the defining statement.
2071 2013-10-25 Richard Biener <rguenther@suse.de>
2073 PR tree-optimization/58626
2074 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2075 anti_dd, output_dd and input_dd.
2076 (struct rdg_edge): Remove level and relation members.
2077 (RDGE_LEVEL, RDGE_RELATION): Remove.
2078 (dot_rdg_1): Adjust.
2079 (create_rdg_edge_for_ddr): Remove.
2080 (create_rdg_edges_for_scalar): Adjust.
2081 (create_edge_for_control_dependence): Likewise.
2082 (create_rdg_edges): Split into ...
2083 (create_rdg_flow_edges): ... this
2084 (create_rdg_cd_edges): ... and this.
2086 (build_rdg): Likewise, do not compute data dependences or
2088 (pg_add_dependence_edges): New function.
2090 (distribute_loop): First apply all non-dependence based
2091 partition mergings. Then compute dependences between partitions
2092 and merge and order partitions according to them.
2094 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2096 PR rtl-optimization/58831
2097 * alias.c (init_alias_analysis): At the beginning of each iteration,
2098 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2100 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2102 * recog.c (search_ofs): New static variable moved from...
2103 (peep2_find_free_register): ...here.
2104 (peephole2_optimize): Initialize it.
2106 2013-10-25 Tobias Burnus <burnus@net-b.de>
2108 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2110 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2112 * config/i386/i386.h (TARGET_MPX): New define.
2113 (TARGET_MPX_P): Ditto.
2115 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2117 * config/i386/constraints.md (B): New.
2120 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2121 * config/i386/i386-modes.def (BND32): New.
2123 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2124 * config/i386/i386.c (isa_opts): Add mmpx.
2125 (regclass_map): Add bound registers.
2126 (dbx_register_map): Likewise.
2127 (dbx64_register_map): Likewise.
2128 (svr4_dbx_register_map): Likewise.
2130 (ix86_option_override_internal): Support MPX ISA.
2131 (ix86_conditional_register_usage): Support bound registers.
2132 (print_reg): Likewise.
2133 (ix86_code_end): Add MPX bnd prefix.
2134 (output_set_got): Likewise.
2135 (ix86_output_call_insn): Likewise.
2136 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2137 (ix86_print_operand_punct_valid_p): Likewise.
2138 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2140 (ix86_class_likely_spilled_p): Add bound regs support.
2141 (ix86_hard_regno_mode_ok): Likewise.
2142 (x86_order_regs_for_local_alloc): Likewise.
2143 (ix86_bnd_prefixed_insn_p): New.
2144 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2145 (FIXED_REGISTERS): Add bound registers.
2146 (CALL_USED_REGISTERS): Likewise.
2147 (REG_ALLOC_ORDER): Likewise.
2148 (HARD_REGNO_NREGS): Likewise.
2150 (VALID_BND_REG_MODE): New.
2151 (FIRST_BND_REG): New.
2152 (LAST_BND_REG): New.
2153 (reg_class): Add BND_REGS.
2154 (REG_CLASS_NAMES): Likewise.
2155 (REG_CLASS_CONTENTS): Likewise.
2157 (ANY_BND_REG_P): New.
2159 (HI_REGISTER_NAMES): Add bound registers.
2160 * config/i386/i386.md (UNSPEC_BNDMK): New.
2161 (UNSPEC_BNDMK_ADDR): New.
2162 (UNSPEC_BNDSTX): New.
2163 (UNSPEC_BNDLDX): New.
2164 (UNSPEC_BNDLDX_ADDR): New.
2165 (UNSPEC_BNDCL): New.
2166 (UNSPEC_BNDCU): New.
2167 (UNSPEC_BNDCN): New.
2168 (UNSPEC_MPX_FENCE): New.
2171 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
2172 (length_immediate): Likewise.
2173 (prefix_0f): Likewise.
2175 (prefix_rep): Check for bnd prefix.
2176 (length_nobnd): New.
2177 (length): Use length_nobnd if specified.
2182 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
2185 (simple_return_internal): Likewise.
2186 (simple_return_pop_internal): Likewise.
2187 (*indirect_jump): Add MPX bnd prefix.
2188 (*tablejump_1): Likewise.
2189 (simple_return_internal_long): Likewise.
2190 (simple_return_indirect_internal): Likewise.
2194 (*mov<mode>_internal_mpx): New.
2195 (<mode>_<bndcheck>): New.
2196 (*<mode>_<bndcheck>): New.
2201 * config/i386/predicates.md (lea_address_operand): Rename to...
2202 (address_no_seg_operand): ... this.
2203 (address_mpx_no_base_operand): New.
2204 (address_mpx_no_index_operand): New.
2205 (bnd_mem_operator): New.
2206 * config/i386/i386.opt (mmpx): New.
2207 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
2208 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
2210 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2212 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2213 * tree.def (POINTER_BOUNDS_TYPE): New.
2214 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2215 (POINTER_BOUNDS_MODE): New.
2216 (make_pointer_bounds_mode): New.
2217 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2218 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2219 (layout_type): Support POINTER_BOUNDS_TYPE.
2220 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2221 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2222 (type_contains_placeholder_1): Likewise.
2223 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2224 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2225 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2227 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2229 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2230 mode of address computation when calling memory_address_addr_space.
2232 2013-08-24 Richard Henderson <rth@twiddle.net>
2235 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2236 instead of create_convert_operand_to.
2237 (maybe_emit_sync_lock_test_and_set): Likewise.
2238 (expand_atomic_compare_and_swap): Likewise.
2239 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2241 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2243 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2244 Change param_is to use the struct and not the pointer to the struct.
2246 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2248 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2249 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2250 Move to targhooks.c.
2251 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2252 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2253 Relocate from gimplify.c.
2254 * targhooks.h: Add 2 prototypes.
2255 * tree.h: Delete 2 prototypes.
2257 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2259 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2260 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2262 2013-10-24 Cong Hou <congh@google.com>
2264 * convert.c (convert_to_real): Guard those unsafe math function
2265 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2268 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2271 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2273 (cgraph_create_edge): Use the new argument.
2274 (cgraph_create_indirect_edge): Likewise.
2276 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2278 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2280 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2281 changes statep->state.
2283 2013-10-24 Tobias Burnus <burnus@net-b.de>
2286 * tree-cfg.c (replace_loop_annotate): New function.
2287 (execute_build_cfg): Call it.
2288 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2289 * internal-fn.c (expand_ANNOTATE): New function.
2290 * internal-fn.def (ANNOTATE): Define as new internal function.
2291 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2292 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2293 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2294 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2295 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2297 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2298 Marcus Shawcroft <marcus.shawcroft@arm.com>
2300 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2301 Special case reload SP+C into none GENERAL_REGS.
2303 2013-10-24 Michael Matz <matz@suse.de>
2305 * gengtype.c (is_file_equal): Check that files will be same length.
2307 2013-10-25 Christian Bruel <christian.bruel@st.com>
2309 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2310 * gcc/config/sh/t-sh (sh-mem.o): New rule.
2311 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
2312 (sh_expand_cmpstr): New function.
2313 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2314 * gcc/config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2315 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2316 (rotlhi3_8): Rename.
2318 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2320 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2321 * Makefile.in (CXXDEPMODE): Assign and change users.
2322 (CCDEPMODE): Delete.
2323 * configure: Regenerate.
2325 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2327 * gengtype-parse.c (require_without_advance): New.
2328 (type): For GTY-marked types that are not GTY((user)), parse any
2329 base classes, requiring them to be single-inheritance, and not
2330 be templates. For non-GTY-marked types and GTY((user)),
2331 continue to skip over any C++ inheritance specification.
2332 * gengtype-state.c (state_writer::write_state_struct_type):
2333 Write base class of type (if any).
2334 (read_state_struct_type): Read base class of type (if any).
2335 * gengtype.c (new_structure): Add a "base_class" parameter.
2336 (create_optional_field_): Update for new parameter to new_structure.
2337 (adjust_field_rtx_def): Likewise.
2338 (adjust_field_tree_exp): Likewise.
2339 * gengtype.h (struct type): Add "base_class" field to the s
2341 (new_structure): Add "base" parameter.
2343 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2346 * config/i386/i386.c (ix86_option_override_internal):
2347 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2348 (ix86_valid_target_attribute_tree):
2349 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2350 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2352 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2354 * tree-ssa-loop.h: Remove include files.
2355 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2356 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2357 * graphite-clast-to-gimple.c: Likewise.
2358 * graphite-scop-detection.c: Likewise.
2359 * graphite-sese-to-poly.c: Likewise.
2360 * ipa-inline-analysis.c: Likewise.
2361 * ipa-pure-const.c: Likewise.
2362 * loop-init.c: Likewise.
2363 * passes.c: Likewise.
2364 * predict.c: Likewise.
2365 * tree-cfg.c: Likewise.
2366 * tree-cfgcleanup.c: Likewise.
2367 * tree-chrec.c: Likewise.
2368 * tree-data-ref.c: Likewise.
2369 * tree-loop-distribution.c: Likewise.
2370 * tree-parloops.c: Likewise.
2371 * tree-predcom.c: Likewise.
2372 * tree-scalar-evolution.c: Likewise.
2373 * tree-ssa-address.c: Likewise.
2374 * tree-ssa.c: Likewise.
2375 * tree-ssa-dce.c: Likewise.
2376 * tree-ssa-loop.c: Likewise.
2377 * tree-ssa-loop-im.c: Likewise.
2378 * tree-ssa-loop-ivcanon.c: Likewise.
2379 * tree-ssa-loop-ivopts.c: Likewise.
2380 * tree-ssa-loop-manip.c: Likewise.
2381 * tree-ssa-loop-niter.c: Likewise.
2382 * tree-ssa-loop-prefetch.c: Likewise.
2383 * tree-ssa-loop-unswitch.c: Likewise.
2384 * tree-ssa-reassoc.c: Likewise.
2385 * tree-vect-data-refs.c: Likewise.
2386 * tree-vect-loop.c: Likewise.
2387 * tree-vect-loop-manip.c: Likewise.
2388 * tree-vectorizer.c: Likewise.
2389 * tree-vect-stmts.c: Likewise.
2390 * tree-vrp.c: Likewise.
2392 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2394 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2396 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2398 PR tree-optimization/58775
2399 PR tree-optimization/58791
2400 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2401 (insert_stmt_after): Rewritten, don't move the stmt, but really
2403 (get_stmt_uid_with_default): Remove.
2404 (build_and_add_sum): Use insert_stmt_after and
2405 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2406 (update_range_test): Set uid on stmts added by
2407 force_gimple_operand_gsi. Don't immediately modify statements
2408 in inter-bb optimization, just update oe->op values.
2409 (optimize_range_tests): Return bool whether any changed have been made.
2410 (update_ops): New function.
2411 (struct inter_bb_range_test_entry): New type.
2412 (maybe_optimize_range_tests): Perform statement changes here.
2413 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2414 ensure_ops_are_available): Remove.
2415 (find_insert_point): Rewritten.
2416 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2417 return LHS of the (new resp. old) stmt. Don't call
2418 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2419 instead of last, move new stmt at the right place.
2420 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2421 (negate_value): Likewise. Set uids.
2422 (break_up_subtract_bb): Initialize uids.
2423 (reassociate_bb): Adjust rewrite_expr_tree caller.
2424 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2426 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2429 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2430 TARGET_32BIT final condition.
2431 (mulsi3_internal2 and splitter): Same.
2433 2013-10-23 Jeff Law <law@redhat.com>
2435 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2436 through joiner blocks with abnormal outgoing edges.
2438 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2439 Add parameter JOINERS, to allow/disallow threading through joiner
2441 (thread_block): New. Call thread_block_1.
2442 (mark_threaded_blocks): Remove code to filter out certain cases
2443 of threading through joiner blocks.
2444 (thread_through_all_blocks): Document how we can have a dangling
2445 edge AUX field and clear it.
2447 2013-10-23 Ian Lance Taylor <iant@google.com>
2449 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2450 (C++ Dialect Options): Likewise.
2451 (Optimize Options): Likewise.
2453 2013-10-23 Tom de Vries <tom@codesourcery.com>
2455 PR tree-optimization/58805
2456 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2458 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2460 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2461 sequence based on get_range_info returned range.
2463 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2465 * tree-ssa.h: Remove all #include's
2466 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2467 * alias.c: Move required includes from tree-ssa.h.
2469 * builtins.c: Likewise.
2470 * calls.c: Likewise.
2471 * cfgexpand.c: Likewise.
2472 * cfghooks.c: Likewise.
2473 * cfgloop.c: Likewise.
2474 * cfgloopmanip.c: Likewise.
2475 * cgraph.c: Likewise.
2476 * cgraphbuild.c: Likewise.
2477 * cgraphclones.c: Likewise.
2478 * cgraphunit.c: Likewise.
2480 * except.c: Likewise.
2482 * final.c: Likewise.
2483 * fold-const.c: Likewise.
2484 * ggc-page.c: Likewise.
2485 * gimple-builder.c: Likewise.
2486 * gimple-fold.c: Likewise.
2487 * gimple-iterator.c: Likewise.
2488 * gimple-low.c: Likewise.
2489 * gimple-pretty-print.c: Likewise.
2490 * gimple-ssa-strength-reduction.c: Likewise.
2491 * gimple-streamer-in.c: Likewise.
2492 * gimple-streamer-out.c: Likewise.
2493 * gimplify.c: Likewise.
2494 * graphite-blocking.c: Likewise.
2495 * graphite-clast-to-gimple.c: Likewise.
2496 * graphite-dependences.c: Likewise.
2497 * graphite-interchange.c: Likewise.
2498 * graphite-optimize-isl.c: Likewise.
2499 * graphite-poly.c: Likewise.
2500 * graphite-scop-detection.c: Likewise.
2501 * graphite-sese-to-poly.c: Likewise.
2502 * graphite.c: Likewise.
2503 * ipa-cp.c: Likewise.
2504 * ipa-inline-analysis.c: Likewise.
2505 * ipa-inline-transform.c: Likewise.
2506 * ipa-inline.c: Likewise.
2507 * ipa-prop.c: Likewise.
2508 * ipa-pure-const.c: Likewise.
2509 * ipa-reference.c: Likewise.
2510 * ipa-split.c: Likewise.
2511 * ipa-utils.c: Likewise.
2512 * loop-init.c: Likewise.
2513 * lto-cgraph.c: Likewise.
2514 * lto-section-in.c: Likewise.
2515 * lto-section-out.c: Likewise.
2516 * lto-streamer-in.c: Likewise.
2517 * lto-streamer-out.c: Likewise.
2518 * lto-streamer.c: Likewise.
2519 * omp-low.c: Likewise.
2520 * passes.c: Likewise.
2521 * predict.c: Likewise.
2522 * print-tree.c: Likewise.
2523 * profile.c: Likewise.
2525 * targhooks.c: Likewise.
2526 * tracer.c: Likewise.
2527 * trans-mem.c: Likewise.
2528 * tree-call-cdce.c: Likewise.
2529 * tree-cfg.c: Likewise.
2530 * tree-cfgcleanup.c: Likewise.
2531 * tree-chrec.c: Likewise.
2532 * tree-complex.c: Likewise.
2533 * tree-data-ref.c: Likewise.
2534 * tree-dfa.c: Likewise.
2535 * tree-eh.c: Likewise.
2536 * tree-emutls.c: Likewise.
2537 * tree-if-conv.c: Likewise.
2538 * tree-inline.c: Likewise.
2539 * tree-into-ssa.c: Likewise.
2540 * tree-loop-distribution.c: Likewise.
2541 * tree-mudflap.c: Likewise.
2542 * tree-nested.c: Likewise.
2543 * tree-nrv.c: Likewise.
2544 * tree-object-size.c: Likewise.
2545 * tree-outof-ssa.c: Likewise.
2546 * tree-parloops.c: Likewise.
2547 * tree-phinodes.c: Likewise.
2548 * tree-predcom.c: Likewise.
2549 * tree-pretty-print.c: Likewise.
2550 * tree-profile.c: Likewise.
2551 * tree-scalar-evolution.c: Likewise.
2552 * tree-sra.c: Likewise.
2553 * tree-ssa-address.c: Likewise.
2554 * tree-ssa-alias.c: Likewise.
2555 * tree-ssa-ccp.c: Likewise.
2556 * tree-ssa-coalesce.c: Likewise.
2557 * tree-ssa-copy.c: Likewise.
2558 * tree-ssa-copyrename.c: Likewise.
2559 * tree-ssa-dce.c: Likewise.
2560 * tree-ssa-dom.c: Likewise.
2561 * tree-ssa-dse.c: Likewise.
2562 * tree-ssa-forwprop.c: Likewise.
2563 * tree-ssa-ifcombine.c: Likewise.
2564 * tree-ssa-live.c: Likewise.
2565 * tree-ssa-loop-ch.c: Likewise.
2566 * tree-ssa-loop-im.c: Likewise.
2567 * tree-ssa-loop-ivcanon.c: Likewise.
2568 * tree-ssa-loop-ivopts.c: Likewise.
2569 * tree-ssa-loop-manip.c: Likewise.
2570 * tree-ssa-loop-niter.c: Likewise.
2571 * tree-ssa-loop-prefetch.c: Likewise.
2572 * tree-ssa-loop-unswitch.c: Likewise.
2573 * tree-ssa-loop.c: Likewise.
2574 * tree-ssa-math-opts.c: Likewise.
2575 * tree-ssa-operands.c: Likewise.
2576 * tree-ssa-phiopt.c: Likewise.
2577 * tree-ssa-phiprop.c: Likewise.
2578 * tree-ssa-pre.c: Likewise.
2579 * tree-ssa-propagate.c: Likewise.
2580 * tree-ssa-reassoc.c: Likewise.
2581 * tree-ssa-sccvn.c: Likewise.
2582 * tree-ssa-sink.c: Likewise.
2583 * tree-ssa-strlen.c: Likewise.
2584 * tree-ssa-structalias.c: Likewise.
2585 * tree-ssa-tail-merge.c: Likewise.
2586 * tree-ssa-ter.c: Likewise.
2587 * tree-ssa-threadedge.c: Likewise.
2588 * tree-ssa-threadupdate.c: Likewise.
2589 * tree-ssa-uncprop.c: Likewise.
2590 * tree-ssa-uninit.c: Likewise.
2591 * tree-ssa.c: Likewise.
2592 * tree-ssanames.c: Likewise.
2593 * tree-stdarg.c: Likewise.
2594 * tree-streamer-in.c: Likewise.
2595 * tree-switch-conversion.c: Likewise.
2596 * tree-tailcall.c: Likewise.
2597 * tree-vect-data-refs.c: Likewise.
2598 * tree-vect-generic.c: Likewise.
2599 * tree-vect-loop-manip.c: Likewise.
2600 * tree-vect-loop.c: Likewise.
2601 * tree-vect-patterns.c: Likewise.
2602 * tree-vect-slp.c: Likewise.
2603 * tree-vect-stmts.c: Likewise.
2604 * tree-vectorizer.c: Likewise.
2605 * tree-vrp.c: Likewise.
2608 * value-prof.c: Likewise.
2609 * var-tracking.c: Likewise.
2610 * varpool.c: Likewise.
2611 * vtable-verify.c: Likewise.
2613 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2615 * config/tilegx/tilegx.c: Include "tree.h".
2617 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2619 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2620 the info, not after it.
2621 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2622 dump_ssaname_info call.
2623 (pp_gimple_stmt_1): Adjust caller.
2624 (dump_phi_nodes): Likewise. Don't print "# " here.
2626 2013-10-22 Jan Hubicka <jh@suse.cz>
2628 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
2630 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
2631 * i386.c (expand_small_movmem_or_setmem): New function.
2632 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
2634 (alg_usable_p): Add support for value ranges; cleanup.
2635 (ix86_expand_set_or_movmem): Add support for misaligned moves.
2637 2013-10-22 Sterling Augustine <saugustine@google.com>
2639 * doc/invoke.texi: Document -ggnu-pubnames.
2640 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
2642 * dwarf2out.c: Include gdb/gdb-index.h.
2643 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
2644 debug_generate_pub_sections.
2645 (is_java, output_pubtables, output_pubname): New functions.
2646 (include_pubname_in_output): Handle debug_generate_pub_sections at
2648 (size_of_pubnames): Use new local space_for_flags based on
2649 debug_generate_pub_sections.
2650 (output_pubnames): Unify pubnames and pubtypes output logic.
2651 Genericize comments. Call output_pubname.
2652 (dwarf2out_finish): Move logic to output_pubnames and call it.
2654 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
2657 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
2658 Remove CCCmode handling.
2659 <case LTU>: Return 'c' suffix for CCCmode.
2660 <case GEU>: Return 'nc' suffix for CCCmode.
2661 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
2662 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
2663 (*sub<mode>3_cc_overflow): Ditto.
2664 (*subsi3_zext_cc_overflow): Ditto.
2666 2013-10-22 Steve Ellcey <sellcey@mips.com>
2668 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
2669 (AND (NOT OP1) (NOT OP2)).
2671 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
2673 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2674 meaning of merge-high and merge-low masks for little endian; avoid
2675 use of vector-pack masks for little endian for mismatched modes.
2677 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2679 * config/tilepro/tilepro.c: Include "tree.h".
2681 2013-10-22 Andreas Schwab <schwab@suse.de>
2683 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
2686 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2688 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
2689 [Ed Smith-Rowland]): New entries.
2690 ([Stephen M. Webb]): Update.
2692 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
2694 * tree-ssa-ter.h: Remove duplicate copy of file contents.
2696 2013-10-21 Marek Polacek <polacek@redhat.com>
2699 * fold-const.c (fold_range_test): Return 0 if the type is not
2702 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
2704 * system.h: Move hwint.h include further down.
2705 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
2706 gcc_checking_asserts.
2707 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
2709 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2711 Fix volatile issues in optimize_bit_field_compare.
2712 * fold-const.c (optimize_bit_field_compare): Bail out if
2713 lvolatilep or rvolatilep.
2715 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2717 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
2718 and get_inner_reference returning different pmode for non-volatile
2719 bit-field members dependent on flag_strict_volatile_bitfields.
2720 * stor-layout.c (layout_decl): Remove special handling of
2721 flag_strict_volatile_bitfields.
2722 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
2723 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
2725 2013-10-21 Paulo Matos <pmatos@broadcom.com>
2727 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
2730 2013-10-21 Jeff Law <law@redhat.com>
2732 * tree-ssa-threadedge.c (thread_through_normal_block): New
2733 argument VISITED. Remove VISISTED as a local variable. When we
2734 have a threadable jump, verify the destination of the jump has not
2736 (thread_across_edge): Allocate VISITED bitmap once at function
2737 scope and use it throughout. Make sure to set appropriate bits in
2738 VISITED for E (start of jump thread path).
2739 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
2740 through a joiner if any edge on the path has a recorded jump thread.
2742 2013-10-21 Ian Lance Taylor <iant@google.com>
2744 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
2745 don't imply that attributes can solve all problems.
2746 (Directory Options): Fix typo.
2748 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2750 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
2751 extend and extend_add.
2753 2013-10-21 Richard Biener <rguenther@suse.de>
2755 PR tree-optimization/58794
2756 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
2757 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
2759 2013-10-21 Richard Biener <rguenther@suse.de>
2762 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
2763 to (T) X for sign-changing conversions (or no conversion).
2765 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
2767 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
2769 2013-10-20 Jan Hubicka <jh@suse.cz>
2771 * config/i386/i386-tune.def: Add comment; organize into categories
2773 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2775 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
2776 argument. Update function comment.
2777 (expand_set_or_movmem_via_rep): New function combining
2778 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
2779 (expand_movmem_via_rep_mov): Remove.
2780 expand_setmem_via_rep_stos): Remove.
2781 (expand_movmem_epilogue): Update calls correspondingly.
2782 (expand_setmem_epilogue_via_loop): Likewise.
2784 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
2785 (expand_set_or_movmem_prologue): New function combining
2786 expand_movmem_prologue and expand_setmem_prologue.
2787 (expand_movmem_prologue): Remove.
2788 (expand_setmem_prologue): Remove.
2789 (expand_set_or_movmem_constant_prologue): New function combining
2790 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
2791 (expand_constant_movmem_prologue): Remove.
2792 (expand_constant_setmem_prologue): Remove.
2793 (promote_duplicated_reg): Allow vector-const0 value.
2794 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
2795 and ix86_expand_setmem.
2796 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
2797 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
2799 2013-10-21 Diego Novillo <dnovillo@google.com>
2801 * asan.c: Include tree.h
2802 * bb-reorder.c: Likewise.
2803 * cfgcleanup.c: Likewise.
2804 * cfgloopmanip.c: Likewise.
2805 * data-streamer-in.c: Likewise.
2806 * data-streamer-out.c: Likewise.
2807 * data-streamer.c: Likewise.
2808 * dwarf2cfi.c: Likewise.
2809 * graphite-blocking.c: Likewise.
2810 * graphite-clast-to-gimple.c: Likewise.
2811 * graphite-dependences.c: Likewise.
2812 * graphite-interchange.c: Likewise.
2813 * graphite-optimize-isl.c: Likewise.
2814 * graphite-poly.c: Likewise.
2815 * graphite-scop-detection.c: Likewise.
2816 * graphite-sese-to-poly.c: Likewise.
2817 * graphite.c: Likewise.
2818 * ipa-devirt.c: Likewise.
2819 * ipa-profile.c: Likewise.
2822 * loop-init.c: Likewise.
2823 * loop-unroll.c: Likewise.
2824 * lower-subreg.c: Likewise.
2825 * lto/lto-object.c: Likewise.
2826 * recog.c: Likewise.
2827 * reginfo.c: Likewise.
2828 * tree-loop-distribution.c: Likewise.
2829 * tree-parloops.c: Likewise.
2830 * tree-ssa-strlen.c: Likewise.
2831 * tree-streamer.c: Likewise.
2832 * value-prof.c: Likewise.
2833 * target-globals.c: Likewise.
2834 * expr.h: Include tree-core.h instead of tree.h.
2835 * gimple.h: Likewise.
2836 * ipa-prop.h: Likewise.
2837 * ipa-utils.h: Likewise.
2838 * lto-streamer.h: Likewise.
2839 * streamer-hooks.h: Likewise.
2840 * ipa-reference.h: Include cgraph.h instead of tree.h.
2841 * cgraph.h: Include basic-block.h instead of tree.h.
2842 * tree-streamer.h: Do not include tree.h.
2843 * genattrtab.c (write_header): Generate inclusion of
2845 * genautomata.c (main): Likewise.
2846 * genemit.c: Likewise.
2847 * genopinit.c: Likewise.
2848 * genoutput.c (output_prologue): Likewise.
2849 * genpeep.c: Likewise.
2851 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2853 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
2855 (vec_unpacku_hi_v8hi): Likewise.
2856 (vec_unpacku_lo_v16qi): Likewise.
2857 (vec_unpacku_lo_v8hi): Likewise.
2859 2013-10-20 Jan Hubicka <jh@suse.cz>
2861 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
2862 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
2863 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
2865 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
2867 * config/mips/mips.h (ISA_HAS_WSBH): Define.
2868 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
2870 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
2872 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
2875 * system.h: Undef m_slot.
2877 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2879 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
2880 all elements for both endian flavors.
2882 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2885 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
2886 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
2887 and SI_REG for 64bit SYSV ABI targets.
2889 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2891 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
2892 to multi_reg_return. Clarify that we are skipping USEs of multiple
2893 return registers. Use bool type where appropriate.
2895 2013-10-18 Jan Hubicka <jh@suse.cz>
2897 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
2899 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
2900 (X86_TUNE_PUSH_MEMORY): Likewise.
2901 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
2902 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
2903 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
2905 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
2906 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
2908 (ix86_option_override_internal): Update to use tune features instead
2911 2013-10-18 Cong Hou <congh@google.com>
2913 PR tree-optimization/58508
2914 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
2915 statement that contains data refs with zero-step.
2917 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2919 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
2920 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
2921 * gimple-low.c (gimple_check_call_arg,
2922 gimple_check_call_matching_types): Move to cgraph.c.
2923 * gimple-low.h: Remove prototype.
2924 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
2925 Relocate from gimple-low.c.
2926 * cgraph.h: Add prototype. Don't include basic-block.h.
2927 * gimplify.c: Add gimple-low to include list.
2928 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
2929 * tree-eh.c: Add gimple-low to include list.
2930 * tree-nested.c: Likewise.
2931 * cfgexpand.c: Add tree-ssa-address.h to include list.
2933 * gimple-fold.c: Likewise.
2934 * gimple-ssa-strength-reduction.c: Likewise.
2935 * trans-mem.c: Likewise.
2936 * tree-mudflap.c: Likewise.
2937 * tree-ssa-loop-ivopts.c: Likewise.
2938 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
2939 (degenerate_phi_result): Move to tree-phinodes.c.
2940 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
2941 * tree-ssa-threadedge.c: Likewise.
2942 * tree-vrp.c: Likewise.
2943 * tree-phinodes.c (degenerate_phi_result): Relocate here.
2944 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
2945 * tree-phinodes.h (degenerate_phi_result): Add prototype.
2946 * tree-ssa-copy.c: Include tree-ssa-dom.h.
2947 * tree-ssa-forwprop.c: Likewise.
2948 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
2949 pass_data_cleanup_cfg_post_optimizing,
2950 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
2951 * tree-optimize.c: Delete File.
2952 * graphite.c: Include tree-cfgcleanup.h.
2953 * passes.c: Likewise.
2954 * tree-cfg.c: Likewise.
2955 * tree-profile.c: Likewise.
2956 * tree-ssa-dse.c: Likewise.
2957 * tree-ssa-loop-ivcanon.c: Likewise.
2958 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
2959 * tree-outof-ssa.c: Include sbitmap.h.
2960 * tree-ssa-live.c: Likewise.
2961 * tree-ssa-propagate.c: Likewise.
2962 * tree-ssa-structalias.c: Likewise.
2963 * tree-stdarg.c: Likewise.
2964 * Makefile.in (OBJS): Delete tree-optimize.o.
2965 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
2966 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
2967 * varasm.c: Include basic-block.h.
2968 * cfgloop.h: Include function.h instead of basic-block.h
2969 (bb_loop_depth): Move to cfgloop.c.
2970 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
2972 2013-10-18 Teresa Johnson <tejohnson@google.com>
2974 * predict.c (probably_never_executed): Compare frequency-based
2975 count to number of training runs.
2976 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
2978 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2980 * config/arm/arm.c (cortexa9_extra_costs): New table.
2981 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
2983 2013-10-18 Jeff Law <law@redhat.com>
2985 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
2987 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
2988 (register_jump_thread): Add "registered_jump_thread" debug
2990 * dbgcnt.def (registered_jump_thread): New debug counter.
2992 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2994 * config/rs6000/rs6000.c: Include cgraph.h.
2996 2013-10-18 Teresa Johnson <tejohnson@google.com>
2998 * tree-ssa-tail-merge.c (replace_block_by): Update edge
2999 weights during merging.
3001 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3003 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
3004 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
3005 * tree-ssa-operands.h: Remove prototype.
3006 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
3007 * gimple.c (virtual_operand_p): Relocate from gimple.c.
3008 * gimple.h: Add prototype.
3009 * gimple-ssa.h: Include tree-ssa-operands.h.
3010 * tree-dump.c: Add tree-cfg.h to include list.
3011 * tree-ssa-alias.c: Add ipa-reference.h to include list.
3012 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
3013 * config/i386/i386.c: Don't include tree-flow.h.
3014 * config/rs6000/rs6000.c: Likewise.
3016 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3018 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3020 2013-10-18 Richard Biener <rguenther@suse.de>
3022 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3023 or TYPE_UNSIGNED of integral types.
3024 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3025 NULL_TREE for zero-precision integral types.
3027 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3029 * config/aarch64/arm_neon.h
3030 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3032 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3035 * opth-gen.awk: Define target_flags_explicit.
3037 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3039 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3040 fields to the reg_addr array that describes the valid addressing
3041 mode for any register, general purpose registers, floating point
3042 registers, and Altivec registers.
3043 (FIRST_RELOAD_REG_CLASS): Likewise.
3044 (LAST_RELOAD_REG_CLASS): Likewise.
3045 (struct reload_reg_map_type): Likewise.
3046 (reload_reg_map_type): Likewise.
3047 (RELOAD_REG_VALID): Likewise.
3048 (RELOAD_REG_MULTIPLE): Likewise.
3049 (RELOAD_REG_INDEXED): Likewise.
3050 (RELOAD_REG_OFFSET): Likewise.
3051 (RELOAD_REG_PRE_INCDEC): Likewise.
3052 (RELOAD_REG_PRE_MODIFY): Likewise.
3053 (reg_addr): Likewise.
3054 (mode_supports_pre_incdec_p): New helper functions to say whether
3055 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3056 (mode_supports_pre_modify_p): Likewise.
3057 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3058 print the valid address mode bits for each mode.
3059 (rs6000_debug_print_mode): Likewise.
3060 (rs6000_debug_reg_global): Likewise.
3061 (rs6000_setup_reg_addr_masks): New function to set up the address
3062 mask bits for each type.
3063 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3064 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3065 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3066 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3067 PRE_MODIFY are supported.
3068 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3069 registers, instead of {src,dest}_av_p.
3070 (rs6000_print_options_internal): Tweak the debug output slightly.
3072 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3074 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3077 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3079 * tree-flow.h (struct omp_region): Move to omp-low.c.
3080 Remove omp_ prototypes and variables.
3081 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3082 (copy_var_decl): Relocate prototype from tree-flow.h.
3083 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3084 * tree.h: Move prototype to omp-low.h.
3085 * omp-low.h: New File. Relocate prototypes here.
3086 * omp-low.c (struct omp_region): Make local here.
3087 (root_omp_region): Make static.
3088 (copy_var_decl) Move to gimple.c.
3089 (new_omp_region): Make static.
3090 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3091 * tree-cfg.c: Include omp-low.h.
3092 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3093 * gimplify.c: Include omp-low.h.
3094 * tree-parloops.c: Likewise.
3096 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3098 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3099 operands are in memory, prefer to force non-matched operand 1 to
3102 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3105 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3106 restrict TImode addresses to single indirect registers if both
3107 -mquad-memory and -mvsx-timode are used.
3108 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3109 we should emit load/store quad. Remove using %y for quad memory
3112 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3113 constraints to allow load/store quad on machines where TImode is
3114 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3115 for TImode to load integer constants.
3117 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3119 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3121 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3123 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3124 handling of STACK_REG.
3126 2013-10-17 Richard Biener <rguenther@suse.de>
3128 PR tree-optimization/58143
3129 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3131 (rewrite_to_defined_overflow): Likewise.
3132 (move_computations_dom_walker::before_dom): Rewrite stmts
3133 with undefined signed overflow that are not always executed
3134 into unsigned arithmetic.
3136 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3139 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3140 explicit isa flag to be an options variable, instead of using
3141 global_options_set. Remove define from rs6000.h.
3142 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3144 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3145 Initialize rs6000_isa_flags_explicit.
3146 (rs6000_function_specific_save): Add gcc_options* parameter, so
3147 that the powerpc builds after the 2013-10-15 changes.
3148 (rs6000_function_specific_restore): Likewise.
3150 2013-10-16 DJ Delorie <dj@redhat.com>
3152 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3153 op is a REG before checking REGNO.
3154 (rl78_alloc_physical_registers): Verify pattern is a SET before
3157 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3159 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
3161 (vec_unpacks_lo_v4sf): Likewise.
3162 (vec_unpacks_float_hi_v4si): Likewise.
3163 (vec_unpacks_float_lo_v4si): Likewise.
3164 (vec_unpacku_float_hi_v4si): Likewise.
3165 (vec_unpacku_float_lo_v4si): Likewise.
3167 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3169 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
3170 (vsx_concat_v2sf): Likewise.
3172 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
3174 * config/aarch64/aarch64.md
3175 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
3177 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
3179 PR tree-optimization/58697
3180 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
3181 estimated_loop_iterations_int.
3182 (max_stmt_executions_int): Call get_max_loop_iterations_int.
3183 (get_max_loop_iterations_int): New. HWINT version of
3184 get_max_loop_iterations.
3185 * cfgloop.h: Add prototypes.
3186 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
3187 * loop-unroll.c (decide_peel_once_rolling): Call
3188 get_estimated_loop_iterations_int.
3189 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
3190 * tree-ssa-loop-niter.h: Tweak prototypes.
3192 2013-10-16 David Malcolm <dmalcolm@redhat.com>
3194 * gengtype-parse.c (struct_field_seq): Ignore access-control
3195 keywords ("public:" etc).
3197 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3199 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
3200 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
3202 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
3204 * config/arm/arm.opt (mlra): New option.
3205 * config/arm/arm.c (arm_lra_p): New function.
3206 (TARGET_LRA_P): Define.
3208 2013-10-16 Paulo Matos <pmatos@broadcom.com>
3210 * tree-core.h (tree_code_name): Remove.
3211 * tree.h (get_tree_code_name): New prototype.
3212 * tree.c (tree_code_name): Make static.
3213 (get_tree_code_name): New function.
3214 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3215 tree_class_check_failed, tree_range_check_failed,
3216 tree_not_class_check_failed, omp_clause_check_failed,
3217 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3218 wrapper get_tree_code_name instead of calling tree_code_name directly.
3219 * tree-vrp.c (dump_asserts_for): Likewise.
3220 * tree-dump.c (dequeue_and_dump): Likewise.
3221 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3222 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3223 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3224 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3225 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3226 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3227 dump_gimple_omp_for): Likewise.
3228 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3229 * tree-ssa-pre.c (print_pre_expr): Likewise.
3230 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3231 * print-tree.c (print_node_brief, print_node): Likewise.
3232 * gimple.c (gimple_check_failed): Likewise.
3233 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3234 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3235 * config/mep/mep.c (mep_validate_vliw): Likewise.
3236 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3237 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3239 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3241 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3244 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3246 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3247 (MULTILIB_MATCHES): Add multilib for -march=v8.
3249 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3252 * optc-save-gen.awk: Add extra parameter to the save and restore
3254 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3255 * tree.c (build_optimization_node): New parameter. Add extra parameter
3256 to call to cl_optimization_save.
3257 (build_target_option_node): New parameter. Add extra parameter
3258 to call to cl_target_option_save.
3259 * tree.h (build_optimization_node): New parameter.
3260 (build_target_option_node): New parameter.
3261 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3262 build_optimization_node and build_target_option_node.
3263 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3264 (handle_pragma_push_options): Ditto.
3265 * toplev.c (process_options): Ditto.
3266 * opts.c (init_options_struct): Check for opts_set non-null.
3267 * target.def (target_option.save): New parameter.
3268 (target_option.restore): New parameter.
3269 * tm.texi: Generate.
3270 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3271 (ix86_pragma_target_parse): Ditto.
3272 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3274 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3275 to build_optimization_node and build_target_option_node.
3276 (rs6000_valid_attribute_p): Ditto.
3277 (rs6000_pragma_target_parse): Ditto.
3278 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3280 * config/i386/i386.h:
3281 TARGET_64BIT_P: New Macro
3282 TARGET_MMX_P: New Macro.
3283 TARGET_3DNOW_P: New Macro.
3284 TARGET_3DNOW_A_P: New Macro.
3285 TARGET_SSE_P: New Macro.
3286 TARGET_SSE2_P: New Macro.
3287 TARGET_SSE3_P: New Macro.
3288 TARGET_SSSE3_P: New Macro.
3289 TARGET_SSE4_1_P: New Macro.
3290 TARGET_SSE4_2_P: New Macro.
3291 TARGET_AVX_P: New Macro.
3292 TARGET_AVX2_P: New Macro.
3293 TARGET_AVX512F_P: New Macro.
3294 TARGET_AVX512PF_P: New Macro.
3295 TARGET_AVX512ER_P: New Macro.
3296 TARGET_AVX512CD_P: New Macro.
3297 TARGET_FMA_P: New Macro.
3298 TARGET_SSE4A_P: New Macro.
3299 TARGET_FMA4_P: New Macro.
3300 TARGET_XOP_P: New Macro.
3301 TARGET_LWP_P: New Macro.
3302 TARGET_ABM_P: New Macro.
3303 TARGET_BMI_P: New Macro.
3304 TARGET_BMI2_P: New Macro.
3305 TARGET_LZCNT_P: New Macro.
3306 TARGET_TBM_P: New Macro.
3307 TARGET_POPCNT_P: New Macro.
3308 TARGET_SAHF_P: New Macro.
3309 TARGET_MOVBE_P: New Macro.
3310 TARGET_CRC32_P: New Macro.
3311 TARGET_AES_P: New Macro.
3312 TARGET_PCLMUL_P: New Macro.
3313 TARGET_CMPXCHG16B_P: New Macro.
3314 TARGET_FSGSBASE_P: New Macro.
3315 TARGET_RDRND_P: New Macro.
3316 TARGET_F16C_P: New Macro.
3317 TARGET_RTM_P: New Macro.
3318 TARGET_HLE_P: New Macro.
3319 TARGET_RDSEED_P: New Macro.
3320 TARGET_PRFCHW_P: New Macro.
3321 TARGET_ADX_P: New Macro.
3322 TARGET_FXSR_P: New Macro.
3323 TARGET_XSAVE_P: New Macro.
3324 TARGET_XSAVEOPT_P: New Macro.
3325 TARGET_LP64_P: New Macro.
3326 TARGET_X32_P: New Macro.
3327 TARGET_FPMATH_DEFAULT_P: New Macro.
3328 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3329 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3331 Change ix86_tune_string to access opts->x_ix86_tune_string.
3332 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3333 Change ix86_arch_string to access opts->x_ix86_arch_string.
3334 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3335 Change ix86_pmode to access opts->x_ix86_pmode.
3336 Change ix86_abi to access opts->x_ix86_abi.
3337 Change ix86_cmodel to access opts->x_ix86_cmodel.
3338 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3339 Change ix86_isa_flags_explicit to access
3340 opts->x_ix86_isa_flags_explicit.
3341 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3342 Change ix86_regparm to access opts->x_ix86_regparm.
3343 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3344 Change ix86_preferred_stack_boundary_arg to access
3345 opts->x_ix86_preferred_stack_boundary_arg.
3346 Change ix86_force_align_arg_pointer to access
3347 opts->x_ix86_force_align_arg_pointer.
3348 Change ix86_incoming_stack_boundar_arg to access
3349 opts->x_ix86_incoming_stack_boundar_arg.
3350 Change ix86_fpmath to access opts->x_ix86_fpmath.
3351 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3352 Change ix86_recip_name to access opts->x_ix86_recip_name.
3353 Change ix86_stack_protector_guard to access
3354 opts->x_ix86_stack_protector_guard.
3355 Change ix86_tune_memcpy_strategy to access
3356 opts->x_ix86_tune_memcpy_strategy.
3357 Change ix86_tune_memset_strategy to access
3358 opts->x_ix86_tune_memset_strategy.
3359 Change global_options to access opts.
3360 Change global_options_set to access opts_set.
3361 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3362 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3363 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3364 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3365 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3366 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3367 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3368 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3369 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3370 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3371 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3372 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3373 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3374 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3375 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3376 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3377 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3378 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3379 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3380 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3381 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3382 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3383 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3384 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3385 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3386 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3387 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3388 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3389 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3390 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3391 Change TARGET_AES to TARGET_AES_P (opts->...).
3392 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3393 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3394 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3395 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3396 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3397 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3398 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3399 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3400 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3401 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3402 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3403 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3404 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3405 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3406 Change TARGET_X32 to TARGET_X32_P (opts->...).
3407 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3408 Change TARGET_FLOAT_RETURNS_IN_80387 to
3409 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3410 (ix86_function_specific_save): New parameter. Use opts-> fields
3411 to replace global fields.
3412 (ix86_function_specific_restore): Ditto.
3413 (ix86_valid_target_attribute_inner_p): New parameters.
3415 Fix call to ix86_handle_option and set_option.
3416 (ix86_valid_target_attribute_tree): New parameters.
3417 Change global_options to access opts.
3418 Change global_options_set to access opts_set.
3419 Fix call to ix86_valid_target_attribute_inner_p.
3420 Change ix86_tune_string to access opts->x_ix86_tune_string.
3421 Change ix86_arch_string to access opts->x_ix86_arch_string.
3422 Change ix86_fpmath to access opts->x_ix86_fpmath
3423 Fix call to ix86_option_override_internal.
3424 Fix call to ix86_add_new_builtins.
3425 Fix calls to build_optimization_node and build_target_option_node.
3426 (ix86_valid_target_attribute_p): Remove access to global_options.
3427 Use new gcc_options structure func_options.
3428 Fix call to ix86_valid_target_attribute_tree.
3429 Fix call to build_optimization_node.
3430 (get_builtin_code_for_version): Fix call to
3431 ix86_valid_target_attribute_tree.
3433 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3435 * Makefile.in (PICFLAG): New.
3436 (enable_host_shared): New.
3437 (INTERNAL_CFLAGS): Use PICFLAG.
3438 (LIBIBERTY): Use pic build of libiberty.a if configured with
3439 --enable-host-shared.
3440 * configure.ac: Add --enable-host-shared, setting up new
3442 * configure: Regenerate.
3443 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3444 (--enable-host-shared): New option.
3446 2013-10-15 Richard Biener <rguenther@suse.de>
3448 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3449 for built-in functions.
3451 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3453 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3454 (optimize_range_tests_1): New function,
3455 extracted from optimize_range_tests.
3456 (optimize_range_tests_xor): Similarly.
3457 (optimize_range_tests_diff): New function.
3458 (optimize_range_tests): Use optimize_range_tests_1.
3460 2013-10-15 Cong Hou <congh@google.com>
3462 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3463 requirement of the reduction pattern so that one operand of the
3464 reduction operation can come from outside of the loop.
3466 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3468 * config/arm/neon-schedgen.ml: Remove.
3469 * config/arm/cortex-a9-neon.md: Remove comment regarding
3472 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3474 * config/arm/types: Remove old neon types.
3476 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3478 * config/arm/cortex-a7.md
3479 (cortex_a7_neon_type): New.
3480 (cortex_a7_neon_mul): Update for new types.
3481 (cortex_a7_neon_mla): Likewise.
3482 (cortex_a7_neon): Likewise.
3484 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3486 * config/arm/cortex-a15-neon.md
3487 (cortex_a15_neon_type): New,
3489 (cortex_a15_neon_int_1): Remove.
3490 (cortex_a15_neon_int_2): Likewise.
3491 (cortex_a15_neon_int_3): Likewise.
3492 (cortex_a15_neon_int_4): Likewise.
3493 (cortex_a15_neon_int_5): Likewise.
3494 (cortex_a15_neon_vqneg_vqabs): Likewise.
3495 (cortex_a15_neon_vmov): Likewise.
3496 (cortex_a15_neon_vaba): Likewise.
3497 (cortex_a15_neon_vaba_qqq): Likewise.
3498 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3499 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3500 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3502 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3503 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3504 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3505 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3506 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3507 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3508 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3509 (cortex_a15_neon_shift_1): Likewise.
3510 (cortex_a15_neon_shift_2): Likewise.
3511 (cortex_a15_neon_shift_3): Likewise.
3512 (cortex_a15_neon_vshl_ddd): Likewise.
3513 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3514 (cortex_a15_neon_vsra_vrsra): Likewise.
3515 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3516 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3517 (cortex_a15_neon_bp_3cycle): Likewise.
3518 (cortex_a15_neon_ldm_2): Likewise.
3519 (cortex_a15_neon_stm_2): Likewise.
3520 (cortex_a15_neon_mcr): Likewise.
3521 (cortex_a15_neon_mrc): Likewise.
3522 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3523 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3524 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3525 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3526 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3527 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3528 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3529 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3530 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3531 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3532 (cortex_a15_neon_bp_simple): Likewise.
3533 (cortex_a15_neon_bp_2cycle): Likewise.
3534 (cortex_a15_neon_bp_3cycle): Likewise.
3535 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3536 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3537 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3538 (cortex_a15_neon_vld2_4_regs): Likewise.
3539 (cortex_a15_neon_vld3_vld4): Likewise.
3540 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3541 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3542 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3543 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3544 (cortex_a15_neon_vst3_vst4): Rename to...
3545 (cortex_a15_neon_vst4): This, update for new attributes.
3546 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3547 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3548 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3549 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3550 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3551 (cortex_a15_neon_ldm_2): Likewise.
3552 (cortex_a15_neon_stm_2): Likewise.
3553 (cortex_a15_neon_mcr): Likewise.
3554 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3555 (cortex_a15_neon_mrc): Likewise.
3556 (cortex_a15_neon_mrrc): Likewise.
3558 (cortex_a15_neon_abd): New.
3559 (cortex_a15_neon_abd_q): Likewise.
3560 (cortex_a15_neon_aba): Likewise.
3561 (cortex_a15_neon_aba_q): Likewise.
3562 (cortex_a15_neon_acc): Likewise.
3563 (cortex_a15_neon_acc_q): Likewise.
3564 (cortex_a15_neon_arith_basic): Likewise.
3565 (cortex_a15_neon_arith_complex): Likewise.
3566 (cortex_a15_neon_multiply): Likewise.
3567 (cortex_a15_neon_multiply_q): Likewise.
3568 (cortex_a15_neon_mla): Likewise.
3569 (cortex_a15_neon_mla_q): Likewise.
3570 (cortex_a15_neon_sat_mla_long): Likewise.
3571 (cortex_a15_neon_shift_acc): Likewise.
3572 (cortex_a15_neon_shift_imm_basic): Likewise.
3573 (cortex_a15_neon_shift_imm_complex): Likewise.
3574 (cortex_a15_neon_shift_reg_basic): Likewise.
3575 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3576 (cortex_a15_neon_shift_reg_complex): Likewise.
3577 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3578 (cortex_a15_neon_fp_negabs): Likewise
3579 (cortex_a15_neon_fp_arith): Likewise
3580 (cortex_a15_neon_fp_arith_q): Likewise
3581 (cortex_a15_neon_fp_cvt_int): Likewise
3582 (cortex_a15_neon_fp_cvt_int_q): Likewise
3583 (cortex_a15_neon_fp_cvt_16): Likewise
3584 (cortex_a15_neon_fp_mul): Likewise
3585 (cortex_a15_neon_fp_mul_q): Likewise
3586 (cortex_a15_neon_fp_mla): Likewise
3587 (cortex_a15_neon_fp_mla_q): Likewise
3588 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3589 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3590 (cortex_a15_neon_bitops): Likewise.
3591 (cortex_a15_neon_bitops_q): Likewise.
3592 (cortex_a15_neon_from_gp): Likewise.
3593 (cortex_a15_neon_from_gp_q): Likewise.
3594 (cortex_a15_neon_tbl3_tbl4): Likewise.
3595 (cortex_a15_neon_zip_q): Likewise.
3596 (cortex_a15_neon_to_gp): Likewise.
3597 (cortex_a15_neon_load_a): Likewise.
3598 (cortex_a15_neon_load_b): Likewise.
3599 (cortex_a15_neon_load_c): Likewise.
3600 (cortex_a15_neon_load_d): Likewise.
3601 (cortex_a15_neon_load_e): Likewise.
3602 (cortex_a15_neon_load_f): Likewise.
3603 (cortex_a15_neon_store_a): Likewise.
3604 (cortex_a15_neon_store_b): Likewise.
3605 (cortex_a15_neon_store_c): Likewise.
3606 (cortex_a15_neon_store_d): Likewise.
3607 (cortex_a15_neon_store_e): Likewise.
3608 (cortex_a15_neon_store_f): Likewise.
3609 (cortex_a15_neon_store_g): Likewise.
3610 (cortex_a15_neon_store_h): Likewise.
3611 (cortex_a15_vfp_to_from_gp): Likewise.
3613 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3615 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3617 (cortex_a9_neon_vshl_ddd): Remove.
3618 (cortex_a9_neon_vst3_vst4): Likewise.
3619 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3621 (cortex_a9_neon_bit_ops_q): New.
3623 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3624 (cortex_a9_neon_int_2): Likewise.
3625 (cortex_a9_neon_int_3): Likewise.
3626 (cortex_a9_neon_int_4): Likewise.
3627 (cortex_a9_neon_int_5): Likewise.
3628 (cortex_a9_neon_vqneg_vqabs): Likewise.
3629 (cortex_a9_neon_vmov): Likewise.
3630 (cortex_a9_neon_vaba): Likewise.
3631 (cortex_a9_neon_vaba_qqq): Likewise.
3632 (cortex_a9_neon_shift_1): Likewise.
3633 (cortex_a9_neon_shift_2): Likewise.
3634 (cortex_a9_neon_shift_3): Likewise.
3635 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3636 (cortex_a9_neon_vsra_vrsra): Likewise.
3637 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3638 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3639 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3641 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3642 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3643 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3645 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3646 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3647 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3648 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3649 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3650 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3651 (cortex_a9_neon_fp_vsum): Likewise.
3652 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3653 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3654 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3655 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3656 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3657 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3658 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3659 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3660 (cortex_a9_neon_bp_simple): Likewise.
3661 (cortex_a9_neon_bp_2cycle): Likewise.
3662 (cortex_a9_neon_bp_3cycle): Likewise.
3663 (cortex_a9_neon_ldr): Likewise.
3664 (cortex_a9_neon_str): Likewise.
3665 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3666 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3667 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3668 (cortex_a9_neon_vld2_4_regs): Likewise.
3669 (cortex_a9_neon_vld3_vld4): Likewise.
3670 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3671 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3672 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3673 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3674 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3675 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3676 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3677 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3678 (cortex_a9_neon_mcr): Likewise.
3679 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3680 (cortex_a9_neon_mrc): Likewise.
3681 (cortex_a9_neon_mrrc): Likewise.
3683 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3685 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
3687 (cortex_a8_neon_vshl_ddd): Remove.
3688 (cortex_a8_neon_vst3_vst4): Likewise.
3689 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3691 (cortex_a8_neon_bit_ops_q): New.
3693 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
3694 (cortex_a8_neon_int_2): Likewise..
3695 (cortex_a8_neon_int_3): Likewise.
3696 (cortex_a8_neon_int_5): Likewise.
3697 (cortex_a8_neon_vqneg_vqabs): Likewise.
3698 (cortex_a8_neon_int_4): Likewise.
3699 (cortex_a8_neon_vaba): Likewise.
3700 (cortex_a8_neon_vaba_qqq): Likewise.
3701 (cortex_a8_neon_shift_1): Likewise.
3702 (cortex_a8_neon_shift_2): Likewise.
3703 (cortex_a8_neon_shift_3): Likewise.
3704 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3705 (cortex_a8_neon_vsra_vrsra): Likewise.
3706 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3707 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3708 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3710 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3711 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3712 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3714 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3715 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3716 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3717 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3718 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3719 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3720 (cortex_a8_neon_fp_vsum): Likewise.
3721 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3722 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3723 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3724 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3725 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3726 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3727 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3728 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3729 (cortex_a8_neon_bp_simple): Likewise.
3730 (cortex_a8_neon_bp_2cycle): Likewise.
3731 (cortex_a8_neon_bp_3cycle): Likewise.
3732 (cortex_a8_neon_ldr): Likewise.
3733 (cortex_a8_neon_str): Likewise.
3734 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3735 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3736 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3737 (cortex_a8_neon_vld2_4_regs): Likewise.
3738 (cortex_a8_neon_vld3_vld4): Likewise.
3739 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3740 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3741 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3742 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3743 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3744 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3745 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3746 (cortex_a8_neon_mcr): Likewise.
3747 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3748 (cortex_a8_neon_mrc): Likewise.
3749 (cortex_a8_neon_mrrc): Likewise.
3751 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3753 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
3755 * config/aarch64/aarch64-simd.md (neon_type): Remove.
3756 (aarch64_simd_dup<mode>): Add "type" attribute.
3757 (aarch64_dup_lane<mode>): Likewise.
3758 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
3759 (*aarch64_simd_mov<mode>): Likewise.
3760 (aarch64_simd_mov_from_<mode>low): Likewise.
3761 (aarch64_simd_mov_from_<mode>high): Likewise.
3762 (orn<mode>3): Likewise.
3763 (bic<mode>3): Likewise.
3764 (add<mode>3): Likewise.
3765 (sub<mode>3): Likewise.
3766 (mul<mode>3): Likewise.
3767 (*aarch64_mul3_elt<mode>): Likewise.
3768 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3769 (*aarch64_mul3_elt_to_128df): Likewise.
3770 (*aarch64_mul3_elt_to_64v2df): Likewise.
3771 (neg<mode>2): Likewise.
3772 (abs<mode>2): Likewise.
3773 (abd<mode>_3): Likewise.
3774 (aba<mode>_3): Likewise.
3775 (fabd<mode>_3): Likewise.
3776 (*fabd_scalar<mode>3): Likewise.
3777 (and<mode>3): Likewise.
3778 (ior<mode>3): Likewise.
3779 (xor<mode>3): Likewise.
3780 (one_cmpl<mode>2): Likewise.
3781 (aarch64_simd_vec_set<mode>): Likewise.
3782 (aarch64_simd_lshr<mode>): Likewise.
3783 (aarch64_simd_ashr<mode>): Likewise.
3784 (aarch64_simd_imm_shl<mode>): Likewise.
3785 (aarch64_simd_reg_sshl<mode): Likewise.
3786 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
3787 (aarch64_simd_reg_shl<mode>_signed): Likewise.
3788 (aarch64_simd_vec_setv2di): Likewise.
3789 (aarch64_simd_vec_set<mode>): Likewise.
3790 (aarch64_mla<mode>): Likewise.
3791 (*aarch64_mla_elt<mode>): Likewise.
3792 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3793 (aarch64_mls<mode>): Likewise.
3794 (*aarch64_mls_elt<mode>): Likewise.
3795 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3796 (<su><maxmin><mode>3): Likewise.
3797 (move_lo_quad_<mode>): Likewise.
3798 (aarch64_simd_move_hi_quad_<mode>): Likewise.
3799 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
3800 (vec_pack_trunc_<mode>): Likewise.
3801 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
3802 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
3803 (*aarch64_<su>mlal_lo<mode>): Likewise.
3804 (*aarch64_<su>mlal_hi<mode>): Likewise.
3805 (*aarch64_<su>mlsl_lo<mode>): Likewise.
3806 (*aarch64_<su>mlsl_hi<mode>): Likewise.
3807 (*aarch64_<su>mlal<mode>): Likewise.
3808 (*aarch64_<su>mlsl<mode>): Likewise.
3809 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
3810 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
3811 (add<mode>3): Likewise.
3812 (sub<mode>3): Likewise.
3813 (mul<mode>3): Likewise.
3814 (div<mode>3): Likewise.
3815 (neg<mode>2): Likewise.
3816 (abs<mode>2): Likewise.
3817 (fma<mode>4): Likewise.
3818 (*aarch64_fma4_elt<mode>): Likewise.
3819 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3820 (*aarch64_fma4_elt_to_128df): Likewise.
3821 (*aarch64_fma4_elt_to_64v2df): Likewise.
3822 (fnma<mode>4): Likewise.
3823 (*aarch64_fnma4_elt<mode>): Likewise.
3824 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3825 (*aarch64_fnma4_elt_to_128df): Likewise.
3826 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3827 (<frint_pattern><mode>2): Likewise.
3828 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3829 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3830 (vec_unpacks_lo_v4sf): Likewise.
3831 (aarch64_float_extend_lo_v2df): Likewise.
3832 (vec_unpacks_hi_v4sf): Likewise.
3833 (aarch64_float_truncate_lo_v2sf): Likewise.
3834 (aarch64_float_truncate_hi_v4sf): Likewise.
3835 (aarch64_vmls<mode>): Likewise.
3836 (<su><maxmin><mode>3): Likewise.
3837 (<maxmin_uns><mode>3): Likewise.
3838 (reduc_<sur>plus_<mode>): Likewise.
3839 (reduc_<sur>plus_v2di): Likewise.
3840 (reduc_<sur>plus_v2si): Likewise.
3841 (reduc_<sur>plus_<mode>): Likewise.
3842 (aarch64_addpv4sf): Likewise.
3843 (clz<mode>2): Likewise.
3844 (reduc_<maxmin_uns>_<mode>): Likewise.
3845 (reduc_<maxmin_uns>_v2di): Likewise.
3846 (reduc_<maxmin_uns>_v2si): Likewise.
3847 (reduc_<maxmin_uns>_<mode>): Likewise.
3848 (reduc_<maxmin_uns>_v4sf): Likewise.
3849 (aarch64_simd_bsl<mode>_internal): Likewise.
3850 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
3851 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
3852 (aarch64_get_lane<mode>): Likewise.
3853 (*aarch64_combinez<mode>): Likewise.
3854 (aarch64_combine<mode>): Likewise.
3855 (aarch64_simd_combine<mode>): Likewise.
3856 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
3857 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
3858 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
3859 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
3860 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
3861 (aarch64_<sur>h<addsub><mode>): Likewise.
3862 (aarch64_<sur><addsub>hn<mode>): Likewise.
3863 (aarch64_<sur><addsub>hn2<mode>): Likewise.
3864 (aarch64_pmul<mode>): Likewise.
3865 (aarch64_<su_optab><optab><mode>): Likewise.
3866 (aarch64_<sur>qadd<mode>): Likewise.
3867 (aarch64_sqmovun<mode>): Likewise.
3868 (aarch64_<sur>qmovn<mode>): Likewise.
3869 (aarch64_s<optab><mode>): Likewise.
3870 (aarch64_sq<r>dmulh<mode>): Likewise.
3871 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3872 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
3873 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3874 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
3875 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3876 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3877 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
3878 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
3879 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
3880 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3881 (aarch64_sqdmull<mode>): Likewise.
3882 (aarch64_sqdmull_lane<mode>_internal): Likewise.
3883 (aarch64_sqdmull_n<mode>): Likewise.
3884 (aarch64_sqdmull2<mode>_internal): Likewise.
3885 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
3886 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3887 (aarch64_<sur>shl<mode>): Likewise.
3888 (aarch64_<sur>q<r>shl<mode>
3889 (aarch64_<sur>shll_n<mode>): Likewise.
3890 (aarch64_<sur>shll2_n<mode>): Likewise.
3891 (aarch64_<sur>shr_n<mode>): Likewise.
3892 (aarch64_<sur>sra_n<mode>): Likewise.
3893 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
3894 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
3895 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
3896 (aarch64_cm<optab><mode>): Likewise.
3897 (aarch64_cm<optab>di): Likewise.
3898 (aarch64_cm<optab><mode>): Likewise.
3899 (aarch64_cm<optab>di): Likewise.
3900 (aarch64_cmtst<mode>): Likewise.
3901 (aarch64_cmtstdi): Likewise.
3902 (aarch64_cm<optab><mode>): Likewise.
3903 (*aarch64_fac<optab><mode>): Likewise.
3904 (aarch64_addp<mode>): Likewise.
3905 (aarch64_addpdi): Likewise.
3906 (sqrt<mode>2): Likewise.
3907 (vec_load_lanesoi<mode>): Likewise.
3908 (vec_store_lanesoi<mode>): Likewise.
3909 (vec_load_lanesci<mode>): Likewise.
3910 (vec_store_lanesci<mode>): Likewise.
3911 (vec_load_lanesxi<mode>): Likewise.
3912 (vec_store_lanesxi<mode>): Likewise.
3913 (*aarch64_mov<mode>): Likewise.
3914 (aarch64_ld2<mode>_dreg): Likewise.
3915 (aarch64_ld2<mode>_dreg): Likewise.
3916 (aarch64_ld3<mode>_dreg): Likewise.
3917 (aarch64_ld3<mode>_dreg): Likewise.
3918 (aarch64_ld4<mode>_dreg): Likewise.
3919 (aarch64_ld4<mode>_dreg): Likewise.
3920 (aarch64_tbl1<mode>): Likewise.
3921 (aarch64_tbl2v16qi): Likewise.
3922 (aarch64_combinev16qi): Likewise.
3923 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
3924 (aarch64_st2<mode>_dreg): Likewise.
3925 (aarch64_st2<mode>_dreg): Likewise.
3926 (aarch64_st3<mode>_dreg): Likewise.
3927 (aarch64_st3<mode>_dreg): Likewise.
3928 (aarch64_st4<mode>_dreg): Likewise.
3929 (aarch64_st4<mode>_dreg): Likewise.
3930 (*aarch64_simd_ld1r<mode>): Likewise.
3931 (aarch64_frecpe<mode>): Likewise.
3932 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
3933 (aarch64_frecps<mode>): Likewise.
3935 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3937 * config/arm/iterators.md (V_elem_ch): New.
3939 (VQH_type): Likewise.
3940 * config/arm/arm.md (is_neon_type): New.
3941 (conds): Use is_neon_type.
3942 (anddi3_insn): Update type attribute.
3943 (xordi3_insn): Likewise.
3944 (one_cmpldi2): Likewise.
3945 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
3946 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
3947 (*movmisalign<mode>_neon_store): Likewise.
3948 (*movmisalign<mode>_neon_load): Likewise.
3949 (vec_set<mode>_internal): Likewise.
3950 (vec_set<mode>_internal): Likewise.
3951 (vec_setv2di_internal): Likewise.
3952 (vec_extract<mode>): Likewise.
3953 (vec_extract<mode>): Likewise.
3954 (vec_extractv2di): Likewise.
3955 (*add<mode>3_neon): Likewise.
3956 (adddi3_neon): Likewise.
3957 (*sub<mode>3_neon): Likewise.
3958 (subdi3_neon): Likewise.
3959 (fma<VCVTF:mode>4): Likewise.
3960 (fma<VCVTF:mode>4_intrinsic): Likewise.
3961 (*fmsub<VCVTF:mode>4): Likewise.
3962 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3963 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3964 (ior<mode>3): Likewise.
3965 (and<mode>3): Likewise.
3966 (orn<mode>3_neon): Likewise.
3967 (orndi3_neon): Likewise.
3968 (bic<mode>3_neon): Likewise.
3969 (bicdi3_neon): Likewise.
3970 (xor<mode>3): Likewise.
3971 (one_cmpl<mode>2): Likewise.
3972 (abs<mode>2): Likewise.
3973 (neg<mode>2): Likewise.
3974 (negdi2_neon): Likewise.
3975 (*umin<mode>3_neon): Likewise.
3976 (*umax<mode>3_neon): Likewise.
3977 (*smin<mode>3_neon): Likewise.
3978 (*smax<mode>3_neon): Likewise.
3979 (vashl<mode>3): Likewise.
3980 (vashr<mode>3_imm): Likewise.
3981 (vlshr<mode>3_imm): Likewise.
3982 (ashl<mode>3_signed): Likewise.
3983 (ashl<mode>3_unsigned): Likewise.
3984 (neon_load_count): Likewise.
3985 (ashldi3_neon_noclobber): Likewise.
3986 (ashldi3_neon): Likewise.
3987 (signed_shift_di3_neon): Likewise.
3988 (unsigned_shift_di3_neon): Likewise.
3989 (ashrdi3_neon_imm_noclobber): Likewise.
3990 (lshrdi3_neon_imm_noclobber): Likewise.
3991 (<shift>di3_neon): Likewise.
3992 (widen_ssum<mode>3): Likewise.
3993 (widen_usum<mode>3): Likewise.
3994 (quad_halves_<code>v4si): Likewise.
3995 (quad_halves_<code>v4sf): Likewise.
3996 (quad_halves_<code>v8hi): Likewise.
3997 (quad_halves_<code>v16qi): Likewise.
3998 (reduc_splus_v2di): Likewise.
3999 (neon_vpadd_internal<mode>): Likewise.
4000 (neon_vpsmin<mode>): Likewise.
4001 (neon_vpsmax<mode>): Likewise.
4002 (neon_vpumin<mode>): Likewise.
4003 (neon_vpumax<mode>): Likewise.
4004 (*ss_add<mode>_neon): Likewise.
4005 (*us_add<mode>_neon): Likewise.
4006 (*ss_sub<mode>_neon): Likewise.
4007 (*us_sub<mode>_neon): Likewise.
4008 (neon_vadd<mode>_unspec): Likewise.
4009 (neon_vaddl<mode>): Likewise.
4010 (neon_vaddw<mode>): Likewise.
4011 (neon_vhadd<mode>): Likewise.
4012 (neon_vqadd<mode>): Likewise.
4013 (neon_vaddhn<mode>): Likewise.
4014 (neon_vmul<mode>): Likewise.
4015 (neon_vfms<VCVTF:mode>): Likewise.
4016 (neon_vmlal<mode>): Likewise.
4017 (neon_vmls<mode>): Likewise.
4018 (neon_vmlsl<mode>): Likewise.
4019 (neon_vqdmulh<mode>): Likewise.
4020 (neon_vqdmlal<mode>): Likewise.
4021 (neon_vqdmlsl<mode>): Likewise.
4022 (neon_vmull<mode>): Likewise.
4023 (neon_vqdmull<mode>): Likewise.
4024 (neon_vsub<mode>_unspec): Likewise.
4025 (neon_vsubl<mode>): Likewise.
4026 (neon_vsubw<mode>): Likewise.
4027 (neon_vqsub<mode>): Likewise.
4028 (neon_vhsub<mode>): Likewise.
4029 (neon_vsubhn<mode>): Likewise.
4030 (neon_vceq<mode>): Likewise.
4031 (neon_vcge<mode>): Likewise.
4032 (neon_vcgeu<mode>): Likewise.
4033 (neon_vcgt<mode>): Likewise.
4034 (neon_vcgtu<mode>): Likewise.
4035 (neon_vcle<mode>): Likewise.
4036 (neon_vclt<mode>): Likewise.
4037 (neon_vcage<mode>): Likewise.
4038 (neon_vcagt<mode>): Likewise.
4039 (neon_vtst<mode>): Likewise.
4040 (neon_vabd<mode>): Likewise.
4041 (neon_vabdl<mode>): Likewise.
4042 (neon_vaba<mode>): Likewise.
4043 (neon_vabal<mode>): Likewise.
4044 (neon_vmax<mode>): Likewise.
4045 (neon_vmin<mode>): Likewise.
4046 (neon_vpaddl<mode>): Likewise.
4047 (neon_vpadal<mode>): Likewise.
4048 (neon_vpmax<mode>): Likewise.
4049 (neon_vpmin<mode>): Likewise.
4050 (neon_vrecps<mode>): Likewise.
4051 (neon_vrsqrts<mode>): Likewise.
4052 (neon_vqabs<mode>): Likewise.
4053 (neon_vqneg<mode>): Likewise.
4054 (neon_vcls<mode>): Likewise.
4055 (clz<mode>2): Likewise.
4056 (popcount<mode>2): Likewise.
4057 (neon_vrecpe<mode>): Likewise.
4058 (neon_vrsqrte<mode>): Likewise.
4059 (neon_vget_lane<mode>_sext_internal): Likewise.
4060 (neon_vget_lane<mode>_zext_internal): Likewise.
4061 (neon_vdup_n<mode>): Likewise.
4062 (neon_vdup_n<mode>): Likewise.
4063 (neon_vdup_nv2di): Likewise.
4064 (neon_vdup_lane<mode>_interal): Likewise.
4065 (*neon_vswp<mode>): Likewise.
4066 (neon_vcombine<mode>): Likewise.
4067 (float<mode><V_cvtto>2): Likewise.
4068 (floatuns<mode><V_cvtto>2): Likewise.
4069 (fix_trunc<mode><V_cvtto>2): Likewise.
4070 (fixuns_trunc<mode><V_cvtto>2
4071 (neon_vcvt<mode>): Likewise.
4072 (neon_vcvt<mode>): Likewise.
4073 (neon_vcvtv4sfv4hf): Likewise.
4074 (neon_vcvtv4hfv4sf): Likewise.
4075 (neon_vcvt_n<mode>): Likewise.
4076 (neon_vcvt_n<mode>): Likewise.
4077 (neon_vmovn<mode>): Likewise.
4078 (neon_vqmovn<mode>): Likewise.
4079 (neon_vqmovun<mode>): Likewise.
4080 (neon_vmovl<mode>): Likewise.
4081 (neon_vmul_lane<mode>): Likewise.
4082 (neon_vmul_lane<mode>): Likewise.
4083 (neon_vmull_lane<mode>): Likewise.
4084 (neon_vqdmull_lane<mode>): Likewise.
4085 (neon_vqdmulh_lane<mode>): Likewise.
4086 (neon_vqdmulh_lane<mode>): Likewise.
4087 (neon_vmla_lane<mode>): Likewise.
4088 (neon_vmla_lane<mode>): Likewise.
4089 (neon_vmlal_lane<mode>): Likewise.
4090 (neon_vqdmlal_lane<mode>): Likewise.
4091 (neon_vmls_lane<mode>): Likewise.
4092 (neon_vmls_lane<mode>): Likewise.
4093 (neon_vmlsl_lane<mode>): Likewise.
4094 (neon_vqdmlsl_lane<mode>): Likewise.
4095 (neon_vext<mode>): Likewise.
4096 (neon_vrev64<mode>): Likewise.
4097 (neon_vrev32<mode>): Likewise.
4098 (neon_vrev16<mode>): Likewise.
4099 (neon_vbsl<mode>_internal): Likewise.
4100 (neon_vshl<mode>): Likewise.
4101 (neon_vqshl<mode>): Likewise.
4102 (neon_vshr_n<mode>): Likewise.
4103 (neon_vshrn_n<mode>): Likewise.
4104 (neon_vqshrn_n<mode>): Likewise.
4105 (neon_vqshrun_n<mode>): Likewise.
4106 (neon_vshl_n<mode>): Likewise.
4107 (neon_vqshl_n<mode>): Likewise.
4108 (neon_vqshlu_n<mode>): Likewise.
4109 (neon_vshll_n<mode>): Likewise.
4110 (neon_vsra_n<mode>): Likewise.
4111 (neon_vsri_n<mode>): Likewise.
4112 (neon_vsli_n<mode>): Likewise.
4113 (neon_vtbl1v8qi): Likewise.
4114 (neon_vtbl2v8qi): Likewise.
4115 (neon_vtbl3v8qi): Likewise.
4116 (neon_vtbl4v8qi): Likewise.
4117 (neon_vtbl1v16qi): Likewise.
4118 (neon_vtbl2v16qi): Likewise.
4119 (neon_vcombinev16qi): Likewise.
4120 (neon_vtbx1v8qi): Likewise.
4121 (neon_vtbx2v8qi): Likewise.
4122 (neon_vtbx3v8qi): Likewise.
4123 (neon_vtbx4v8qi): Likewise.
4124 (*neon_vtrn<mode>_insn): Likewise.
4125 (*neon_vzip<mode>_insn): Likewise.
4126 (*neon_vuzp<mode>_insn): Likewise.
4127 (neon_vld1<mode>): Likewise.
4128 (neon_vld1_lane<mode>): Likewise.
4129 (neon_vld1_lane<mode>): Likewise.
4130 (neon_vld1_dup<mode>): Likewise.
4131 (neon_vld1_dup<mode>): Likewise.
4132 (neon_vld1_dupv2di): Likewise.
4133 (neon_vst1<mode>): Likewise.
4134 (neon_vst1_lane<mode>): Likewise.
4135 (neon_vst1_lane<mode>): Likewise.
4136 (neon_vld2<mode>): Likewise.
4137 (neon_vld2<mode>): Likewise.
4138 (neon_vld2_lane<mode>): Likewise.
4139 (neon_vld2_lane<mode>): Likewise.
4140 (neon_vld2_dup<mode>): Likewise.
4141 (neon_vst2<mode>): Likewise.
4142 (neon_vst2<mode>): Likewise.
4143 (neon_vst2_lane<mode>): Likewise.
4144 (neon_vst2_lane<mode>): Likewise.
4145 (neon_vld3<mode>): Likewise.
4146 (neon_vld3qa<mode>): Likewise.
4147 (neon_vld3qb<mode>): Likewise.
4148 (neon_vld3_lane<mode>): Likewise.
4149 (neon_vld3_lane<mode>): Likewise.
4150 (neon_vld3_dup<mode>): Likewise.
4151 (neon_vst3<mode>): Likewise.
4152 (neon_vst3qa<mode>): Likewise.
4153 (neon_vst3qb<mode>): Likewise.
4154 (neon_vst3_lane<mode>): Likewise.
4155 (neon_vst3_lane<mode>): Likewise.
4156 (neon_vld4<mode>): Likewise.
4157 (neon_vld4qa<mode>): Likewise.
4158 (neon_vld4qb<mode>): Likewise.
4159 (neon_vld4_lane<mode>): Likewise.
4160 (neon_vld4_lane<mode>): Likewise.
4161 (neon_vld4_dup<mode>): Likewise.
4162 (neon_vst4<mode>): Likewise.
4163 (neon_vst4qa<mode>): Likewise.
4164 (neon_vst4qb<mode>): Likewise.
4165 (neon_vst4_lane<mode>): Likewise.
4166 (neon_vst4_lane<mode>): Likewise.
4167 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4168 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4169 (neon_vec_<US>mult_lo_<mode>): Likewise.
4170 (neon_vec_<US>mult_hi_<mode>): Likewise.
4171 (neon_vec_<US>shiftl_<mode>): Likewise.
4172 (neon_unpack<US>_<mode>): Likewise.
4173 (neon_vec_<US>mult_<mode>): Likewise.
4174 (vec_pack_trunc_<mode>): Likewise.
4175 (neon_vec_pack_trunc_<mode>): Likewise.
4176 (neon_vabd<mode>_2): Likewise.
4177 (neon_vabd<mode>_3): Likewise.
4179 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4181 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
4182 (load_pair): Update type attribute.
4183 (store_pair): Update type attribute.
4184 * config/aarch64/iterators.md (q): New.
4186 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4188 * config/arm/types.md: Add new types for Neon insns.
4190 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4191 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4192 Sergey Lega <sergey.s.lega@intel.com>
4193 Anna Tikhonova <anna.tikhonova@intel.com>
4194 Ilya Tocar <ilya.tocar@intel.com>
4195 Andrey Turetskiy <andrey.turetskiy@intel.com>
4196 Ilya Verbin <ilya.verbin@intel.com>
4197 Kirill Yukhin <kirill.yukhin@intel.com>
4198 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4200 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
4201 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
4202 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
4204 (srcp14<mode>): Ditto.
4205 (rsqrt14<mode>): Ditto.
4206 (rsqrt14<mode>): Ditto.
4207 (avx512f_vmscalef<mode>): Ditto.
4208 (avx512f_scalef<mode>): Ditto.
4209 (avx512f_getexp<mode>): Ditto.
4210 (avx512f_sgetexp<mode>): Ditto.
4211 (avx512f_fixupimm<mode>): Ditto.
4212 (avx512f_sfixupimm<mode>): Ditto.
4213 (avx512f_rndscale<mode>): Ditto.
4214 (*avx512er_exp2<mode>): Ditto.
4215 (*avx512er_rcp28<mode>): Ditto.
4216 (avx512er_rsqrt28<mode>): Ditto.
4217 (avx512f_getmant<mode>): Ditto.
4218 (avx512f_getmant<mode>): Ditto.
4219 (avx512f_rndscale<mode>): Fix formatting.
4221 2013-10-15 Martin Jambor <mjambor@suse.cz>
4223 * ipa-utils.h (ipa_edge_within_scc): Declare.
4224 * ipa-cp.c (edge_within_scc): Moved...
4225 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4227 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4228 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4229 Sergey Lega <sergey.s.lega@intel.com>
4230 Anna Tikhonova <anna.tikhonova@intel.com>
4231 Ilya Tocar <ilya.tocar@intel.com>
4232 Andrey Turetskiy <andrey.turetskiy@intel.com>
4233 Ilya Verbin <ilya.verbin@intel.com>
4234 Kirill Yukhin <kirill.yukhin@intel.com>
4235 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4237 * config/i386/predicates.md (const_8_to_15_operand): New.
4238 (const_16_to_31_operand): Ditto.
4239 * config/i386/sse.md (V8FI): New.
4241 (reduc_splus_v8df): Ditto.
4242 (reduc_splus_v16sf): Ditto.
4243 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4244 (vec_extract_hi_<mode>): Ditto.
4245 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4246 (vec_set_lo_<mode>): Ditto.
4247 (vec_set_hi_<mode>): Ditto.
4248 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4249 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4250 (avx512f_pshufd_1): Ditto.
4251 (avx512f_broadcast<mode>): Ditto.
4252 (avx512f_broadcast<mode>): Ditto.
4253 (define_split): Split vec_extract_lo into move.
4254 (ssequartermode): Ditto.
4255 (ssedoublemode): Extened with wider modes.
4256 (vec_extract_lo_<mode>): Ditto.
4258 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4259 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4260 Sergey Lega <sergey.s.lega@intel.com>
4261 Anna Tikhonova <anna.tikhonova@intel.com>
4262 Ilya Tocar <ilya.tocar@intel.com>
4263 Andrey Turetskiy <andrey.turetskiy@intel.com>
4264 Ilya Verbin <ilya.verbin@intel.com>
4265 Kirill Yukhin <kirill.yukhin@intel.com>
4266 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4268 * config/i386/predicates.md (register_or_constm1_operand): New.
4269 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4270 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4271 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4272 UNSPEC_SCATTER_PREFETCH
4274 (avx512f_ucmp<mode>3): Ditto.
4275 (avx512f_vternlog<mode>): Ditto.
4276 (avx512f_align<mode>): Ditto.
4277 (<shift_insn><mode>3): Ditto.
4278 (avx512f_<rotate>v<mode>): Ditto.
4279 (avx512f_<rotate><mode>): Ditto.
4280 (avx512f_eq<mode>3): Ditto.
4281 (avx512f_eq<mode>3_1): Ditto.
4282 (avx512f_gt<mode>3): Ditto.
4283 (avx512f_testm<mode>3): Ditto.
4284 (avx512f_testnm<mode>3): Ditto.
4285 (avx512pf_gatherpf<mode>): Ditto.
4286 (*avx512pf_gatherpf<mode>_mask): Ditto.
4287 (*avx512pf_gatherpf<mode>): Ditto.
4288 (avx512pf_scatterpf<mode>): Ditto.
4289 (*avx512pf_scatterpf<mode>_mask): Ditto.
4290 (*avx512pf_scatterpf<mode>): Ditto.
4291 (avx512f_vec_dup_gpr<mode>): Ditto.
4292 (clz<mode>2): Ditto.
4293 (conflict<mode>): Ditto.
4294 (REDUC_SMINMAX_MODE): Extened with wider modes.
4295 (reduc_<code>_<mode>): Ditto.
4296 (vlshr<mode>3): Ditto.
4297 (vashl<mode>3): Ditto.
4299 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4300 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4301 Sergey Lega <sergey.s.lega@intel.com>
4302 Anna Tikhonova <anna.tikhonova@intel.com>
4303 Ilya Tocar <ilya.tocar@intel.com>
4304 Andrey Turetskiy <andrey.turetskiy@intel.com>
4305 Ilya Verbin <ilya.verbin@intel.com>
4306 Kirill Yukhin <kirill.yukhin@intel.com>
4307 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4309 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4312 (avx512fmaskmode): Ditto.
4313 (bcstscalarsuff): Ditto.
4314 (avx512f_blendm<mode>): Ditto.
4315 (cmp_imm_predicate): Ditto.
4316 (avx512f_cmp<mode>3): Ditto.
4317 (avx512f_vec_dup<mode>): Ditto.
4318 (avx512f_vec_dup_mem<mode>): Ditto.
4319 (avx512f_vpermi2var<mode>3): Ditto.
4320 (avx512f_vpermt2var<mode>3): Ditto.
4321 (vec_init<mode>): Ditto.
4322 (avx512f_gathersi<mode>): Ditto.
4323 (*avx512f_gathersi<mode>): Ditto.
4324 (*avx512f_gathersi<mode>_2): Ditto.
4325 (avx512f_gatherdi<mode>): Ditto.
4326 (*avx512f_gatherdi<mode>): Ditto.
4327 (*avx512f_gatherdi<mode>_2): Ditto.
4328 (avx512f_scattersi<mode>): Ditto.
4329 (*avx512f_scattersi<mode>): Ditto.
4330 (avx512f_scatterdi<mode>): Ditto.
4331 (*avx512f_scatterdi<mode>): Ditto.
4332 (sseintprefix): Extened with wider modes.
4333 (VEC_GATHER_IDXSI): Ditto.
4334 (VEC_GATHER_IDXDI): Ditto.
4335 (VEC_GATHER_SRCDI): Ditto.
4337 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4338 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4340 * config/arm/t-aprofile: New file.
4341 * config.gcc: Handle --with-multilib-list option.
4343 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4345 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4348 2013-10-15 Richard Biener <rguenther@suse.de>
4350 * tree-loop-distribution.c (build_empty_rdg): Inline into
4352 (rdg_flag_vertex): Inline into single user.
4353 (rdg_flag_vertex_and_dependent): Likewise.
4354 (build_rdg_partition_for_vertex): Remove processed bitmap.
4355 (rdg_build_partitions): Simplify.
4357 2013-10-15 Richard Biener <rguenther@suse.de>
4359 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4360 Restructure forwarding through conversions and copies to
4361 avoid performing copy-propagation the wrong way. Adjust
4362 recursion invocations.
4363 (forward_propagate_addr_expr): Add argument stating if we
4364 are recursing from a single-use.
4365 (ssa_forward_propagate_and_combine): Adjust.
4367 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4369 * dumpfile.h (gcc::dump_manager): New class, to hold state
4370 relating to dumpfile management.
4371 (get_dump_file_name): Remove in favor of method of dump_manager.
4372 (dump_initialized_p): Likewise.
4373 (dump_start): Likewise.
4374 (dump_finish): Likewise.
4375 (dump_switch_p): Likewise.
4376 (dump_register): Likewise.
4377 (get_dump_file_info): Likewise.
4378 * context.c (gcc::context::context): Construct the dump_manager
4380 * context.h (gcc::context::get_dumps): New.
4381 (gcc::context::m_dumps): New.
4382 * coverage.c (coverage_init): Port to dump_manager API.
4383 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4384 (extra_dump_files_in_use): Likewise.
4385 (extra_dump_files_alloced): Likewise.
4386 (gcc::dump_manager::dump_manager): New.
4387 (dump_register): Convert to...
4388 (gcc::dump_manager::dump_register): ...method, replacing
4389 function-static next_dump with m_next_dump field.
4390 (get_dump_file_info): Convert to...
4391 (gcc::dump_manager::get_dump_file_info): ...method.
4392 (get_dump_file_name): Convert to...
4393 (gcc::dump_manager::get_dump_file_name): ...method.
4394 (dump_start): Convert to...
4395 (gcc::dump_manager::dump_start): ...method.
4396 (dump_finish): Convert to...
4397 (gcc::dump_manager::dump_finish): ...method.
4398 (dump_begin): Replace body with...
4399 (gcc::dump_manager::dump_begin): ...new method.
4400 (dump_phase_enabled_p): Convert to...
4401 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4402 (dump_phase_enabled_p): Convert to...
4403 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4404 (dump_initialized_p): Convert to...
4405 (gcc::dump_manager::dump_initialized_p): ...method.
4406 (dump_flag_name): Replace body with...
4407 (gcc::dump_manager::dump_flag_name): ...new method.
4408 (dump_enable_all): Convert to...
4409 (gcc::dump_manager::dump_enable_all): ...new method.
4410 (opt_info_enable_passes): Convert to...
4411 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4412 (dump_switch_p_1): Convert to...
4413 (gcc::dump_manager::dump_switch_p_1): ...new method.
4414 (dump_switch_p): Convert to...
4415 (gcc::dump_manager::dump_switch_p): ...new method.
4416 (opt_info_switch_p): Port to dump_manager API.
4417 (enable_rtl_dump_file): Likewise.
4418 * opts-global.c (handle_common_deferred_options): Port to new
4420 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4421 (pass_manager::register_one_dump_file): Likewise.
4422 (pass_manager::register_pass): Likewise.
4423 (pass_init_dump_file): Likewise.
4424 (pass_fini_dump_file): Likewise.
4425 * statistics.c (statistics_early_init): Likewise.
4427 2013-10-14 Richard Biener <rguenther@suse.de>
4429 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4430 iterative_hash_canonical_type, gimple_canonical_type_hash,
4431 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4432 gimple_register_canonical_type, print_gimple_types_stats,
4433 free_gimple_type_tables): Move to lto/lto.c
4434 (gt-gimple.h): Do not include.
4435 * gimple.h (gimple_register_canonical_type,
4436 print_gimple_types_stats, free_gimple_type_tables): Remove.
4437 * Makefile.in (GTFILES): Remove gimple.c.
4439 2013-10-14 Travis Snoozy <quandary@remstate.com>
4442 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4443 scanning for msp430x targets.
4445 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4448 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4449 (registers_ok_for_ldd_peep): Move around.
4450 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4451 * config/sparc/sparc.md (widening peepholes): Use it.
4453 2013-10-14 Richard Biener <rguenther@suse.de>
4457 * gimple.c (iterative_hash_canonical_type): Make sure to
4458 record the hash into the correct hashtable slot.
4460 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4462 PR rtl-optimization/58662
4463 * combine.c (try_combine): Take into account death nodes on I2 when
4464 splitting a PARALLEL of two independent SETs. Fix dump message.
4466 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4469 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4470 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4471 * config/sh/t-sh (sh_treg_combine.o): New entry.
4472 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4473 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4474 (register_sh_passes): New function. Register sh_treg_combine pass.
4475 (sh_option_override): Invoke it.
4476 (sh_canonicalize_comparison): Handle op0_preserve_value.
4477 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4478 opportunities. Canonicalize branch condition.
4479 (nott): Allow only if pseudos can be created for non-SH2A.
4481 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4484 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4485 (ix86_expand_movmem): Replace copy_addr_to_reg with
4486 ix86_copy_addr_to_reg.
4487 (ix86_expand_setmem): Likewise.
4489 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4491 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4492 provided by ix86_fp_compare_mode instead of CCFPUmode.
4494 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4496 * config/aarch64/arm_neon.h
4497 (vtbx<1,3>_<psu>8): Fix register constriants.
4499 2013-10-11 Jeff Law <law@redhat.com>
4501 PR tree-optimization/58640
4502 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4503 threading paths that cross over two loop entry points.
4505 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4507 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4508 handle vector float as well.
4509 (*vsx_le_perm_load_v4si): Likewise.
4510 (*vsx_le_perm_store_v2di): Likewise.
4511 (*vsx_le_perm_store_v4si): Likewise.
4513 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4515 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4516 directly to circumvent subtract from splat{31} workaround.
4517 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4519 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4520 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4521 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4522 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4523 separate big and little endian logic.
4524 (*altivec_vperm_<mode>_internal): New define_insn.
4525 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4526 separate big and little endian logic.
4527 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4528 (vec_permv16qi): Add little endian logic.
4530 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4532 * doc/extend.texi (returns_nonnull): Remove arguments.
4534 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4535 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4536 Sergey Lega <sergey.s.lega@intel.com>
4537 Anna Tikhonova <anna.tikhonova@intel.com>
4538 Ilya Tocar <ilya.tocar@intel.com>
4539 Andrey Turetskiy <andrey.turetskiy@intel.com>
4540 Ilya Verbin <ilya.verbin@intel.com>
4541 Kirill Yukhin <kirill.yukhin@intel.com>
4542 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4544 * config/i386/sse.md (VI48F_256_512): New.
4545 (avx2_permvar<mode>): Change to ...
4546 (<avx2_avx512f>_permvar<mode>): This.
4548 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4549 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4550 Sergey Lega <sergey.s.lega@intel.com>
4551 Anna Tikhonova <anna.tikhonova@intel.com>
4552 Ilya Tocar <ilya.tocar@intel.com>
4553 Andrey Turetskiy <andrey.turetskiy@intel.com>
4554 Ilya Verbin <ilya.verbin@intel.com>
4555 Kirill Yukhin <kirill.yukhin@intel.com>
4556 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4558 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4559 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4560 * config/i386/sse.md (VI4_AVX): New.
4561 (sf2simodelower): Ditto.
4562 (sse2_cvtps2dq): Change to ...
4563 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4565 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4566 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4567 Sergey Lega <sergey.s.lega@intel.com>
4568 Anna Tikhonova <anna.tikhonova@intel.com>
4569 Ilya Tocar <ilya.tocar@intel.com>
4570 Andrey Turetskiy <andrey.turetskiy@intel.com>
4571 Ilya Verbin <ilya.verbin@intel.com>
4572 Kirill Yukhin <kirill.yukhin@intel.com>
4573 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4575 * config/i386/sse.md (V_512): New.
4577 (vcond<V_512:mode><VF_512:mode>): Ditto.
4578 (vcond<V_512:mode><VI_512:mode>): Ditto.
4579 (vcondu<V_512:mode><VI_512:mode>): Ditto.
4581 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4582 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4583 Sergey Lega <sergey.s.lega@intel.com>
4584 Anna Tikhonova <anna.tikhonova@intel.com>
4585 Ilya Tocar <ilya.tocar@intel.com>
4586 Andrey Turetskiy <andrey.turetskiy@intel.com>
4587 Ilya Verbin <ilya.verbin@intel.com>
4588 Kirill Yukhin <kirill.yukhin@intel.com>
4589 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4591 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4592 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4594 (fma<mode>4): Removed condition.
4595 (fms<mode>4): Ditto.
4596 (fnma<mode>4): Ditto.
4597 (fnms<mode>4): Ditto.
4598 (fma4i_fmadd_<mode>): Ditto.
4599 (*fma_fmadd_<mode>): Ditto.
4600 (*fma_fmsub_<mode>): Ditto.
4601 (*fma_fnmadd_<mode>): Ditto.
4602 (*fma_fnmsub_<mode>): Ditto.
4603 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4604 (*fma_fmaddsub_<mode>): Ditto.
4605 (*fma_fmsubadd_<mode>): Ditto.
4606 (*fmai_fmadd_<mode>): Ditto.
4607 (*fmai_fmsub_<mode>): Ditto.
4608 (*fmai_fnmadd_<mode>): Ditto.
4609 (*fmai_fnmsub_<mode>): Ditto.
4611 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4612 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4613 Sergey Lega <sergey.s.lega@intel.com>
4614 Anna Tikhonova <anna.tikhonova@intel.com>
4615 Ilya Tocar <ilya.tocar@intel.com>
4616 Andrey Turetskiy <andrey.turetskiy@intel.com>
4617 Ilya Verbin <ilya.verbin@intel.com>
4618 Kirill Yukhin <kirill.yukhin@intel.com>
4619 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4621 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4622 (VI124_256): Changed to ...
4623 (VI124_256_48_512): This.
4624 (ssepackmode): Extended with wider modes.
4625 (<code><mode>3): Changed iterator.
4626 (*avx2_<code><mode>3): Ditto.
4627 (vec_pack_trunc_<mode>): Ditto.
4629 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4630 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4631 Sergey Lega <sergey.s.lega@intel.com>
4632 Anna Tikhonova <anna.tikhonova@intel.com>
4633 Ilya Tocar <ilya.tocar@intel.com>
4634 Andrey Turetskiy <andrey.turetskiy@intel.com>
4635 Ilya Verbin <ilya.verbin@intel.com>
4636 Kirill Yukhin <kirill.yukhin@intel.com>
4637 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4639 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
4640 (VI8F_256_512): Ditto.
4641 (abs<mode>2): Changed iterator.
4642 (avx2_perm<mode>): Changed to ...
4643 (<avx2_avx512f>_perm<mode>): This.
4644 (avx2_perm<mode>_1): Changed to ...
4645 (<avx2_avx512f>_perm<mode>_1): This.
4647 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4648 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4649 Sergey Lega <sergey.s.lega@intel.com>
4650 Anna Tikhonova <anna.tikhonova@intel.com>
4651 Ilya Tocar <ilya.tocar@intel.com>
4652 Andrey Turetskiy <andrey.turetskiy@intel.com>
4653 Ilya Verbin <ilya.verbin@intel.com>
4654 Kirill Yukhin <kirill.yukhin@intel.com>
4655 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4657 * config/i386/sse.md (VI48_AVX512F): New.
4658 (VI48_AVX2): Changed to ...
4659 (VI48_AVX2_48_AVX512F): This.
4660 (avx2_ashrv<mode>): Changed to ...
4661 (<avx2_avx512f>_ashrv<mode>): This.
4662 (avx2_<shift_insn>v<mode>): Changed to ...
4663 (<avx2_avx512f>_<shift_insn>v<mode>): This.
4665 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4666 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4667 Sergey Lega <sergey.s.lega@intel.com>
4668 Anna Tikhonova <anna.tikhonova@intel.com>
4669 Ilya Tocar <ilya.tocar@intel.com>
4670 Andrey Turetskiy <andrey.turetskiy@intel.com>
4671 Ilya Verbin <ilya.verbin@intel.com>
4672 Kirill Yukhin <kirill.yukhin@intel.com>
4673 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4675 * config/i386/sse.md (VI4_AVX512F): New.
4676 (VI8_AVX2_AVX512F): Ditto.
4677 (mul<mode>3): Extended with wider modes.
4678 (*<sse4_1_avx2>_mul<mode>3): Ditto.
4679 (mul<mode>3): Ditto.
4680 (vec_widen_<s>mult_odd_<mode>): Ditto.
4682 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4683 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4684 Sergey Lega <sergey.s.lega@intel.com>
4685 Anna Tikhonova <anna.tikhonova@intel.com>
4686 Ilya Tocar <ilya.tocar@intel.com>
4687 Andrey Turetskiy <andrey.turetskiy@intel.com>
4688 Ilya Verbin <ilya.verbin@intel.com>
4689 Kirill Yukhin <kirill.yukhin@intel.com>
4690 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4692 * config/i386/sse.md (VI2_AVX512F): New.
4693 (VI124_AVX512F): Ditto.
4694 (sseunpackmode): Extended with wider modes.
4695 (sseunpackfltmode): Ditto.
4696 (vec_unpacks_float_hi_<mode>): Ditto.
4697 (vec_unpacks_float_lo_<mode>): Ditto.
4698 (vec_unpacku_float_hi_<mode>): Ditto.
4699 (vec_unpacku_float_lo_<mode>): Ditto.
4700 (vec_unpacks_lo_<mode>): Ditto.
4701 (vec_unpacks_hi_<mode>): Ditto.
4702 (vec_unpacku_lo_<mode>): Ditto.
4703 (vec_unpacku_hi_<mode>): Ditto.
4705 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4706 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4707 Sergey Lega <sergey.s.lega@intel.com>
4708 Anna Tikhonova <anna.tikhonova@intel.com>
4709 Ilya Tocar <ilya.tocar@intel.com>
4710 Andrey Turetskiy <andrey.turetskiy@intel.com>
4711 Ilya Verbin <ilya.verbin@intel.com>
4712 Kirill Yukhin <kirill.yukhin@intel.com>
4713 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4715 * config/i386/i386.md (multdiv): New.
4716 (multdiv_mnemonic): Ditto.
4717 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
4718 (<sse>_vm<multdiv_mnemonic><mode>3): This.
4719 (<sse>_vmdiv<mode>3): Removed.
4721 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4722 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4723 Sergey Lega <sergey.s.lega@intel.com>
4724 Anna Tikhonova <anna.tikhonova@intel.com>
4725 Ilya Tocar <ilya.tocar@intel.com>
4726 Andrey Turetskiy <andrey.turetskiy@intel.com>
4727 Ilya Verbin <ilya.verbin@intel.com>
4728 Kirill Yukhin <kirill.yukhin@intel.com>
4729 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4731 * config/i386/sse.md (V): Extended with wider modes.
4733 (ssehalfvecmode): Ditto.
4735 (ssepackfltmode): Ditto.
4736 (avx_vec_concat<mode>): Ditto.
4737 (V_256_512): New iterator.
4738 (VF2_512_256): Ditto.
4739 (si2dfmode): New attribute.
4740 (si2dfmodelower): Ditto.
4742 (concat_tg_mode): Ditto.
4743 (floatv4siv4df2): Changed to ...
4744 (float<si2dfmodelower><mode>2): This.
4745 (avx_cvtps2pd256): Changed to ...
4746 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
4747 (vec_pack_trunc_v4df): Changed to ...
4748 (vec_pack_trunc_<mode>): This.
4749 (avx_vpermil<mode>): Changed to ...
4750 (<sse2_avx_avx512f>_vpermil<mode>): This.
4751 (<fixsuffix>fix_truncv8dfv8si2): New.
4752 (vec_pack_sfix_trunc_v8df): Ditto.
4753 (avx512f_rndscale<mode>): Ditto.
4754 (avx512f_roundpd512): Ditto.
4755 (vec_pack_ufix_trunc_<mode>): Updated iterator.
4757 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4758 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4759 Sergey Lega <sergey.s.lega@intel.com>
4760 Anna Tikhonova <anna.tikhonova@intel.com>
4761 Ilya Tocar <ilya.tocar@intel.com>
4762 Andrey Turetskiy <andrey.turetskiy@intel.com>
4763 Ilya Verbin <ilya.verbin@intel.com>
4764 Kirill Yukhin <kirill.yukhin@intel.com>
4765 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4767 * config/i386/i386.md (any_fix): New iterator.
4768 (fixsuffix): New attribute.
4769 * config/i386/sse.md (VF1): Extened with wider modes.
4773 (sseintvecmodelower): Ditto.
4774 (ssescalarmode): Ditto.
4775 (ssescalarnum): Ditto.
4777 (ssexmmmode): Ditto.
4778 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
4779 (<sse>_rcp<mode>2): Change iterator.
4780 (rsqrt<mode>2): Ditto.
4781 (<sse>_rsqrt<mode>2): Ditto.
4782 (avx2_vec_dup<mode>): Ditto.
4783 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
4784 (round<mode>2_sfix): Ditto.
4785 (avx2_pbroadcast<mode>): Ditto.
4786 (*andnot<mode>3): Handle XI mode.
4787 (*<code><mode>3): Ditto.
4788 (AVXTOSSEMODE): Removed.
4789 (avx_vpermil<mode>): Changed to ...
4790 (<sse2_avx_avx512f>_vpermil<mode>): This.
4792 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4793 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4794 Sergey Lega <sergey.s.lega@intel.com>
4795 Anna Tikhonova <anna.tikhonova@intel.com>
4796 Ilya Tocar <ilya.tocar@intel.com>
4797 Andrey Turetskiy <andrey.turetskiy@intel.com>
4798 Ilya Verbin <ilya.verbin@intel.com>
4799 Kirill Yukhin <kirill.yukhin@intel.com>
4800 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4802 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
4803 (<sse>_comi): Ditto.
4804 (<sse>_ucomi): Ditto.
4805 (sse_cvtss2siq_2): Ditto.
4806 (sse2_cvtsd2si): Ditto.
4807 (sse2_cvtsd2siq): Ditto.
4808 (sse2_cvttsd2si): Ditto.
4809 (sse2_cvttsd2siq): Ditto.
4810 (<shift_insn><mode>3): Ditto.
4811 (sse2_cvtsi2sdq): Update constraint and prefix.
4812 (sse_cvtsi2ss): Update prefix.
4813 (sse_cvtsi2ssq): Ditto.
4815 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4817 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4818 ignore internal calls.
4820 2013-10-11 Richard Biener <rguenther@suse.de>
4822 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4823 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4824 TREE_OVERFLOW is set.
4826 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4828 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4830 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4831 * gimple.h: Likewise.
4833 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4835 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4838 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4840 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4841 * configure: Regenerate.
4843 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4845 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4846 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4847 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4848 similarly pp_right_paren.
4849 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4850 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
4851 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
4852 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
4854 (omp_declare_simd_clauses_equal,
4855 omp_remove_redundant_declare_simd_attrs): New functions.
4856 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
4857 (walk_tree_1): Handle new OpenMP 4.0 clauses.
4858 * tree.h (OMP_LOOP_CHECK): Define.
4859 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
4860 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
4861 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
4862 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
4863 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
4864 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
4865 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
4866 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
4867 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
4868 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
4869 OMP_CLAUSE_SIMDLEN_EXPR): Define.
4870 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
4871 (omp_remove_redundant_declare_simd_attrs): New prototype.
4872 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
4873 GIMPLE_OMP_TEAMS): New codes.
4874 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
4875 * omp-low.c (struct omp_context): Add cancel_label and cancellable
4877 (target_nesting_level): New variable.
4878 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
4879 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
4880 for collapse > 1 static schedule unless ordered.
4881 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
4882 (determine_parallel_type): Adjust get_ws_args_for caller.
4883 (install_var_field): Handle mask & 4 for double indirection.
4884 (scan_sharing_clauses): Ignore shared clause on teams construct.
4885 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
4886 (create_omp_child_function): If inside target or declare target
4887 constructs, set "omp declare target" attribute on the child function.
4888 (find_combined_for): New function.
4889 (scan_omp_parallel): Handle combined loops.
4890 (scan_omp_target, scan_omp_teams): New functions.
4891 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
4892 restrictions and set ctx->cancellable for cancellable constructs.
4893 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
4894 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
4895 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
4896 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
4897 (omp_clause_aligned_alignment): New function.
4898 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
4899 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
4900 on teams constructs. Handle user defined reductions and new
4902 (lower_reduction_clauses): Don't set placeholder to address of ref
4903 if it has already the right type.
4904 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
4905 (expand_parallel_call): Use the new non-_start suffixed builtins,
4906 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
4907 and GOMP_parallel_end after the call.
4908 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
4909 (expand_omp_for_init_counts): Handle combined loops.
4910 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
4912 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
4913 end of cancellable loops.
4914 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
4915 Likewise. Handle collapse > 1 loops.
4916 (expand_omp_simd): Handle combined loops.
4917 (expand_omp_for): Add inner_stmt argument, adjust callers of
4918 expand_omp_for* functions, use expand_omp_for_static*chunk even
4919 for collapse > 1 unless ordered.
4920 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
4921 of cancellable sections.
4922 (expand_omp_single): Remove need_barrier variable, just rely on
4923 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
4924 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
4925 (expand_omp_atomic_load, expand_omp_atomic_store,
4926 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
4927 (expand_omp_target): New function.
4928 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
4929 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
4930 (build_omp_regions_1): Immediately close region for
4931 GF_OMP_TARGET_KIND_UPDATE.
4932 (maybe_add_implicit_barrier_cancel): New function.
4933 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
4935 (lower_omp_single): Likewise. Add clobber after the barrier.
4936 (lower_omp_taskgroup): New function.
4937 (lower_omp_for): Handle combined loops. Adjust
4938 lower_rec_input_clauses caller. Handle cancellation.
4939 (lower_depend_clauses): New function.
4940 (lower_omp_taskreg): Lower depend clauses. Adjust
4941 lower_rec_input_clauses caller. Add clobber after the call. Handle
4943 (lower_omp_target, lower_omp_teams): New functions.
4944 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
4945 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
4946 and GOMP_cancellation_point calls.
4947 (lower_omp): Fold stmts inside of target region.
4948 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
4949 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4950 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
4951 (BT_FN_VOID_OMPFN_PTR_UINT,
4952 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
4953 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
4954 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
4955 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
4956 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
4957 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
4958 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
4959 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
4960 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
4961 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
4962 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
4963 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
4964 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
4965 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
4966 BUILT_IN_GOMP_PARALLEL_END.
4967 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
4968 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4969 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
4970 GF_OMP_FOR_KIND_DISTRIBUTE.
4971 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
4972 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
4973 (dump_gimple_omp_return): Print lhs if it has any.
4974 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
4975 gimple_omp_atomic_seq_cst_p.
4976 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
4977 and GIMPLE_OMP_TEAMS.
4978 * langhooks.c (lhd_omp_mappable_type): New function.
4979 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
4980 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
4982 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
4983 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
4984 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
4985 (struct gimplify_omp_ctx): Add combined_loop field.
4986 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
4987 on stmts inside of target region.
4988 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
4989 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
4990 ORT_TARGET and ORT_TARGET_DATA.
4991 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
4993 (omp_notice_threadprivate_variable): Complain about threadprivate
4994 variables in target region.
4995 (omp_notice_variable): Complain about vars with non-mappable type
4996 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
4997 (omp_check_private): Ignore ORT_TARGET* regions.
4998 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
4999 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
5000 (find_combined_omp_for): New function.
5001 (gimplify_omp_for): Handle gimplification of combined loops.
5002 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
5004 (gimplify_omp_target_update): New function.
5005 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
5006 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
5007 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
5008 (gimplify_body): If fndecl has "omp declare target" attribute, add
5009 implicit ORT_TARGET context around it.
5010 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
5011 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
5012 * tree-nested.c (convert_nonlocal_reference_stmt,
5013 convert_local_reference_stmt, convert_gimple_call): Handle
5014 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5015 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
5016 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5017 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5018 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5019 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5020 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5021 BUILT_IN_GOMP_LOOP_END_CANCEL,
5022 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5023 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5024 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5025 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5026 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5027 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5028 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5029 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5030 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5031 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5032 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5033 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5034 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5035 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5036 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5037 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5038 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5039 gimple_build_omp_teams): New functions.
5040 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5041 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5042 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5043 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5044 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5045 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5046 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5047 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5048 GF_OMP_ATOMIC_SEQ_CST): New.
5049 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5050 gimple_build_omp_teams): New prototypes.
5051 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5052 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5053 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5054 GIMPLE_OMP_SINGLE as end of range.
5055 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5056 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5057 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5058 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5059 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5060 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5061 gimple_omp_target_kind, gimple_omp_target_set_kind,
5062 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5063 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5064 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5065 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5066 gimple_omp_teams_set_clauses): New inlines.
5067 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5068 and GIMPLE_OMP_TASKGROUP.
5069 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5070 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5071 enum omp_clause_proc_bind_kind): New.
5072 (union omp_clause_subcode): Add depend_kind, map_kind and
5073 proc_bind_kind fields.
5074 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5075 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5076 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5077 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5078 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5080 2013-10-10 Teresa Johnson <tejohnson@google.com>
5082 * predict.c (tree_estimate_probability): Add new parameter
5083 for estimate_bb_frequencies.
5084 (estimate_bb_frequencies): Add new parameter to force estimation.
5085 (rebuild_frequencies): When max frequency in function is small,
5086 recompute counts from frequencies.
5087 * predict.h (estimate_bb_frequencies): New parameter.
5089 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5091 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5092 optimizations are disabled.
5094 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5096 * coverage.c (coverage_finish): Fix leak of da_file_name.
5098 2013-10-10 Jan Hubicka <jh@suse.cz>
5100 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5101 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5104 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5107 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5108 if any labels are in FALLTHRU_BB, use a special label emitted
5109 immediately after the asm goto insn rather than label_rtx
5111 (expand_asm_stmt): Adjust caller.
5112 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5113 edge if the last insn in predecessor is a jump with single successor,
5114 but it isn't simplejump_p.
5116 2013-10-10 Richard Biener <rguenther@suse.de>
5118 PR tree-optimization/58656
5119 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5121 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5123 * gimplify.c: Include expr.h and tm_p.h for targets with special
5124 va-arg padding requirements.
5126 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5128 * tree-flow.h: Move some prototypes to gimple.h.
5129 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5130 * gimple.h: Relocate some prototypes from tree-flow.h
5131 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5133 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5134 (build_va_arg_indirect_ref): Relocate and make static.
5135 (std_gimplify_va_arg_expr): Relocate here.
5136 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5137 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5139 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5141 * doc/md.texi: Document the mnemonic attribute.
5143 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5146 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5147 (set_attr_alternative x ...) when searching for user defined
5150 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5152 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5154 2013-10-09 Easwaran Raman <eraman@google.com>
5156 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5157 * cfgexpand.c (defer_stack_allocation): ...use here
5158 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5160 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
5162 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
5163 (operand_equal_for_value_replacement): New function, extracted from
5164 value_replacement and enhanced to catch more cases.
5165 (value_replacement): Use operand_equal_for_value_replacement.
5167 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5169 * loop-doloop.c (doloop_modify, doloop_optimize): Use
5170 get_max_loop_iterations.
5172 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5174 * config/arm/aarch-common.c (arm_early_load_addr_dep):
5175 Place comment above function.
5177 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5179 * tree-flow.h: Remove all remaining prototypes, enums and structs that
5180 are not related to tree-cfg.c.
5181 * tree-ssa-address.h: New file. Relocate prototypes.
5182 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
5183 (addr_for_mem_ref): New. Combine call to get_address_description and
5184 return addr_for_mem_ref.
5185 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
5186 * tree-ssa-live.h: Adjust prototypes.
5187 * passes.c: Include tree-ssa-live.h.
5188 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
5189 * graphite.c (graphite_transform_loops): Make static.
5190 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5191 make_pass_graphite, pass_data_graphite_transforms,
5192 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
5193 * ipa-pure-const.c (warn_function_noreturn): Make static.
5194 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5195 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5196 Relocate from tree-cfg.c
5197 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
5199 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5200 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5201 Move to ipa-pure-const.c.
5202 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
5203 Relocate from tree-optimize.c.
5204 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
5205 make_pass_fixup_cfg): Move to tree-cfg.c.
5206 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
5207 Relocate some prototypes.
5208 * tree-data-ref.h (tree_check_data_deps) Add prototype.
5209 * tree-dump.c (dump_function_to_file): Remove prototype.
5210 Add tree-flow.h to the include file.
5211 * tree-dump.h: Remove prototype.
5212 * tree-parloops.h: New File. Add prototypes.
5213 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5214 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5215 from tree-ssa-loop.c.
5216 * tree-predcom.c (run_tree_predictive_commoning,
5217 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5218 Relocate here from tree-ssa-loop.c.
5219 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5220 ssa_name_values.release ().
5221 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5222 (ssa_name_values): Relocate from tree-flow.h.
5223 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5224 * tree-ssa-loop.c (run_tree_predictive_commoning,
5225 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5226 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5227 make_pass_graphite, pass_data_graphite_transforms,
5228 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5229 tree_parallelize_loops, pass_data_parallelize_loops,
5230 make_pass_parallelize_loops): Move to other files.
5231 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5233 * tree.h: Remove prototypes from tree-address.c.
5235 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5237 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5238 (struct int_tree_map): Move to tree-hasher.h
5239 (SCALE, LABEL, PERCENT): Move to gimple.h
5240 * tree-flow-inline.h: Delete. Move functions to other files.
5241 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5242 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5243 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5244 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5245 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5246 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5247 * tree-hasher.h: Don't include tree-flow.h.
5248 (struct int_tree_map): Relocate from tree-flow.h.
5249 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5250 tree-flow-inline.h and make static.
5251 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5253 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5254 tree-flow-inline.h and make static.
5255 * tree.h (is_global_var, may_be_aliased): Relocate from
5257 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5258 * value-prof.c: No longer include tree-flow-inline.h.
5259 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5261 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5263 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5264 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5265 (enum move_pos): Move to tree-ssa-loop-im.h
5266 * cfgloop.h: Move some prototypes.
5267 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5268 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5269 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5270 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5271 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5272 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5273 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5274 (enum move_pos): Relocate here.
5275 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5277 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5278 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5279 make_pass_lim): Relocate here from tree-ssa-loop.c.
5280 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5282 (loop_edge_to_cancel, unloop_loops): Make static.
5283 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5284 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5285 (tree_complete_unroll, gate_tree_complete_unroll,
5286 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5287 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5288 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5289 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5290 (stmt_invariant_in_loop_p): Remove unused function.
5291 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5292 * tree-ssa-loop-manip.h: New file. Add prototypes.
5293 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5294 (gcov_type_to_double_int): Move to cfgloop.h.
5295 (double_int_cmp, bound_index,
5296 estimate_numbers_of_iterations_loop): Make static.
5297 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5298 (max_loop_iterations): Factor out get_max_loop_iterations.
5299 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5301 * tree-ssa-loop-niter.h: New file. Add prototypes.
5302 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5303 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5304 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5305 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5306 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5307 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5308 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5309 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5310 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5311 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5312 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5313 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5314 pass_data_complete_unroll, make_pass_complete_unroll,
5315 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5316 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5317 tree-ssa-loop-ivcanon.c.
5318 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5319 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5320 tree-ssa-loop-prefetch.c.
5321 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5323 (get_lsm_tmp_name): Relocate and add suffix parameter.
5324 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5325 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5326 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5327 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5328 (get_estimated_loop_iterations): Factor out accessor from
5329 estimated_loop_iterations in tree-ssa-loop-niter.c.
5330 (get_max_loop_iterations): Factor out accessor from
5331 _max_loop_iterations in tree-ssa-niter.c.
5332 * loop-unroll.c (decide_unroll_constant_iterations,
5333 decide_unroll_runtime_iterations, decide_peel_simple,
5334 decide_unroll_stupid): Use new get_* accessors.
5336 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5338 PR tree-optimization/20318
5339 * doc/extend.texi (returns_nonnull): New function attribute.
5340 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5342 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5343 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5345 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5348 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5349 false if both components are bitfields.
5351 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5353 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5354 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5355 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5356 * config/aarch64/aarch64.h
5357 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5358 * config/aarch64/aarch64-simd-builtins.def
5359 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5361 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5363 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5365 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5367 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5369 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5371 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5372 (vneg_f64): New intrinsic.
5373 (vneg_s8): Asm replaced with C.
5374 (vneg_s16): Likewise.
5375 (vneg_s32): Likewise.
5376 (vneg_s64): New intrinsic.
5377 (vnegq_f32): Asm replaced with C.
5378 (vnegq_f64): Likewise.
5379 (vnegq_s8): Likewise.
5380 (vnegq_s16): Likewise.
5381 (vnegq_s32): Likewise.
5382 (vnegq_s64): Likewise.
5384 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5386 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5388 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5390 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5391 packed stack special handling.
5392 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5393 back to fixed stack slots for FPRs saved due to stdarg.
5395 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5397 * config/s390/s390.c (s390_frame_info): Restructure function.
5399 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5401 * config/s390/s390.c (struct s390_frame_layout): New field
5403 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5404 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5405 regs_ever_clobbered to char*.
5406 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5407 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5408 (s390_register_info): Enable FPR save slots. Move/Copy some
5409 functionality into ...
5410 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5411 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5413 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5414 not imply a stack frame.
5415 (s390_init_frame_layout): Remove variable clobbered_regs.
5416 (s390_update_register_info): Remove function.
5417 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5418 cannot be used for register renaming.
5419 (s390_hard_regno_scratch_ok): New function.
5420 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5421 (s390_initial_elimination_offset): Change offset calculation of
5422 the return address pointer.
5423 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5424 (restore_gprs): Set frame related flag.
5425 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5426 (s390_emit_prologue): Call s390_register_info instead of
5427 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5428 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5429 (s390_optimize_prologue): Call s390_optimize_register_info.
5430 Try to remove also FPR slot save/restore INSNs. Remove frame
5431 related flags from restore INSNs.
5433 2013-10-08 DJ Delorie <dj@redhat.com>
5435 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5438 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5439 avoid conflict with the MI use of %c.
5440 * config/rl78/rl78-real.md: change %c to %C throughout.
5441 * config/rl78/rl78-virt.md: Likewise.
5443 2013-10-08 Jan Hubicka <jh@suse.cz>
5445 * config/i386/i386.c (ix86_option_override_internal): Switch
5446 to SSE math for -ffast-math when target ISA supports SSE2.
5448 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5450 * tree-flow.h: Remove some prototypes.
5451 * tree.h: Remove some protypes, add a couple.
5452 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5453 using_eh_for_cleanups_p): Add interface routines for front ends.
5454 * tree-eh.h: New file. Add protoptyes.
5455 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5456 (add_stmt_to_eh_lp_fn): Make static.
5457 (lower_try_finally): Use new using_eh_for_cleanups_p.
5458 * emit-rtl.c: Include tree-eh.h.
5459 * gimple.h: Include tree-eh.h.
5461 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5463 PR tree-optimization/58480
5464 * tree-vrp.c (infer_nonnull_range): New function.
5465 (infer_value_range): Call infer_nonnull_range.
5467 2013-10-08 Dehao Chen <dehao@google.com>
5469 PR tree-optimization/58619
5470 * tree-inline.c (copy_phis_for_bb): Combine location data
5473 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5476 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5477 RTX_FRAME_RELATED_P on INSN.
5479 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5481 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5482 (altivec_expand_vec_perm_const): Call it.
5484 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5486 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5487 sequences for LE VSX loads and stores at expand time.
5488 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5490 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5491 (rs6000_gen_le_vsx_permute): New.
5492 (rs6000_gen_le_vsx_load): New.
5493 (rs6000_gen_le_vsx_store): New.
5494 (rs6000_gen_le_vsx_move): New.
5495 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5496 (*vsx_le_perm_load_v4si): New.
5497 (*vsx_le_perm_load_v8hi): New.
5498 (*vsx_le_perm_load_v16qi): New.
5499 (*vsx_le_perm_store_v2di): New.
5500 (*vsx_le_perm_store_v4si): New.
5501 (*vsx_le_perm_store_v8hi): New.
5502 (*vsx_le_perm_store_v16qi): New.
5503 (*vsx_xxpermdi2_le_<mode>): New.
5504 (*vsx_xxpermdi4_le_<mode>): New.
5505 (*vsx_xxpermdi8_le_V8HI): New.
5506 (*vsx_xxpermdi16_le_V16QI): New.
5507 (*vsx_lxvd2x2_le_<mode>): New.
5508 (*vsx_lxvd2x4_le_<mode>): New.
5509 (*vsx_lxvd2x8_le_V8HI): New.
5510 (*vsx_lxvd2x16_le_V16QI): New.
5511 (*vsx_stxvd2x2_le_<mode>): New.
5512 (*vsx_stxvd2x4_le_<mode>): New.
5513 (*vsx_stxvd2x8_le_V8HI): New.
5514 (*vsx_stxvd2x16_le_V16QI): New.
5516 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5518 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5520 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5522 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5523 loop to work also for 31bit ABI.
5524 Save the stack pointer for frame_size > 0.
5526 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5528 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5529 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5530 constraint letters from expanders.
5531 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5532 retry count to general_operand.
5533 ("tabort"): Give operand 0 a mode.
5534 ("tabort_1"): Add mode and constraint letter for operand 0.
5535 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5537 2013-10-04 Jeff Law <law@redhat.com>
5539 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5541 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5543 (thread_across_edge): Here. Call it.
5545 2013-10-04 Cary Coutant <ccoutant@google.com>
5547 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5548 discarding a location list expression (or a piece of one).
5550 2013-10-03 Jan Hubicka <jh@suse.cz>
5552 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5553 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5554 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5556 2013-10-03 Jan Hubicka <jh@suse.cz>
5558 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5559 accumulate-outgoing-args when producing unwind info.
5561 2013-10-03 Wei Mi <wmi@google.com>
5563 * lra-constraints.c (insert_move_for_subreg): New function
5564 extracted from simplify_operand_subreg.
5565 (simplify_operand_subreg): Add reload for paradoxical subreg.
5567 2013-10-03 Rong Xu <xur@google.com>
5569 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5570 the candidate of builtin_expect such that we should fix the
5571 size/time estimation.
5572 (estimate_function_body_sizes): Do the acutally size/time fix-up
5575 2013-10-03 Rong Xu <xur@google.com>
5577 * predict.c (tree_predict_by_opcode): Get the probability
5578 for builtin_expect from param builtin_expect_probability.
5579 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5580 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5581 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5583 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5586 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5587 * calls.c (alloca_call_p): Use get_callee_fndecl.
5588 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5589 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5591 (vrp_visit_stmt): Remove duplicated code.
5593 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5595 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5596 ceildf2, btruncdf2, instead of vsx_* name.
5598 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5599 iterators to only do V2DF and V4SF here. Move the DF code to
5600 rs6000.md where it is combined with SF mode. Replace <VSv> with
5601 just 'v' since only vector operations are handled with these insns
5602 after moving the DF support to rs6000.md.
5603 (vsx_sub<mode>3): Likewise.
5604 (vsx_mul<mode>3): Likewise.
5605 (vsx_div<mode>3): Likewise.
5606 (vsx_fre<mode>2): Likewise.
5607 (vsx_neg<mode>2): Likewise.
5608 (vsx_abs<mode>2): Likewise.
5609 (vsx_nabs<mode>2): Likewise.
5610 (vsx_smax<mode>3): Likewise.
5611 (vsx_smin<mode>3): Likewise.
5612 (vsx_sqrt<mode>2): Likewise.
5613 (vsx_rsqrte<mode>2): Likewise.
5614 (vsx_fms<mode>4): Likewise.
5615 (vsx_nfma<mode>4): Likewise.
5616 (vsx_copysign<mode>3): Likewise.
5617 (vsx_btrunc<mode>2): Likewise.
5618 (vsx_floor<mode>2): Likewise.
5619 (vsx_ceil<mode>2): Likewise.
5620 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5621 (vsx_sminsf3): Likewise.
5622 (vsx_fmadf4): Likewise.
5623 (vsx_fmsdf4): Likewise.
5624 (vsx_nfmadf4): Likewise.
5625 (vsx_nfmsdf4): Likewise.
5626 (vsx_cmpdf_internal1): Likewise.
5628 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
5629 simpler to select whether a target has SPE or traditional floating
5630 point support in iterators.
5631 (TARGET_DF_SPE): Likewise.
5632 (TARGET_SF_FPR): Likewise.
5633 (TARGET_DF_FPR): Likewise.
5634 (TARGET_SF_INSN): Macros to say whether floating point support
5635 exists for a given operation for expanders.
5636 (TARGET_DF_INSN): Likewise.
5638 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
5639 combining of SF/DF mode operations, using both traditional and VSX
5647 (abs<mode>2): Combine SF/DF modes using traditional floating point
5648 instructions. Add support for using the upper DF registers with
5649 VSX support, and SF registers with power8-vector support. Update
5650 expanders for operations supported by both the SPE and traditional
5651 floating point units.
5652 (abs<mode>2_fpr): Likewise.
5653 (nabs<mode>2): Likewise.
5654 (nabs<mode>2_fpr): Likewise.
5655 (neg<mode>2): Likewise.
5656 (neg<mode>2_fpr): Likewise.
5657 (add<mode>3): Likewise.
5658 (add<mode>3_fpr): Likewise.
5659 (sub<mode>3): Likewise.
5660 (sub<mode>3_fpr): Likewise.
5661 (mul<mode>3): Likewise.
5662 (mul<mode>3_fpr): Likewise.
5663 (div<mode>3): Likewise.
5664 (div<mode>3_fpr): Likewise.
5665 (sqrt<mode>3): Likewise.
5666 (sqrt<mode>3_fpr): Likewise.
5667 (fre<Fs>): Likewise.
5668 (rsqrt<mode>2): Likewise.
5669 (cmp<mode>_fpr): Likewise.
5670 (smax<mode>3): Likewise.
5671 (smin<mode>3): Likewise.
5672 (smax<mode>3_vsx): Likewise.
5673 (smin<mode>3_vsx): Likewise.
5674 (negsf2): Delete SF operations that are merged with DF.
5681 (fmasf4_fpr): Likewise.
5682 (fmssf4_fpr): Likewise.
5683 (nfmasf4_fpr): Likewise.
5684 (nfmssf4_fpr): Likewise.
5685 (sqrtsf2): Likewise.
5686 (rsqrtsf_internal1): Likewise.
5687 (smaxsf3): Likewise.
5688 (sminsf3): Likewise.
5689 (cmpsf_internal1): Likewise.
5690 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
5691 (negdf2): Delete DF operations that are merged with SF.
5693 (nabsdf2): Likewise.
5699 (rsqrtdf_internal1): Likewise.
5700 (fmadf4_fpr): Likewise.
5701 (fmsdf4_fpr): Likewise.
5702 (nfmadf4_fpr): Likewise.
5703 (nfmsdf4_fpr): Likewise.
5704 (sqrtdf2): Likewise.
5705 (smaxdf3): Likewise.
5706 (smindf3): Likewise.
5707 (cmpdf_internal1): Likewise.
5708 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
5709 (btrunc<mode>2): Delete separate expander, and combine with the
5710 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
5711 (btrunc<mode>2_fpr): Likewise.
5712 (ceil<mode>2): Likewise.
5713 (ceil<mode>2_fpr): Likewise.
5714 (floor<mode>2): Likewise.
5715 (floor<mode>2_fpr): Likewise.
5716 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
5717 Add support for using the upper registers with VSX and
5718 power8-vector. Move insns to be closer to the define_expands. On
5719 VSX systems, prefer the traditional form of FMA over the VSX
5720 version, since the traditional form allows the target not to
5721 overlap with the inputs.
5722 (fms<mode>4_fpr): Likewise.
5723 (nfma<mode>4_fpr): Likewise.
5724 (nfms<mode>4_fpr): Likewise.
5726 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5727 Richard Earnshaw <richard.earnshaw@arm.com>
5729 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
5730 (struct mult_cost_table): Likewise.
5731 (struct mem_cost_table): Likewise.
5732 (struct fp_cost_table): Likewise.
5733 (struct vector_cost_table): Likewise.
5734 (cpu_cost_table): Likewise.
5735 * config/arm/arm.opt (mold-rts-costs): New option.
5736 (mnew-generic-costs): Likewise.
5737 * config/arm/arm.c (generic_extra_costs): New table.
5738 (cortexa15_extra_costs): Likewise.
5739 (arm_slowmul_tune): Use NULL as new costs.
5740 (arm_fastmul_tune): Likewise.
5741 (arm_strongarm_tune): Likewise.
5742 (arm_xscale_tune): Likewise.
5743 (arm_9e_tune): Likewise.
5744 (arm_v6t2_tune): Likewise.
5745 (arm_cortex_a5_tune): Likewise.
5746 (arm_cortex_a9_tune): Likewise.
5747 (arm_v6m_tune): Likewise.
5748 (arm_fa726te_tune): Likewise.
5749 (arm_cortex_a15_tune): Use cortex15_extra_costs.
5750 (arm_cortex_tune): Use generict_extra_costs.
5751 (shifter_op_p): New function.
5752 (arm_unspec_cost): Likewise.
5753 (LIBCALL_COST): Define.
5754 (arm_new_rtx_costs): New function.
5755 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
5756 table is available. Use old costs otherwise unless mnew-generic-costs
5758 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
5759 (cpu_cost_table): Declare.
5761 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
5764 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
5765 (*subs_mul_imm_<mode>)
5766 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
5767 (*sub_<shift>_<mode>)
5768 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
5769 Remove k constraint.
5771 2013-10-03 Ian Bolton <ian.bolton@arm.com>
5773 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
5775 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
5777 2013-10-02 Teresa Johnson <tejohnson@google.com>
5779 * predict.c (probably_never_executed): New function.
5780 (probably_never_executed_bb_p): Invoke probably_never_executed.
5781 (probably_never_executed_edge_p): Ditto.
5782 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
5783 Treat profile insanities conservatively.
5785 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
5787 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
5789 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
5791 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
5792 use it. Use smaller increase for scratch. Don't increase reject
5793 for early clobber scratch.
5794 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
5795 setting eliminated regs except setting fp from hfp.
5796 (lra_eliminate): Check lra_insn_recog_data on NULL.
5798 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5801 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
5802 setting -mvsx-timode by default until the underlying problem is fixed.
5803 (RS6000_CPU, power7 defaults): Likewise.
5805 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5807 * config/x-linux (host-linux.o): Remove header dependencies.
5808 Use $(COMPILE) and $(POSTCOMPILE).
5809 * config/t-linux-android (linux-android.o): Ditto.
5811 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5813 * Makefile.in (expmed.o-warn): Remove.
5815 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5817 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5818 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5820 2013-10-02 Teresa Johnson <tejohnson@google.com>
5822 * dojump.c (do_jump_1): Divide probability between
5823 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5825 2013-10-02 Tom Tromey <tromey@redhat.com>
5827 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5829 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5831 * tree-flow.h: Remove some prototypes.
5832 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5833 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5834 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5835 mark_virtual_phi_result_for_renaming): Relocate here.
5836 * tree-into-ssa.h: Add prototypes.
5837 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5838 single_pred_before_succ_order.
5839 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5840 (nonfreeing_call_p) Move to gimple.c.
5841 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5843 * basic-block.h (single_pred_before_succ_order): Add prototype.
5844 * gimple.c (nonfreeing_call_p): Relocate here.
5845 * gimple.h: Add prototype.
5846 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5847 * tree-ssa-dom.h: New file. Relocate prototypes here.
5848 * tree-ssa.h: Include tree-ssa-dom.h.
5850 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5852 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
5853 Use $(COMPILE) and $(POSTCOMPILE).
5855 * config/alpha/x-alpha (driver-alpha.o): Ditto.
5857 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5859 * tree-flow.h: Remove some prototypes.
5860 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
5861 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5862 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
5863 propagate_tree_value*): Move from here to...
5864 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
5865 propagate_tree_value*): Relocate here.
5866 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5867 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
5868 * gimple-fold.c: Remove gimple-fold.h from include list.
5869 * tree-vrp.c: Remove gimple-fold.h from include list.
5870 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
5871 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
5872 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
5873 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
5874 * sese.c: Add tree-ssa-propagate.h to include list.
5876 2013-10-02 Richard Biener <rguenther@suse.de>
5878 * tree-loop-distribution.c: Include tree-vectorizer.h for
5880 (enum partition_kind): Remove PKIND_REDUCTION.
5881 (struct partition_s): Remove has_writes member, add reduction_p member.
5882 (partition_alloc): Adjust.
5883 (partition_builtin_p): Likewise.
5884 (partition_has_writes): Remove.
5885 (partition_reduction_p): New function.
5886 (partition_merge_into): Likewise.
5887 (generate_code_for_partition): Commonize builtin partition
5889 (rdg_cannot_recompute_vertex_p): Remove.
5890 (already_processed_vertex_p): Likewise.
5891 (rdg_flag_vertex): Do not set has_writes.
5892 (classify_partition): Adjust.
5893 (rdg_build_partitions): Do not set has_writes, treat all
5894 partitions as useful.
5895 (distribute_loop): Record number of library calls generated. Adjust.
5896 (tree_loop_distribution): Report number of loops and library
5897 calls generated as opt-info.
5899 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5901 * tree-flow.h: Include new .h files. Move prototypes.
5902 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
5903 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
5904 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
5905 * tree-pretty-print.h: Add prototypes from tree-flow.h.
5906 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
5907 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
5908 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
5909 * tree.h (get_ref_base_and_extent): Move prototype out.
5910 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
5912 * gimple-low.h: New File. Add prototypes from tree-flow.h.
5913 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
5914 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
5915 * tree-scalar-evolution.c: Include tree.h.
5916 * sese.c: Include tree.h.
5917 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
5918 * dwarf2out.c: Include tree-dfa.h.
5919 * tree-chrec.c: Include tree.h.
5920 * tree-data-ref.c: Include tree.h.
5922 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
5924 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
5927 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5929 * config/t-sol2 (sol2-c.o): Remove header dependencies.
5930 Use $(COMPILE) and $(POSTCOMPILE).
5931 (sol2-cxx.o): Likewise.
5932 (sol2-stubs.o): Likewise.
5934 * config/x-solaris (host-solaris.o): Likewise.
5936 * config/sparc/t-sparc (sparc.o): Remove.
5937 (sparc-c.o): Remove header dependencies.
5938 Use $(COMPILE) and $(POSTCOMPILE).
5939 * config/sparc/x-sparc: Likewise.
5941 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
5943 * config/arc/arc-opts.h: Add 2013 to Copyright years.
5944 * config/arc/arc700.md: Likewise.
5945 * config/arc/arc-modes.def: Likewise.
5946 * config/arc/arc-simd.h: Likewise.
5947 * config/arc/t-arc-uClibc: Likewise.
5948 * config/arc/t-arc-newlib: Likewise.
5950 2013-10-02 Renlin Li <renlin.li@arm.com>
5952 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5954 (aarch64_expand_epilogue): Likewise.
5956 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5957 Yufeng Zhang <yufeng.zhang@arm.com>
5959 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
5961 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
5962 'base_in' represent a conversion and legal_cast_p_1 holds; set
5963 'base_in' with the returned value from get_unwidened.
5965 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5967 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
5968 plus_constant is not used.
5970 2013-10-01 Wei Mi <wmi@google.com>
5972 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
5973 * config/i386/i386.md: Add define_peephole2 to
5974 break partial reg stall for cvtss2sd/cvtsd2ss.
5976 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5978 * config/arc/arc.c (pass_arc_ifcvt::clone):
5979 Update for ctxt_ -> m_ctxt change.
5981 2013-10-01 Jeff Law <law@redhat.com>
5983 * tree-ssa-threadupdate.c (struct redirection_data): Delete
5984 outgoing_edge and intermediate_edge fields. Instead store the path.
5985 (redirection_data::hash): Hash on the last edge's destination index.
5986 (redirection_data::equal): Check the entire thread path.
5987 (lookup_redirectio_data): Corresponding changes.
5988 (create_edge_and_update_destination_phis): Likewise.
5989 (thread_single_edge): Likewise.
5991 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5992 Diego Novillo <dnovillo@google.com>
5994 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
5995 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
5996 (vld32wh_insn, vld32wl_insn): Delete commented-out old
5997 versions of these patterns.
5999 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
6000 (__builtin_arc_aligned): Likewise.
6002 * config/arc/arc.md: Expand adc_0 comment stating the intended
6003 purpose and why it isn't ready.
6004 Replace commented out call_value_via_label_mixed with a
6005 plain comment about bl_s.
6007 * config/arc/arc.c (stdio.h): Don't include directly.
6008 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
6009 Remove if (1) condition.
6010 (arc_encode_section_info): Fix comment.
6012 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6014 * config/arc/arc.c (arc_conditional_register_usage):
6015 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
6016 Also set reg_alloc_order for DMA config regs.
6018 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6019 Jeremy Bennett <jeremy.bennett@embecosm.com>
6021 * doc/install.texi (--with-cpu): Mention ARC.
6022 (arc-*-elf32): New paragraph.
6023 (arc-linux-uclibc): Likewise.
6024 * doc/md.texi (Machine Constraints): Add ARC part.
6025 * doc/invoke.texi: (menu): Add ARC Options.
6026 (Machine Dependent Options) <ARC Options>: Add synopsis.
6027 (node ARC Options): Add.
6028 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6029 (ARC Built-in Functions): New section defining
6030 generic ARC built-in functions.
6031 (ARC SIMD Built-in Functions): New section defining SIMD specific
6033 (Declaring Attributes of Functions): Extended
6034 description of short_call and long_call attributes for ARC and
6035 added index entries.
6037 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6038 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6039 Joern Rennecke <joern.rennecke@embecosm.com>
6040 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6041 Brendan Kehoe <brendan@zen.org>
6042 Michael Eager <eager@eagercon.com>
6043 Simon Cook <simon.cook@embecosm.com>
6044 Jeremy Bennett <jeremy.bennett@embecosm.com>
6046 * config/arc, common/config/arc: New directories.
6048 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6049 Brendan Kehoe <brendan@zen.org>
6050 Simon Cook <simon.cook@embecosm.com>
6052 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6054 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6056 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6057 * tree-ssa-coalesce.h: New. Move prototype to here.
6058 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6059 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6060 (gimple_can_coalesce_p): Move to...
6061 * gimple.c (gimple_can_coalesce_p): Here.
6063 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6065 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6066 (dump_decl_set): Move to gimple.c.
6067 * gimple.h: Don't include tree-ssa-operands.h.
6068 (dump_decl_set): Add prototype.
6069 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6070 Move to gimple-ssa.h.
6071 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6072 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6073 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6074 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6075 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6077 (dump_decl_set): Relocate here.
6078 * gimple-ssa.h: New file.
6079 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6080 Relocate from gimple.h.
6081 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6082 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6083 swap_tree_operands and remove non-ssa path.
6084 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6085 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6087 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6088 vect_is_simple_reduction_1): Use swap_ssa_operands.
6089 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6090 (enum need_phi_state): Move to tree-into-ssa.c.
6091 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6092 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6093 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6094 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6095 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6096 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6097 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6098 num_imm_uses): Move to ssa-iterators.h.
6099 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6100 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6102 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6103 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6104 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6105 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6106 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6107 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6108 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6109 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6110 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6111 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6112 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6113 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6114 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6115 (set_phi_nodes): Move to tree-phinodes.h.
6116 * tree-ssa-operands.h (enum ssa_op_iter_type,
6117 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6118 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6119 (dump_decl_set): Remove prototype.
6120 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6121 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6122 (set_phi_nodes): Relocate from tree-flow-inline.h.
6123 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6125 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6126 include list. Temporarily add gimple.h to include list.
6127 * ssa-iterators.h: New file.
6128 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6129 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6130 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6131 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6132 Relocate from tree-ssa-operands.h.
6133 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6134 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6135 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6136 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6137 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6138 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6139 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6140 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6141 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6142 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6143 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6144 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6145 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6146 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6147 Relocate from tree-flow-inline.h.
6148 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6150 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6153 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6154 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6155 Insert '\t' to output template.
6156 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6157 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6158 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6159 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
6161 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
6163 * doc/install.texi (Host/target specific installation notes for GCC):
6164 Put @anchor before @heading.
6165 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
6166 Use @email for email addresses.
6168 2013-10-01 Jeff Law <law@redhat.com>
6170 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
6171 a vec. Only delete the path if we create one without successfully
6172 registering a jump thread.
6173 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
6175 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
6176 (paths): New vector of jump threading paths.
6177 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
6178 (THREAD_PATH): New accessor macro for the entire thread path.
6179 (lookup_redirection_data): Get intermediate and final outgoing edge
6180 from the thread path.
6181 (create_edge_and_update_destination_phis): Copy the threading path.
6182 (ssa_fix_duplicate_block_edges): Get edges and block types from the
6183 jump threading path.
6184 (ssa_redirect_edges): Get edges and block types from the jump threading
6185 path. Free the path vector.
6186 (thread_block): Get edges from the jump threading path. Look at the
6187 entire path to see if we thread to a loop exit. If we cancel a jump
6188 thread request, then free the path vector.
6189 (thread_single_edge): Get edges and block types from the jump threading
6190 path. Free the path vector.
6191 (thread_through_loop_header): Get edges and block types from the jump
6192 threading path. Free the path vector.
6193 (mark_threaded_blocks): Iterate over the vector of paths and store
6194 the path on the appropriate edge. Get edges and block types from the
6195 jump threading path.
6196 (mark_threaded_blocks): Get edges and block types from the jump
6197 threading path. Free the path vector.
6198 (thread_through_all_blocks): Use the vector of paths rather than
6199 a vector of 3-edge sets.
6200 (register_jump_thread): Accept pointer to a path vector rather
6201 than the path vector itself. Store the path vector for later use.
6204 2013-10-01 Jakub Jelinek <jakub@redhat.com>
6205 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6208 * config/s390/s390.c (s390_split_branches): Modify check for table
6210 (s390_chunkify_start): Rearrange table jump insn check in order to
6211 deal with compare and branch insns correctly.
6213 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6217 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6218 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6219 define_insn_and_split.
6220 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6221 (shiftsi3_compare): New pattern.
6223 * config/arm/unspecs.md (UNSPEC_RRX): New.
6225 2013-10-01 Alan Modra <amodra@gmail.com>
6227 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6228 casing inout operands.
6230 2013-10-01 Richard Biener <rguenther@suse.de>
6232 PR tree-optimization/58553
6233 * tree-loop-distribution.c (struct partition_s): Add niter member.
6234 (classify_partition): Populate niter member for the partition
6235 and properly identify whether the relevant store happens before
6236 or after the loop exit.
6237 (generate_memset_builtin): Use niter member from the partition.
6238 (generate_memcpy_builtin): Likewise.
6240 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6242 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6243 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6245 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6247 * basic-block.h (edge_list): Prefix member names with "m_".
6248 * context.h (context): Likewise.
6249 * domwalk.h (dom_walker): Likewise.
6250 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6251 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6252 * hash-table.h (hash_table): Likewise.
6253 * machmode.h (bit_field_mode_iterator): Likewise.
6254 * pass_manager.h (pass_list): Likewise.
6255 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6256 * tree-pass.h (pass_data): Likewise.
6257 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6258 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6259 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6260 * asan.c (pass_data_asan): Update accordingly.
6261 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6262 (control_dependences::control_dependences): Likewise.
6263 (control_dependences::~control_dependences): Likewise.
6264 (control_dependences::~control_dependences): Likewise.
6265 (control_dependences::get_edges_dependent_on): Likewise.
6266 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6267 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6268 * context.c (gcc::context::context): Likewise.
6269 * cprop.c (pass_rtl_cprop::clone): Likewise.
6270 * domwalk.c (dom_walker::walk): Likewise.
6271 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6272 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6273 * mode-switching.c (pass_mode_switching::clone): Likewise.
6274 * passes.c (opt_pass::opt_pass): Likewise.
6275 (pass_manager::pass_manager): Likewise.
6276 * predict.c (pass_strip_predict_hints::clone): Likewise.
6277 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6278 (pass_split_all_insns::clone): Likewise.
6279 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6281 (bit_field_mode_iterator::next_mode): Likewise.
6282 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6283 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6284 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6285 * tree-complex.c (pass_lower_complex::clone): Likewise.
6286 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6287 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6288 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6289 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6290 (pass_fold_builtins::clone): Likewise.
6291 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6292 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6293 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6294 (pass_cd_dce::clone): Likewise.
6295 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6296 (pass_phi_only_cprop::clone): Likewise.
6297 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6298 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6299 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6300 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6301 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6302 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6303 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6304 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6305 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6306 * tree-vrp.c (pass_vrp::clone): Likewise.
6307 * tsan.c (pass_tsan::clone): Likewise.
6309 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6312 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6313 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6316 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6317 optimization, punt if sign_bit_p looked through any zero extension.
6319 2013-09-30 Teresa Johnson <tejohnson@google.com>
6321 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6322 Update redirected out edge count in joiner case.
6323 (ssa_redirect_edges): Common the joiner and non-joiner cases
6324 so that joiner case gets profile updates.
6326 2013-09-30 Richard Biener <rguenther@suse.de>
6328 PR tree-optimization/58554
6329 * tree-loop-distribution.c (classify_partition): Require
6330 unconditionally executed stores for memcpy and memset recognition.
6331 (tree_loop_distribution): Calculate dominance info.
6333 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6335 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6336 (NO_PROFILE_COUNTERS): Likewise.
6337 (PROFILE_HOOK): Likewise.
6338 (FUNCTION_PROFILER): Likewise.
6339 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6341 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6343 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6344 calls and defines in TARGET_MACHO conditional.
6345 (load_macho_picbase_di): Likewise.
6346 (reload_macho_picbase): Likewise.
6347 (reload_macho_picbase_si): Likewise.
6348 (reload_macho_picbase_di): Likewise.
6349 (nonlocal_goto_receiver): Likewise.
6351 2013-09-30 Nick Clifton <nickc@redhat.com>
6353 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6354 that use the MSP430X ISA.
6355 (msp430_option_override): Scan -mmcu command line option for any
6356 MCU name that supports the MSP430X ISA.
6357 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6358 -mmcu options which enable the MSP430X ISA.
6360 2013-09-30 Richard Biener <rguenther@suse.de>
6363 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6364 before looking for setjmp-like calls.
6366 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6369 * config/darwin-protos.h (machopic_get_function_picbase): New.
6370 * config/darwin.c (machopic_get_function_picbase): New.
6371 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6372 label for a new func. (load_macho_picbase_di): Likewise.
6373 (reload_macho_picbase): New expand.
6374 (reload_macho_picbase_si): New insn.
6375 (reload_macho_picbase_di): New insn.
6376 (nonlocal_goto_receiver): New define and split.
6377 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6378 (unspecv enum): Add UNSPECV_NLGR.
6380 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6382 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6383 that altivec registers are correctly sized on Darwin.
6385 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6387 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6388 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6389 * config/x-darwin (host-darwin.o): Likewise.
6390 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6391 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6392 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6394 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6396 * doc/invoke.texi: Fix usage of @tie{} command.
6398 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6400 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6402 2013-09-28 Jan Hubicka <jh@suse.cz>
6404 * config/i386/x86-tune.def: Add documentation for each of the options;
6407 2013-09-28 Jan Hubicka <jh@suse.cz>
6409 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6411 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6412 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6413 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6414 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6415 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6417 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6419 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6420 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6421 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6422 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6423 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6424 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6425 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6426 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6427 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6428 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6429 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6430 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6431 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6432 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6433 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6434 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6435 gimple.h, godump.c, graphite-clast-to-gimple.c,
6436 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6437 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6438 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6439 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6440 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6441 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6442 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6443 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6444 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6445 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6446 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6447 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6448 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6449 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6450 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6451 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6452 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6453 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6454 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6455 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6456 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6457 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6458 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6459 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6460 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6461 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6462 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6463 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6464 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6465 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6466 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6467 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6468 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6469 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6470 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6471 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6472 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6473 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6474 whitespace before "(".
6476 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6478 * expr.h (extract_bit_field): Remove packedp parameter.
6479 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6480 from forward declaration.
6481 (store_split_bit_field): Remove packedp arg from calls to
6482 extract_fixed_bit_field.
6483 (extract_bit_field_1): Remove packedp parameter and packedp
6484 argument from recursive calls and calls to extract_fixed_bit_field.
6485 (extract_bit_field): Remove packedp parameter and corresponding
6486 arg to extract_bit_field_1.
6487 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6489 (extract_split_bit_field): Remove packedp arg from call to
6490 extract_fixed_bit_field.
6491 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6492 (copy_blkmode_from_reg): Likewise.
6493 (copy_blkmode_to_reg): Likewise.
6494 (read_complex_part): Likewise.
6495 (store_field): Likewise.
6496 (expand_expr_real_1): Likewise.
6497 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6498 to extract_bit_field.
6499 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6500 call to extract_bit_field.
6501 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6502 call to extract_bit_field.
6503 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6504 and special packedp behavior from -fstrict-volatile-bitfields
6507 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6509 * lra-eliminations.c (init_elim_table): Guard value_p.
6511 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6513 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6514 DFmode, DImode, and SFmode in the upper VSX registers based on the
6515 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6516 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6517 of the VSX register handling.
6519 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6521 (zero_extendsidi2_lfiwzx): Add support for loading into the
6522 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6523 only do VSX memory options on Altivec registers.
6524 (extendsidi2_lfiwax): Likewise.
6525 (extendsfdf2_fpr): Likewise.
6526 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6527 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6528 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6529 (movdi_internal64): Likewise.
6531 2013-09-27 Xinliang David Li <davidxl@google.com>
6533 * opts.c (finish_options): Adjust parameters
6534 according to vect cost model.
6535 (common_handle_option): Set dynamic vect cost
6537 targhooks.c (default_add_stmt_cost): Compute stmt cost
6539 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6540 Use helper function.
6541 * tree-vectorizer.h (unlimited_cost_model): New function.
6542 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6543 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6545 (vect_enhance_data_refs_alignment): Ditto.
6546 * flag-types.h: New enum.
6547 * common/config/i386/i386-common.c (ix86_option_init_struct):
6548 No need to initialize vect_cost_model flag.
6549 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6552 2013-09-27 Diego Novillo <dnovillo@google.com>
6554 * gimple.h (enum ssa_mode): Remove.
6556 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6558 * cfgloop.h (number_of_loops): Fix typo in check for null.
6560 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6563 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6564 are children of outermost saved_cfun's loop, and set it up to
6565 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6566 if loop != loop0 and SESE region contains bbs that belong to loop0.
6568 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6570 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6571 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6572 (set_address_segment, set_address_base, set_address_index)
6573 (set_address_disp): Accept the argument unconditionally.
6574 (baseness): Remove must_be_base_p and must_be_index_p checks.
6575 (decompose_normal_address): Classify as much as possible in the
6578 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6580 * cse.c (count_reg_usage): Handle INT_LIST.
6581 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6582 * reginfo.c (reg_scan_mark_refs): Likewise.
6583 * reload1.c (eliminate_regs_1): Likewise.
6585 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6588 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6591 2013-09-27 Richard Biener <rguenther@suse.de>
6593 PR tree-optimization/58459
6594 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6595 restriction not propagating into loops.
6597 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6599 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6600 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6601 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6603 2013-09-26 Richard Biener <rguenther@suse.de>
6605 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6607 2013-09-26 Richard Biener <rguenther@suse.de>
6609 * alias.h (component_uses_parent_alias_set): Rename to ...
6610 (component_uses_parent_alias_set_from): ... this.
6611 * alias.c (component_uses_parent_alias_set): Rename to ...
6612 (component_uses_parent_alias_set_from): ... this and return
6614 (reference_alias_ptr_type_1): Use the result from
6615 component_uses_parent_alias_set_from instead of stripping
6616 components one at a time.
6617 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6619 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6621 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6622 Move prototypes to...
6623 * tree-ssa-ter.h: New File. Move prototypes here.
6624 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6625 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6626 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6627 from is_replaceable_p.
6628 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
6629 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
6630 (process_replaceable): Use ter_is_replaceable_p.
6631 (find_replaceable_in_bb): Use ter_is_replaceable_p.
6632 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
6633 newly refactored ssa_is_replaceable_p.
6634 * cfgexpand.c: Include tree-outof-ssa.h.
6635 * ssaexpand.h: Delete.
6637 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6639 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
6640 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
6642 (create_gimple_tmp): Delete.
6643 (get_expr_type, build_assign, build_type_cast): Move to...
6644 * gimple-builder.c: New File.
6645 (get_expr_type): Relocate from gimple.c.
6646 (build_assign, build_type_cast): Change to only create ssanames.
6647 * gimple.h: Move prototypes to...
6648 * gimple-builder.h: New File. Here.
6649 * tree-ssa.h: And here.
6650 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
6651 count_uses_and_derefs): Relocate from gimple.c.
6652 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
6653 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
6654 * tree-ssa-math-opts (execute_cse_reciprocals): Use
6655 gimple_replace_ssa_lhs.
6656 * asan.c: Include gimple-builder.h.
6657 * Makefile.in: Add gimple-builder.o.
6659 2013-09-26 Richard Biener <rguenther@suse.de>
6661 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
6662 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
6663 (loe_visit_block): Use gcc_checking_assert.
6664 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
6665 gimple_assign_ssa_name_copy_p.
6666 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
6669 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
6671 * config/rs6000/t-rs6000 (rs6000.o): Remove.
6672 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
6674 2013-09-26 Richard Biener <rguenther@suse.de>
6676 PR tree-optimization/58539
6677 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
6678 the fact that debug statements are not taking part in loop-closed
6681 2013-09-26 Nick Clifton <nickc@redhat.com>
6683 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
6684 time warning message.
6685 (msp430_print_operand_raw): Delete unused letter parameter.
6686 (TARGET_PRINT_OPERAND_ADDRESS): Define.
6687 (msp430_print_operand_address): New function.
6688 (msp430_print_operand): Move address printing code from here to
6690 * config/msp430/msp430.md (movsipsi2): Add comment in generated
6692 (zero_extendpsisi2): Likewise.
6693 (extendpsisi2): New pattern.
6694 (andneghi3): New pattern.
6696 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
6698 * config/aarch64/aarch64.opt (mlra): New option.
6699 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
6700 (TARGET_LRA_P): Define.
6702 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
6704 * expr.c (expand_assignment): Remove obsolete comment.
6706 2013-09-25 Jeff Law <law@redhat.com>
6708 * tree-flow.h (thread_through_all_blocks): Prototype moved into
6709 tree-ssa-threadupdate.h.
6710 (register_jump_thread): Similarly.
6711 * tree-ssa-threadupdate.h: New header file.
6712 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
6713 * tree-vrp.c: Likewise.
6714 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
6715 (thread_around_empty_blocks): Change type of path vector argument to
6716 an edge,type pair from just an edge. Initialize both elements when
6717 appending to a jump threading path. Tweak references to elements
6719 (thread_across_edge): Similarly. Release memory for the elements
6721 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
6722 (dump_jump_thread_path): New function broken out from
6723 register_jump_thread.
6724 (register_jump_thread): Use dump_jump_thread_path. Change type of
6725 path vector entries. Search the path for NULL edges and dump
6726 the path if one is found. Tweak the conversion of path to 3-edge
6727 form to use the block copy type information embedded in the path.
6729 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6731 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
6733 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6734 Vladimir Makarov <vmakarov@redhat.com>
6736 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
6737 from the least significant bit.
6738 (strip_address_mutations): Add bitfield operations handling.
6739 (must_be_index_p): Add shifting and rotate operations handling.
6740 (set_address_base): Use must_be_base_p predicate.
6741 (set_address_index): Use must_be_index_p predicate.
6743 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
6744 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6745 Sergey Lega <sergey.s.lega@intel.com>
6746 Anna Tikhonova <anna.tikhonova@intel.com>
6747 Ilya Tocar <ilya.tocar@intel.com>
6748 Andrey Turetskiy <andrey.turetskiy@intel.com>
6749 Ilya Verbin <ilya.verbin@intel.com>
6750 Kirill Yukhin <kirill.yukhin@intel.com>
6751 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6753 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
6755 (ix86_expand_vector_move_misalign): Support new unaligned load and
6756 stores and use new names.
6757 (CODE_FOR_sse2_storedqu): Rename to ...
6758 (CODE_FOR_sse2_storedquv16qi): ... this.
6759 (CODE_FOR_sse2_loaddqu): Rename to ...
6760 (CODE_FOR_sse2_loaddquv16qi): ... this.
6761 (CODE_FOR_avx_loaddqu256): Rename to ...
6762 (CODE_FOR_avx_loaddquv32qi): ... this.
6763 (CODE_FOR_avx_storedqu256): Rename to ...
6764 (CODE_FOR_avx_storedquv32qi): ... this.
6765 * config/i386/i386.md (fpint_logic): New.
6766 * config/i386/sse.md (VMOVE): Extend for AVX512.
6770 (VI_UNALIGNED_LOADSTORE): Ditto.
6771 (sse2_avx_avx512f): Ditto.
6772 (sse2_avx2): Extend for AVX512.
6773 (sse4_1_avx2): Ditto.
6774 (avx2_avx512f): New.
6775 (sse): Extend for AVX512.
6778 (avxsizesuffix): Ditto.
6779 (sseintvecmode): Ditto.
6781 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
6782 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
6783 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6784 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
6785 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6786 (<sse2_avx_avx512f>_storedqu<mode): ... this.
6787 (<sse>_movnt<mode>): Replace constraint "x" with "v".
6788 (STORENT_MODE): Extend for AVX512.
6789 (*absneg<mode>2): Replace constraint "x" with "v".
6790 (*mul<mode>3): Ditto.
6791 (*ieee_smin<mode>3): Ditto.
6792 (*ieee_smax<mode>3): Ditto.
6793 (avx_cmp<mode>3): Replace VF with VF_128_256.
6794 (*<sse>_maskcmp<mode>3_comm): Ditto.
6795 (<sse>_maskcmp<mode>3): Ditto.
6796 (<sse>_andnot<mode>3): Extend for AVX512.
6797 (<code><mode>3, anylogic): Replace VF with VF_128_256.
6798 (<code><mode>3, fpint_logic): New.
6799 (*<code><mode>3): Extend for AVX512.
6800 (avx512flogicsuff): New.
6801 (avx512f_<logic><mode>): Ditto.
6802 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
6804 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
6805 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
6806 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
6807 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
6808 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
6809 (xop_vpermil2<mode>3): Ditto.
6810 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
6811 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
6812 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
6813 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
6815 2013-09-25 Tom Tromey <tromey@redhat.com>
6817 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6818 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6819 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6820 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6821 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6822 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6823 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6824 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6825 (GRAPHITE_HTAB_H): Remove.
6827 2013-09-25 Tom Tromey <tromey@redhat.com>
6829 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6831 2013-09-25 Tom Tromey <tromey@redhat.com>
6833 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6835 2013-09-25 Tom Tromey <tromey@redhat.com>
6837 * config/i386/t-i386 (i386.o): Remove.
6838 (i386-c.o): Use COMPILE and POSTCOMPILE.
6840 2013-09-25 Tom Tromey <tromey@redhat.com>
6842 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6844 2013-09-25 Tom Tromey <tromey@redhat.com>
6846 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6847 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6848 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6849 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6850 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
6851 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
6852 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
6853 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
6854 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
6855 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
6856 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
6857 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
6858 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
6859 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
6860 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
6861 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
6862 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
6863 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
6864 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
6865 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
6866 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
6867 (tree-ssa-pre.o, tree-ssa-sccvn.o)
6868 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
6869 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
6870 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
6871 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
6872 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
6873 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
6874 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
6875 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
6876 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
6877 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
6878 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
6879 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
6880 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
6881 (graphite-clast-to-gimple.o, graphite-dependences.o)
6882 (graphite-interchange.o, graphite-poly.o)
6883 (graphite-scop-detection.o, graphite-sese-to-poly.o)
6884 (graphite-optimize-isl.o, tree-vect-loop.o)
6885 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
6886 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
6887 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
6888 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
6889 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
6890 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
6891 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
6892 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
6893 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
6894 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
6895 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
6896 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
6897 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
6898 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
6899 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
6900 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
6901 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
6902 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
6903 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
6904 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
6905 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
6906 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
6907 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
6908 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
6909 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
6910 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
6911 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
6912 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
6913 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
6914 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
6915 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
6916 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
6917 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
6918 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
6919 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
6920 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
6921 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
6922 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
6923 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
6924 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
6925 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
6926 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
6927 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
6928 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
6929 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
6930 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
6931 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
6932 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
6933 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
6934 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
6935 (gcov-dump.o): Remove.
6936 (default-c.o): Use COMPILE and POSTCOMPILE.
6937 (CFLAGS-gcc.o): New variable.
6938 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
6940 2013-09-25 Tom Tromey <tromey@redhat.com>
6942 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
6943 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
6944 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
6945 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
6946 (c-family/c-pragma.o, c-family/c-pretty-print.o)
6947 (c-family/c-semantics.o, c-family/c-ada-spec.o)
6948 (c-family/array-notation-common.o, c-family/stub-objc.o)
6949 (c-family/c-ubsan.o): Remove.
6951 2013-09-25 Tom Tromey <tromey@redhat.com>
6953 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
6955 2013-09-25 Tom Tromey <tromey@redhat.com>
6957 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
6958 to add -DENABLE_SHARED_LIBGCC.
6959 (gcc.o): Don't use subshell.
6961 2013-09-25 Tom Tromey <tromey@redhat.com>
6963 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
6964 * configure.ac: Don't invoke AM_PROG_CC_C_O.
6965 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
6966 * configure, config.in: Rebuild.
6968 2013-09-25 Tom Tromey <tromey@redhat.com>
6970 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
6971 (COMPILE, POSTCOMPILE): New variables.
6972 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
6973 (DEPFILES): New variable.
6974 Include ".Po" files.
6975 * configure.ac: Add checks for dependency checking.
6976 * configure, aclocal.m4: Regenerate.
6978 2013-09-25 Tom Tromey <tromey@redhat.com>
6980 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
6981 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
6983 2013-09-25 Tom Tromey <tromey@redhat.com>
6985 * Makefile.in (generated_files): Add options.h,
6986 target-hooks-def.h, insn-opinit.h,
6987 common/common-target-hooks-def.h, pass-instances.def,
6988 c-family/c-target-hooks-def.h.
6990 2013-09-25 Jeff Law <law@redhat.com>
6992 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
6993 than foo[foo.length () - 1] to access last member in a vec.
6994 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
6996 2013-09-25 Richard Biener <rguenther@suse.de>
6999 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
7001 2013-09-25 Jan Hubicka <jh@suse.cz>
7003 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
7006 2013-09-25 Marek Polacek <polacek@redhat.com>
7009 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
7010 when determining the type name.
7012 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
7014 * config/sh/sh.md: Fix formatting.
7016 2013-09-24 Xinliang David Li <davidxl@google.com>
7018 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7019 max peel iterations parameter.
7020 * param.def: New parameter.
7021 * doc/invoke.texi: Document New parameter.
7023 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7025 * gimple-pretty-print.c: Various whitespace tweaks.
7026 * tree-core.h: Likewise.
7027 * tree-pretty-print.c: Likewise.
7028 * tree-ssa-alias.c: Likewise.
7029 * tree-ssa-copy.c: Likewise.
7030 * tree-ssanames.c: Likewise.
7031 * tree-ssanames.h: Likewise.
7032 * tree-vrp.c: Likewise.
7034 2013-09-24 Alan Modra <amodra@gmail.com>
7038 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7039 for output operands that disallow regs. Don't use EXPAND_WRITE on
7042 2013-09-24 Richard Biener <rguenther@suse.de>
7045 * tree.c (reference_alias_ptr_type): Move ...
7046 * alias.c (reference_alias_ptr_type): ... here and implement
7047 in terms of the new reference_alias_ptr_type_1.
7048 (ref_all_alias_ptr_type_p): New helper.
7049 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7050 use ref_all_alias_ptr_type_p.
7051 (get_deref_alias_set): Add flag_strict_aliasing check here.
7052 (reference_alias_ptr_type_1): New function, split out from ...
7053 (get_alias_set): ... here.
7054 (alias_ptr_types_compatible_p): New function.
7055 * alias.h (reference_alias_ptr_type): Declare.
7056 (alias_ptr_types_compatible_p): Likewise.
7057 * tree.h (reference_alias_ptr_type): Remove.
7058 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7059 to compare MEM_REF alias types.
7061 2013-09-24 Richard Biener <rguenther@suse.de>
7063 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7065 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7067 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7068 reload helper function arrays into a single array reg_addr.
7069 (reload_fpr_gpr): Likewise.
7070 (reload_gpr_vsx): Likewise.
7071 (reload_vsx_gpr): Likewise.
7072 (struct rs6000_reg_addr): Likewise.
7073 (reg_addr): Likewise.
7074 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7075 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7076 (rs6000_init_hard_regno_mode_ok): Likewise.
7077 (rs6000_secondary_reload_direct_move): Likewise.
7078 (rs6000_secondary_reload): Likewise.
7080 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7081 constraints: wu, ww, and wy. Repurpose wv constraint added during
7082 power8 changes. Put wg constraint in alphabetical order.
7084 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7085 for future work to add ISA 2.07 VSX single precision support.
7086 (-mvsx-scalar-double): Change default from -1 to 1, update
7087 documentation comment.
7088 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7089 (-mupper-regs-df): New debug switch to control whether DF values
7090 can go in the traditional Altivec registers.
7091 (-mupper-regs-sf): New debug switch to control whether SF values
7092 can go in the traditional Altivec registers.
7094 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7096 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7097 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7098 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7099 (rs6000_debug_legitimate_address_p): Print if we are running
7100 before, during, or after reload.
7101 (rs6000_secondary_reload): Add a comment.
7102 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7104 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7105 constraints. Update documentation string.
7106 (wd constraint): Likewise.
7107 (wf constraint): Likewise.
7108 (wg constraint): Likewise.
7109 (wn constraint): Likewise.
7110 (ws constraint): Likewise.
7111 (wt constraint): Likewise.
7112 (wx constraint): Likewise.
7113 (wz constraint): Likewise.
7114 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7116 (ww constraint): Likewise.
7117 (wy constraint): Likewise.
7118 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7119 in the previous submissions.
7120 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7122 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7124 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7125 (REG_BR_PROB): Say that the probability is stored in an int_list.
7126 * reg-notes.def: Update commentary to mention INT_LIST.
7127 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7128 (INT_LIST): New rtx.
7129 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7130 * rtlanal.c (int_reg_note_p): New function.
7131 (alloc_reg_note): Assert that the note does not have an int argument.
7132 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7133 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7134 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7135 rather than an INSN_LIST. Handle INT_LIST.
7136 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7137 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7138 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7139 Manipulate them as ints rather than rtxes.
7140 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7141 * regmove.c (copy_src_to_dest): Likewise.
7142 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7144 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7145 into the cases that need it.
7146 * config/arm/arm.c (arm_unwind_emit): Likewise.
7148 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7149 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7150 * loop-doloop.c (add_test, doloop_modify): Likewise.
7151 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7152 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7153 * predict.c (combine_predictions_for_insn): Likewise.
7154 * print-rtl.c (print_rtx): Handle INT_LIST.
7155 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7156 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7157 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7158 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7159 (ix86_split_fp_branch, predict_jump): Likewise.
7160 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
7161 * config/sh/sh.c (expand_cbranchsi4): Likewise.
7162 * config/spu/spu.c (ea_load_store_inline): Likewise.
7164 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
7165 value of a REG_BR_PROB note.
7166 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
7167 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
7168 * emit-rtl.c (try_split): Likewise.
7169 * predict.c (br_prob_note_reliable_p): Likewise.
7170 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
7171 * reorg.c (mostly_true_jump): Likewise.
7172 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
7173 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
7174 * config/i386/i386.c (ix86_print_operand): Likewise.
7175 * config/ia64/ia64.c (ia64_print_operand): Likewise.
7176 * config/mmix/mmix.c (mmix_print_operand): Likewise.
7177 * config/rs6000/rs6000.c (output_cbranch): Likewise.
7178 * config/s390/s390.c (s390_expand_tbegin): Likewise.
7179 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
7180 * config/sparc/sparc.c (output_cbranch): Likewise.
7181 * config/spu/spu.c (get_branch_target): Likewise.
7182 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
7183 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
7185 2013-09-23 Jan Hubicka <jh@suse.cz>
7187 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
7189 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
7190 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
7192 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
7193 (gimple_fold_call): Dump inconsistent devirtualizations; add
7194 sanity check for type based devirtualizations.
7195 * ipa-prop.c: Include ipa-utils.h
7196 (ipa_intraprocedural_devirtualization): Add sanity check.
7197 (try_make_edge_direct_virtual_call): Likewise.
7199 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
7201 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
7202 assignment statements.
7204 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
7206 * gimple-pretty-print.c (dump_ssaname_info): New function.
7207 (dump_gimple_phi): Call it.
7208 (pp_gimple_stmt_1): Likewise.
7209 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7210 (range_info_def): Declare.
7211 * tree-pretty-print.c (pp_double_int): New function.
7212 (dump_generic_node): Call it.
7213 * tree-pretty-print.h (pp_double_int): Declare.
7214 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7215 * tree-ssanames.h (range_info_def): New structure.
7216 (value_range_type): Move definition here.
7217 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7219 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7221 (set_range_info): New function.
7222 (get_range_info): Likewise.
7223 (duplicate_ssa_name_range_info): Likewise.
7224 (duplicate_ssa_name_fn): Check pointer type and call
7225 duplicate_ssa_name_range_info.
7226 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7227 * tree-vrp.c (value_range_type): Remove definition, now in
7229 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7230 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7231 (SSA_NAME_RANGE_INFO): New macro.
7233 2013-09-23 Richard Biener <rguenther@suse.de>
7235 PR tree-optimization/58464
7236 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7237 (phi_trans_add): Change to add conditionally on being not present.
7238 (phi_translate_1): Remove recursion detection here.
7239 (phi_translate): Pre-seed the cache with NULL to catch
7240 recursion here in a more generic way.
7241 (bitmap_find_leader): Adjust comment.
7242 (get_representative_for): Dump value-numbers.
7243 (create_expression_by_pieces): Likewise.
7244 (insert_into_preds_of_block): Likewise.
7246 2013-09-23 Christian Bruel <christian.bruel@st.com>
7249 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7250 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7252 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7255 2013-09-20 Renlin Li <renlin.li@arm.com>
7257 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7259 (aarch64_expand_epilogue): Likewise.
7260 (aarch64_legitimize_reload_address): Likewise.
7262 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7264 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7265 NULL_TREE before pushing them onto the vector. Likewise for labels.
7267 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7269 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7270 comparison operators when -fno-trapping-math is in effect.
7271 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7272 comparison operators in TFmode and assert that unsupported operators
7274 (ia64_print_operand): Likewise.
7276 2013-09-21 Jan Hubicka <jh@suse.cz>
7278 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7279 (sse_typeless_stores): Enable for core
7280 (sse_load0_by_pxor): Likewise.
7281 (four_jump_limit): Disable for core.
7282 (pad_returns): Likewise.
7283 (avoid_vector_decode): Likewise.
7284 (fuse_cmp_and_branch): Enable for cores.
7285 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7287 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7290 * config/pa/pa.c (pa_option_override): Disable auto increment and
7291 decrement instructions until reload is completed.
7293 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7294 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7295 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7297 2013-09-20 DJ Delorie <dj@redhat.com>
7298 Nick Clifton <nickc@redhat.com>
7300 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7301 (need_to_save): Save bank 0 on interrupts.
7302 (characterize_address): Strip far address wrappers.
7303 (rl78_as_legitimate_address): Likewise.
7304 (transcode_memory_rtx): Likewise.
7305 (rl78_peep_movhi_p): Disable this peephole after devirt.
7306 (rl78_propogate_register_origins): Forget all origins when a
7308 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7309 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7311 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7312 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7313 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7315 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7317 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7319 * config/pa/pa.md: In "scc" insn patterns, change output template to
7320 handle const0_rtx in reg_or_0_operand operands.
7322 2013-09-20 Martin Husemann <martin@NetBSD.org>
7325 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7326 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7328 2013-09-20 Richard Biener <rguenther@suse.de>
7331 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7332 remembering SSA name version and block index.
7333 (new_scev_info_str): Adjust.
7334 (hash_scev_info): Likewise. Also hash the block index.
7335 (eq_scev_info): Adjust.
7336 (find_var_scev_info): Likewise.
7337 (struct instantiate_cache_entry): Remove.
7338 (struct instantiate_cache_type): Use a htab to map name, block
7340 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7341 (get_instantiated_value_entry): Likewise.
7342 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7343 (instantiate_scev_name): Adjust.
7345 2013-09-20 Jeff Law <law@redhat.com>
7347 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7349 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7351 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7352 Call aarch64_simd_expand_args to update op[argc].
7354 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7356 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7359 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7361 * gengtype.c (file_rules): Added rule for *.cc files.
7362 (get_output_file_with_visibility): Give fatal message when no
7365 2013-09-20 Renlin Li <renlin.li@arm.com>
7367 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7368 (aarch64_expand_epilogue): Likewise.
7369 (aarch64_legitimize_reload_address): Likewise.
7371 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7374 * expr.c (expand_assignment): Remove misalignp code path.
7376 2013-09-20 Marek Polacek <polacek@redhat.com>
7379 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7380 TYPE_PRECISION. Add asserts.
7382 2013-09-20 Richard Biener <rguenther@suse.de>
7384 PR tree-optimization/58453
7385 * tree-loop-distribution.c (distribute_loop): Apply the cost
7386 model for -ftree-loop-distribute-patterns, too.
7388 2013-09-20 Richard Biener <rguenther@suse.de>
7391 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7392 make type comparison less strict.
7394 2013-09-20 Alan Modra <amodra@gmail.com>
7396 * configure: Regenerate.
7397 * aclocal.m4: Regenerate.
7399 2013-09-20 Marek Polacek <polacek@redhat.com>
7402 * doc/extend.texi: Document that attribute used is meant to be used
7403 on variables with static storage duration.
7405 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7407 PR tree-optimization/58472
7408 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7409 simd_lane_access set inv_p = false.
7410 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7411 the simduid magic VAR_DECL.
7413 2013-09-19 Jan Hubicka <jh@suse.cz>
7415 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7417 2013-09-17 Jeff Law <law@redhat.com>
7419 * tree-ssa-dom.c (record_temporary_equivalences): New function
7420 split out of dom_opt_dom_walker::after_dom_children.
7421 (dom_opt_dom_walker::thread_across_edge): Move common code
7422 in here from dom_opt_dom_walker::after_dom_children.
7423 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7425 2013-09-19 Jan Hubicka <jh@suse.cz>
7427 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7428 (TARGET_GENERIC): Use PROCESOR_GENERIC
7429 (enum processor_type): Unify generic32 and 64.
7430 * i386.md (cpu): Likewise.
7431 * x86-tune.def (use_leave): Enable for generic32.
7432 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7433 * athlon.md: Change generic64 to generic in all occurences.
7434 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7435 (ix86_target_macros_internal): Likewise.
7436 * driver-i386.c (host_detect_local_cpu): Likewise.
7437 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7439 (generic_memcpy, generic_memset, generic_cost): This one.
7440 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7441 (m_GENERIC32, m_GENERIC64): Remove.
7442 (m_GENERIC): Turn into one flag.
7443 (processor_target): Unify generic tunnings.
7444 (ix86_option_override_internal): Replace generic32/64 by generic.
7445 (ix86_issue_rate): Likewise.
7446 (ix86_adjust_cost): Likewise.
7448 2013-09-19 Jan Hubicka <jh@suse.cz>
7450 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7451 of speculative flag.
7453 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7455 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7456 GOMP_sections_start, even if !exit_reachable.
7458 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7460 * lra-constraints.c (need_for_all_save_p): Use macro
7461 HARD_REGNO_CALL_PART_CLOBBERED.
7462 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7463 set up pseudo conflict hard regs.
7465 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7468 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7471 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7473 PR rtl-optimization/58438
7474 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7475 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7476 from previous subpasses.
7478 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7480 * arm.c (arm_get_frame_offsets): Validate architecture supports
7481 LDRD/STRD before accepting the tuning preference.
7482 (arm_expand_prologue): Likewise.
7483 (arm_expand_epilogue): Likewise.
7485 2013-09-18 Richard Biener <rguenther@suse.de>
7487 PR tree-optimization/58417
7488 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7489 have chrecs with symbols defined in the loop as operands.
7490 (chrec_fold_multiply): Likewise.
7491 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7492 parameters before folding binary operations.
7493 (struct instantiate_cache_entry_hasher): Remove.
7494 (struct instantiate_cache_type): Use a pointer-map.
7495 (instantiate_cache_type::instantiate_cache_type): New function.
7496 (instantiate_cache_type::get): Likewise.
7497 (instantiate_cache_type::set): Likewise.
7498 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7499 (get_instantiated_value_entry): Likewise.
7500 (global_cache): New global.
7501 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7502 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7503 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7504 (instantiate_scev_name): Adjust.
7505 (instantiate_scev): Construct global instead of local cache.
7506 (resolve_mixers): Likewise.
7508 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7509 Paolo Carlini <paolo.carlini@oracle.com>
7512 * doc/implement-cxx.texi: Fix references to the C++ standards.
7514 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7516 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7517 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7518 simd_lane_access drs, update also DR_ALIGNED_TO.
7520 2013-09-18 Marek Polacek <polacek@redhat.com>
7523 * doc/extend.texi: Document no_sanitize_undefined attribute.
7524 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7525 no_sanitize_undefined attribute.
7527 2013-09-18 Nick Clifton <nickc@redhat.com>
7529 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7530 (ASM_DECLARE_FUNCTION_NAME): Define.
7532 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7534 * compare-elim.c (find_comparison_dom_walker): New class
7535 (find_comparisons_in_bb): Rename to
7536 find_comparison_dom_walker::before_dom_children
7537 (find_comparisons): Adjust
7538 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7540 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7541 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7542 (init_walk_dominator_tree): Remove declaration.
7543 (fini_walk_dominator_tree): Remove declaration.
7544 * fwprop.c (single_def_use_dom_walker): New class
7545 (single_def_use_enter_block): Convert to
7546 single_def_use_dom_walker::before_dom_children.
7547 (single_def_use_leave_block): Convert to
7548 single_def_use_dom_walker::after_dom_children.
7549 (build_single_def_use_links): Adjust.
7550 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7552 (find_candidates_in_block): Convert to
7553 find_candidates_dom_walker::before_dom_children.
7554 (execute_strength_reduction): Adjust.
7555 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7556 (sese_dom_walker): New class.
7557 (sese_dom_walker::sese_dom_walker): New constructor.
7558 (sese_dom_walker::~sese_dom_walker): New destructor.
7559 (build_sese_conditions_before): Convert to
7560 sese_dom_walker::before_dom_children.
7561 (build_sese_conditions_after): Convert to
7562 sese_dom_walker::after_dom_children.
7563 (build_poly_scop): Adjust
7564 * tree-into-ssa.c (rewrite_dom_walker): New class
7565 (rewrite_enter_block): Convert to
7566 rewrite_dom_walker::before_dom_children.
7567 (rewrite_leave_block): Convert to
7568 rewrite_dom_walker::after_dom_children.
7569 (rewrite_update_dom_walker): New class.
7570 (rewrite_update_enter_block): Convert to
7571 rewrite_update_dom_walker::before_dom_children.
7572 (rewrite_update_leave_block): Convert to
7573 rewrite_update_dom_walker::after_dom_children.
7574 (rewrite_blocks, rewrite_into_ssa): Adjust.
7575 (mark_def_dom_walker): New class.
7576 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7577 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7578 (mark_def_sites_blocks): Convert to
7579 mark_def_dom_walker::before_dom_children.
7580 (mark_def_site_blocks): Remove.
7581 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7582 (tree_ssa_dominator_optimize): Adjust.
7583 (dom_thread_across_edge): Convert to method
7584 dom_opt_dom_walker::thread_across_edge.
7585 (dom_opt_enter_block): Convert to member function
7586 dom_opt_dom_walker::before_dom_children.
7587 (dom_opt_leave_block): Convert to member function
7588 dom_opt_dom_walker::after_dom_children.
7589 * tree-ssa-dse.c (dse_dom_walker): New class.
7590 (dse_enter_block): Convert to member function
7591 dse_dom_walker::before_dom_children.
7592 (tree_ssa_dse): Adjust.
7593 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7594 (determine_invariantness_stmt): Convert to method
7595 invariantness_dom_walker::before_dom_children.
7596 (determine_invariantness): Remove
7597 (move_computations_dom_walker): New class.
7598 (move_computations_stmt): Convert to method
7599 move_computations_dom_walker::before_dom_children.
7600 (move_computations, tree_ssa_lim): Adjust.
7601 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7602 (nt_init_block): Convert to method
7603 notrappping_dom_walker::before_dom_children.
7604 (nt_fini_block): Convert to method
7605 method nontrapping_dom_walker::after_dom_children.
7606 (get_non_trapping): Adjust.
7607 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7608 (eliminate_bb): Convert to method
7609 eliminate_dom_walker::before_dom_children.
7610 (eliminate_leave_block): Convert to method
7611 eliminate_dom_walker::after_dom_children.
7612 (eliminate): Adjust.
7613 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7614 (strlen_enter_block): Convert to method
7615 strlen_dom_walker::before_dom_children.
7616 (strlen_leave_block): Convert to method
7617 method strlen_dom_walker::after_dom_children.
7618 (tree_ssa_strlen): Adjust.
7619 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7620 (tree_ssa_uncprop): Adjust.
7621 (uncprop_leave_block): Convert to method
7622 uncprop_dom_walker::after_dom_children.
7623 (uncprop_leave_block): Convert to method
7624 uncprop_dom_walker::before_dom_children.
7626 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7628 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
7629 branch in same basic block. Check both src and dest of the move insn.
7631 2013-09-17 Nick Clifton <nickc@redhat.com>
7633 * config/rl78/rl78-real.md (bf): New pattern.
7635 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
7636 (rl78_print_operand): Do not put a # before a %B.
7637 * config/rl78/rl78.opt: Tweak doc strings.
7639 2013-09-17 DJ Delorie <dj@redhat.com>
7641 * config/rl78/constraints.md (Wcv): Allow up to $r31.
7642 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
7643 (rl78_option_override): Likewise, if -mallregs.
7644 (is_virtual_register): Likewise.
7645 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
7646 (REGNO_OK_FOR_BASE_P): Likewise.
7647 * config/rl78/rl78.opt (-mallregs): New.
7649 2013-09-17 Nick Clifton <nickc@redhat.com>
7651 * config/rl78/rl78.c (need_to_save): Change return type to bool.
7652 For interrupt functions: save all call clobbered registers if the
7653 interrupt handler is not a leaf function.
7654 (rl78_expand_prologue): Always recompute the frame information.
7655 For interrupt functions: only select bank 0 if one of the bank 0
7656 registers is going to be psuhed.
7658 2013-09-17 DJ Delorie <dj@redhat.com>
7660 * config/rl78/constraints.md: For each W* constraint, rename to C*
7661 and create a W* constraint that checks for an optional ES: prefix
7663 * config/rl78/rl78.md (UNS_ES_ADDR): New.
7664 (es_addr): New. Used to wrap far addresses.
7665 * config/rl78/rl78-protos.h (rl78_es_addr): New.
7666 (rl78_es_base): New.
7667 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
7668 wrapped far addresses.
7669 (rl78_print_operand_1): Unwrap far addresses before processing.
7670 (rl78_lo16): Wrap far addresses in unspecs.
7671 (rl78_es_addr): New.
7672 (rl78_es_base): New.
7673 (insn_ok_now): Check for not-yet-wrapped far addresses.
7674 (transcode_memory_rtx): Properly re-wrap far addresses.
7676 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7678 * config/sparc/t-rtems: Add leon3 multilibs.
7680 2013-09-17 Cong Hou <congh@google.com>
7682 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
7683 when checking the dot production pattern. The type of rhs operand
7684 of multiply is now checked correctly.
7686 2013-09-17 Jeff Law <law@redhat.com>
7688 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7689 edge implied equivalences into successor phis.
7690 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
7691 here from tree-ssa-threadedge.c.
7692 (mark_threaded_blocks): When threading through a joiner, if both
7693 successors of the joiner's clone reach the same block, verify the
7694 PHI arguments are equal. If not, cancel the jump threading request.
7695 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
7696 tree-ssa-threadupdate.c
7697 (thread_across_edge): Don't check PHI argument equality when
7698 threading through joiner block here.
7700 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
7702 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
7703 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7704 (warn_uninit, warn_uninitialized_vars,
7705 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7706 Move to tree-ssa-uninit.c.
7707 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7708 (has_undefined_value_p): New. Pass dependant parts of
7709 ssa_undefined_value_p.
7710 (uninit_undefined_value_p): Use has_undefined_value_p.
7711 (warn_uninit, warn_uninitialized_vars,
7712 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7713 Move from tree-ssa.c.
7714 * tree-ssa.h: Adjust prototypes.
7716 2013-09-17 Jan Hubicka <jh@suse.cz>
7719 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
7720 * ipa-inline.c (can_inline_edge_p): Do not downgrade
7721 FUNCTION_NOT_OPTIMIZED.
7722 * ipa-inline-analysis.c (compute_inline_parameters): Function
7723 not optimized is not inlinable unless it is alwaysinline.
7724 (inline_analyze_function): Force calls in not optimized
7725 function not inlinable.
7727 2013-09-17 Jan Hubicka <jh@suse.cz>
7730 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
7732 * ipa.c (function_and_variable_visibility): Likewise.
7733 * ipa-profile.c (ipa_profile): Likewise.
7735 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
7738 * cgraph.c (cgraph_function_body_availability): Check for ifunc
7739 attribute, and don't inline the resolver in this case.
7741 2013-09-17 Teresa Johnson <tejohnson@google.com>
7743 * coverage.c (get_coverage_counts): Add missing newline.
7745 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7747 PR tree-optimization/58088
7748 * fold-const.c (mask_with_trailing_zeros): New function.
7749 (fold_binary_loc): Make sure we don't recurse infinitely
7750 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
7751 Use mask_with_trailing_zeros where appropriate.
7753 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
7755 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
7756 of 'prev' var to get better distance estimation.
7758 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
7760 * tree-inline.h (struct copy_body_data): Add transform_parameter.
7761 * tree-inline.c (is_parameter_of): New predicate.
7762 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
7763 a parameter has been remapped.
7764 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
7765 (optimize_inline_calls): Initialize transform_parameter.
7766 (copy_gimple_seq_and_replace_locals): Likewise.
7767 (tree_function_versioning): Likewise.
7768 (maybe_inline_call_in_expr): Likewise.
7770 2013-09-17 Nick Clifton <nickc@redhat.com>
7772 * config/msp430/msp430-protos.h: Add prototypes for new functions.
7773 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
7775 (is_attr_func): New function.
7776 (msp430_is_interrupt_func): New function.
7777 (is_naked_func): New function.
7778 (is_reentrant_func): New function.
7779 (is_critical_func): New function.
7780 (msp430_start_function): Add annotations for function attributes.
7781 (msp430_attr): New function.
7782 (msp430_attribute_table): New.
7783 (msp430_function_section): New function.
7784 (TARGET_ASM_FUNCTION_SECTION): Define.
7785 (msp430_builtin): New enum.
7786 (msp430_init_builtins): New function.
7787 (msp430_builtin_devl): New function.
7788 (msp430_expand_builtin): New function.
7789 (TARGET_INIT_BUILTINS): Define.
7790 (TARGET_EXPAND_BUILTINS): Define.
7791 (TARGET_BUILTIN_DECL): Define.
7792 (msp430_expand_prologue): Add support for naked, interrupt,
7793 critical and reentrant functions.
7794 (msp430_expand_epilogue): Likewise.
7795 (msp430_print_operand): Handle 'O' character.
7796 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
7798 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
7799 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
7800 (pushm): Use a 'n' rather than an 'i' constraint.
7801 (msp_return): Add generation of the interrupt return instruction.
7802 (disable_interrupts): New pattern.
7803 (enable_interrupts): New pattern.
7804 (push_intr_state): New pattern.
7805 (pop_intr_state): New pattern.
7806 (bic_SR): New pattern.
7807 (bis_SR): New pattern.
7808 * doc/extend.texi: Document MSP430 function attributes and builtin
7811 2013-09-17 Richard Biener <rguenther@suse.de>
7813 PR tree-optimization/58432
7814 * tree-loop-distribution.c (tree_loop_distribution): Also
7815 scan PHIs for outside loop uses and seed a partition from them.
7817 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7819 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7820 (restructure_reference): Call backtrace_base_for_ref.
7822 2013-09-17 Alan Modra <amodra@gmail.com>
7825 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7828 2013-09-16 DJ Delorie <dj@redhat.com>
7830 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7831 vregs location for RL78/G10.
7832 (rl78_expand_prologue): Avoid SEL on G10.
7833 (rl78_expand_epilogue): Likewise.
7834 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7835 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7836 __RL78_G10__ when appropriate.
7837 (ASM_SPEC): Pass -mg10 along to the assembler.
7838 * config/rl78/rl78.md (sel_rb): Disable for G10.
7839 * config/rl78/rl78.opt: Add -mg10 option.
7840 * config/rl78/t-rl78: Add -mg10 multilib.
7842 2013-09-16 Xinliang David Li <davidxl@google.com>
7844 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7845 * omp-low.c (omp_max_vf): Ditto.
7846 (expand_omp_simd): Ditto.
7847 * tree-vectorizer.c (vectorize_loops): Ditto.
7848 (gate_vect_slp): Ditto.
7849 (gate_increase_alignment): Ditto.
7850 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
7851 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
7852 (gate_tree_vectorize): Name change.
7853 (tree_vectorize): Ditto.
7854 (pass_vectorize::gate): Call new function.
7855 (pass_vectorize::execute): Ditto.
7856 * opts.c: O3 default setting change.
7857 (finish_options): Check new flag.
7858 * doc/invoke.texi: Document new flags.
7859 * common.opt: New flags.
7861 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
7863 * doc/tm.texi.in (Cond Exec Macros): Remove node.
7864 (Condition Code): Don't reference it.
7865 * doc/tm.texi: Regenerate.
7867 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
7870 * lra-constraints.c (undo_optional_reloads): Consider all optional
7871 reload even if it did not get a hard reg.
7873 2013-09-16 Teresa Johnson <tejohnson@google.com>
7875 * dumpfile.c (dump_loc): Remove newline emission.
7876 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
7877 emission to dump_printf_loc calls where missing.
7878 (vect_mark_for_runtime_alias_test): Ditto.
7879 (vect_analyze_data_ref_dependence): Ditto.
7880 (vect_analyze_data_ref_dependences): Ditto.
7881 (vect_slp_analyze_data_ref_dependence): Ditto.
7882 (vect_slp_analyze_data_ref_dependences): Ditto.
7883 (vect_compute_data_ref_alignment): Ditto.
7884 (vect_update_misalignment_for_peel): Ditto.
7885 (vect_verify_datarefs_alignment): Ditto.
7886 (vector_alignment_reachable_p): Ditto.
7887 (vect_get_data_access_cost): Ditto.
7888 (vect_enhance_data_refs_alignment): Ditto.
7889 (vect_find_same_alignment_drs): Ditto.
7890 (vect_analyze_data_refs_alignment): Ditto.
7891 (vect_analyze_group_access): Ditto.
7892 (vect_analyze_data_ref_access): Ditto.
7893 (vect_analyze_data_ref_accesses): Ditto.
7894 (vect_prune_runtime_alias_test_list): Ditto.
7895 (vect_analyze_data_refs): Ditto.
7896 (vect_create_addr_base_for_vector_ref): Ditto.
7897 (vect_create_data_ref_ptr): Ditto.
7898 (vect_grouped_store_supported): Ditto.
7899 (vect_grouped_load_supported): Ditto.
7900 * value-prof.c (check_counter): Ditto.
7901 (check_ic_target): Ditto.
7902 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
7903 (vect_recog_widen_mult_pattern): Ditto.
7904 (vect_recog_widen_sum_pattern): Ditto.
7905 (vect_recog_over_widening_pattern): Ditto.
7906 (vect_recog_widen_shift_pattern): Ditto.
7907 (vect_recog_rotate_pattern): Ditto.
7908 (vect_recog_vector_vector_shift_pattern): Ditto.
7909 (vect_recog_divmod_pattern): Ditto.
7910 (vect_recog_mixed_size_cond_pattern): Ditto.
7911 (vect_recog_bool_pattern): Ditto.
7912 (vect_pattern_recog_1): Ditto.
7913 (vect_pattern_recog): Ditto.
7914 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
7915 (vect_is_simple_iv_evolution): Ditto.
7916 (vect_analyze_scalar_cycles_1): Ditto.
7917 (vect_get_loop_niters): Ditto.
7918 (vect_analyze_loop_1): Ditto.
7919 (vect_analyze_loop_form): Ditto.
7920 (vect_analyze_loop_operations): Ditto.
7921 (vect_analyze_loop_2): Ditto.
7922 (vect_analyze_loop): Ditto.
7923 (report_vect_op): Ditto.
7924 (vect_is_slp_reduction): Ditto.
7925 (vect_is_simple_reduction_1): Ditto.
7926 (vect_get_known_peeling_cost): Ditto.
7927 (vect_estimate_min_profitable_iters): Ditto.
7928 (vect_model_reduction_cost): Ditto.
7929 (vect_model_induction_cost): Ditto.
7930 (get_initial_def_for_induction): Ditto.
7931 (vect_create_epilog_for_reduction): Ditto.
7932 (vectorizable_reduction): Ditto.
7933 (vectorizable_induction): Ditto.
7934 (vectorizable_live_operation): Ditto.
7935 (vect_loop_kill_debug_uses): Ditto.
7936 (vect_transform_loop): Ditto.
7937 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
7938 (vect_stmt_relevant_p): Ditto.
7939 (process_use): Ditto.
7940 (vect_mark_stmts_to_be_vectorized): Ditto.
7941 (vect_model_simple_cost): Ditto.
7942 (vect_model_promotion_demotion_cost): Ditto.
7943 (vect_model_store_cost): Ditto.
7944 (vect_get_store_cost): Ditto.
7945 (vect_model_load_cost): Ditto.
7946 (vect_get_load_cost): Ditto.
7947 (vect_init_vector_1): Ditto.
7948 (vect_get_vec_def_for_operand): Ditto.
7949 (vect_finish_stmt_generation): Ditto.
7950 (vectorizable_call): Ditto.
7951 (vectorizable_conversion): Ditto.
7952 (vectorizable_assignment): Ditto.
7953 (vectorizable_shift): Ditto.
7954 (vectorizable_operation): Ditto.
7955 (vectorizable_store): Ditto.
7956 (vectorizable_load): Ditto.
7957 (vectorizable_condition): Ditto.
7958 (vect_analyze_stmt): Ditto.
7959 (vect_transform_stmt): Ditto.
7960 (vect_is_simple_use): Ditto.
7961 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
7962 (vect_can_advance_ivs_p): Ditto.
7963 (vect_update_ivs_after_vectorizer): Ditto.
7964 (vect_do_peeling_for_loop_bound): Ditto.
7965 (vect_gen_niters_for_prolog_loop): Ditto.
7966 (vect_update_inits_of_drs): Ditto.
7967 (vect_create_cond_for_alias_checks): Ditto.
7968 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
7969 (vect_build_slp_tree_1): Ditto.
7970 (vect_supported_load_permutation_p): Ditto.
7971 (vect_analyze_slp_instance): Ditto.
7972 (vect_analyze_slp): Ditto.
7973 (vect_make_slp_decision): Ditto.
7974 (vect_detect_hybrid_slp): Ditto.
7975 (vect_bb_vectorization_profitable_p): Ditto.
7976 (vect_slp_analyze_bb_1): Ditto.
7977 (vect_update_slp_costs_according_to_vf): Ditto.
7978 (vect_get_mask_element): Ditto.
7979 (vect_transform_slp_perm_load): Ditto.
7980 (vect_schedule_slp_instance): Ditto.
7981 (vect_schedule_slp): Ditto.
7982 (vect_slp_transform_bb): Ditto.
7983 * profile.c (read_profile_edge_counts): Ditto.
7984 (compute_branch_probabilities): Ditto.
7985 * coverage.c (get_coverage_counts): Ditto.
7987 2013-09-16 Diego Novillo <dnovillo@google.com>
7989 * tree-core.h: Add missing comment lines from refactoring of tree.h.
7991 2013-09-16 Jan Hubicka <jh@suse.cz>
7993 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7994 abstract functions; for static functions check the presence of body.
7996 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7998 * config/aarch64/aarch64-simd-builtins.def (fma): New.
7999 * config/aarch64/aarch64-simd.md
8000 (aarch64_mla_elt<mode>): New.
8001 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8002 (aarch64_mls_elt<mode>): Likewise.
8003 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8004 (aarch64_fma4_elt<mode>): Likewise.
8005 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8006 (aarch64_fma4_elt_to_128v2df): Likewise.
8007 (aarch64_fma4_elt_to_64df): Likewise.
8008 (fnma<mode>4): Likewise.
8009 (aarch64_fnma4_elt<mode>): Likewise.
8010 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8011 (aarch64_fnma4_elt_to_128v2df): Likewise.
8012 (aarch64_fnma4_elt_to_64df): Likewise.
8013 * config/aarch64/iterators.md (VDQSF): New.
8014 * config/aarch64/arm_neon.h
8015 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
8016 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8018 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8020 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8021 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8022 (aarch64_mul3_elt_to_128df): Likewise.
8023 (aarch64_mul3_elt_to_64v2df): Likewise.
8024 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8026 (VMUL_CHANGE_NLANES) Likewise.
8029 * config/aarch64/arm_neon.h
8030 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8032 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8034 * config/aarch64/arm_neon.h
8035 (vcvtx_high_f32_f64): Fix parameters.
8037 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8038 Uros Bizjak <ubizjak@gmail.com>
8040 * config/alpha.c: Include tree-ssanames.h.
8042 2013-09-16 Richard Biener <rguenther@suse.de>
8044 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8045 (dot_rdg_1): Handle control_dd.
8046 (create_edge_for_control_dependence): New function.
8047 (create_rdg_edges): Add control dependences if asked for.
8048 (build_rdg): Likewise.
8049 (generate_loops_for_partition): If there are not necessary
8050 control stmts remove all their dependencies.
8051 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8052 (distribute_loop): Pass on control dependences.
8053 (tree_loop_distribution): Compute control dependences and remove
8054 restriction on number of loop nodes.
8056 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8058 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8061 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8063 * cse.c (try_const_anchors): Punt on CC modes.
8065 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8067 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8069 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8072 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8075 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8078 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8080 (darwin_objc1_section): Likewise.
8081 (darwin_file_end): Emit Image Info section when required.
8083 2013-09-14 Jan Hubicka <jh@suse.cz>
8085 * tree-into-ssa.c (gate_into_ssa): New.
8086 (pass_data_build_ssa): Use it.
8087 * cgraph.h (expand_thunk): Update prototype.
8088 * cgraphunit.c (analyze_function): Expand thunks early.
8089 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8090 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8091 set lowered flag; do not add new function.
8092 (assemble_thunks_and_aliases): Update.
8093 * tree-ssa.c (gate_init_datastructures): New gate.
8094 (pass_data_init_datastructures): Use it.
8096 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8099 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8100 xmm register set described in the psABI.
8102 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8104 * dwarf2out.c (should_emit_struct_debug): Add check
8105 for type_decl variable is not NULL.
8107 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8109 * config.gcc: Use new winnt-c.c target hooks
8110 * config/t-winnt: New file
8111 * config/winnt-c.c: New file
8112 * doc/tm.texi.in: Document new hook
8113 * doc/tm.texi: Regenerated
8115 2013-09-13 Jan Hubicka <jh@suse.cz>
8118 * ipa-inline.c (check_callers): New function.
8119 (check_caller_edge): Remove.
8120 (want_inline_function_to_all_callers_p): Also permit alises that are
8122 (inline_to_all_callers): Terminate the walk when devirtualization
8123 introduce new calls.
8125 2013-09-13 Jan Hubicka <jh@suse.cz>
8127 * ipa-inline-analysis.c (struct growth_data): Add node.
8128 (do_estimate_growth_1): Fix detection of recursion.
8130 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8132 PR tree-optimization/58392
8133 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8134 to avoid shadowing of outer loop variable. If
8135 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8136 replace_by_duplicate_decl simduid of loops that have it set and
8137 set dest_cfun->has_simduid_loops and/or
8138 dest_cfun->has_force_vect_loops.
8139 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8140 instead of maybe_lookup_decl.
8141 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8142 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8143 call. Copy over force_vect and copy and remap simduid. Set
8144 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8145 (copy_cfg_body): Remove blocks_to_copy argument. Use
8146 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8147 caller. Don't set cfun->has_simduid_loops and/or
8148 cfun->has_force_vect_loops here.
8149 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8151 (expand_call_inline, tree_function_versioning): Adjust copy_body
8154 2013-09-13 Martin Jambor <mjambor@suse.cz>
8157 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8158 the assert if the edge was a speculative one.
8160 2013-09-13 Richard Biener <rguenther@suse.de>
8162 * tree-data-ref.h (known_dependences_p): Move here ...
8163 * tree-loop-distribution.c (known_dependences_p): ... from here.
8164 (dump_rdg_component, debug_rdg_component): Remove.
8166 (generate_loops_for_partition): Use gimple_uid instead of
8167 relying on matching stmt visit order.
8168 (rdg_build_partitions): Take starting stmt vector.
8169 (ldist_gen): Merge into ...
8170 (distribute_loop): ... this function. Do not compute starting
8172 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
8174 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8176 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
8177 Set type attribute correctly. Set predicable_short_it attribute.
8178 (cmpsi_shiftsi): Remove %? from output template.
8180 2013-09-13 Richard Biener <rguenther@suse.de>
8182 * tree-loop-distribution.c (struct rdg_component,
8183 rdg_defs_used_in_other_loops_p, free_rdg_components,
8184 rdg_build_components): Remove.
8185 (stmts_from_loop): Do not record virtual PHIs.
8186 (generate_loops_for_partition): Skip virtual PHIs.
8187 (build_rdg_partition_for_component): Rename to ...
8188 (build_rdg_partition_for_vertex): ... this and adjust.
8189 (rdg_build_partitions): Take a vector of starting vertices
8190 instead of components. Remove unnecessary leftover handling.
8191 (ldist_gen): Do not build components or record other stores.
8192 (distribute_loop): Do not distribute loops containing stmts
8195 2013-09-13 Christian Bruel <christian.bruel@st.com>
8198 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
8200 2013-09-13 Kai Tietz <ktietz@redhat.com>
8202 * config.gcc: Separate cases for mingw and cygwin targets,
8203 and add 64-bit cygwin target case.
8205 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
8206 dll-export inline-functions.
8207 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
8210 2013-09-13 Jeff Law <law@redhat.com>
8214 2013-09-06 Jeff Law <law@redhat.com>
8216 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8217 edge implied equivalences into successor phis.
8219 2013-09-12 DJ Delorie <dj@redhat.com>
8221 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8224 2013-09-12 Brooks Moses <bmoses@google.com>
8227 * Makefile.in: Do not install driver binaries in $(target)/bin.
8229 2013-09-12 DJ Delorie <dj@redhat.com>
8231 * config/rl78/rl78.opt (mrelax): New.
8232 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8233 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8235 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8237 (rl78_expand_epilogue): Likewise.
8239 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8242 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8243 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8244 pointer elimination with using remove_reg_equal_offset_note.
8246 2013-09-12 DJ Delorie <dj@redhat.com>
8248 * config/msp430/: New port.
8249 * config.gcc (msp430): Added.
8250 * doc/invoke.texi: Document MSP430 options.
8251 * doc/install.texi: Document msp430-elf
8252 * doc/md.texi: Document msp430-elf
8253 * doc/contrib.texi: Document msp430-elf
8255 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8258 2013-09-12 Martin Jambor <mjambor@suse.cz>
8261 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8262 reference descriptor is NULL.
8263 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8266 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8268 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8269 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8270 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8271 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8272 make_temp_ssa_name): move to tree-ssanames.h
8273 * tree-ssa-alias.h: Move prototype.
8274 * tree-ssa.h: Include tree-ssanames.h.
8275 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8276 * tree-ssanames.h: New. Move items from tree-flow*.h
8277 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8279 2013-09-12 Richard Biener <rguenther@suse.de>
8281 PR tree-optimization/58404
8282 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8283 propagate non-invariant addresses into dereferences wrapped
8284 in component references.
8286 2013-09-12 Richard Biener <rguenther@suse.de>
8288 PR tree-optimization/58402
8289 * passes.def: Move pass_late_warn_uninitialized later.
8291 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8293 * tree-ssa.h: New. Move content from tree-flow.h and
8295 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8296 Move prototypes belonging to tree-ssa.c.
8297 * tree-flow-inline.h (redirect_edge_var_map_def,
8298 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8300 * gimple.h: Adjust prototypes.
8301 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8303 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8304 * tree.h: Move prototype to tree-ssa.h.
8305 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8306 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8307 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8308 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8309 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8310 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8311 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8312 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8313 graphite-blocking.c, graphite-clast-to-gimple.c,
8314 graphite-dependences.c, graphite-interchange.c,
8315 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8316 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8317 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8318 ipa-reference.c, ipa-split.c, ipa-utils.c,
8319 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8320 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8321 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8322 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8323 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8324 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8325 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8326 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8327 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8328 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8329 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8330 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8331 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8332 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8333 value-prof.c, var-tracking.c,
8334 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8336 2013-09-12 Richard Biener <rguenther@suse.de>
8338 PR tree-optimization/58396
8339 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8340 (build_rdg): Take a loop-nest parameter, fix memleaks.
8341 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8343 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8345 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8348 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8350 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8351 The FMA instruction names should have a 'v' prefix.
8353 2013-09-12 Richard Biener <rguenther@suse.de>
8355 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8356 (dot_rdg): Use popen instead of system in optional code.
8357 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8358 (already_processed_vertex_p): Adjust.
8359 (has_anti_or_output_dependence, predecessor_has_mem_write,
8360 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8361 rdg_flag_uses): Remove.
8362 (rdg_flag_vertex): Simplify.
8363 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8365 (build_rdg_partition_for_component): Process the first vertex
8366 of a component only.
8367 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8369 2013-09-12 Alan Modra <amodra@gmail.com>
8371 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8373 2013-09-11 DJ Delorie <dj@redhat.com>
8374 Nick Clifton <nickc@redhat.com>
8376 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8377 (rl78_stack_based_mem): New.
8378 * config/rl78/constraints.md (Iv08): New.
8387 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8389 (movsi): Change from expand to insn-and-split.
8390 (ashrsi3): Clobber AX.
8394 * config/rl78/rl78.md (CC_REG): Fix.
8395 (addsi3): Allow memory and immediate operands.
8396 (addsi3_internal): Split into...
8397 (addsi3_internal_virt): ...new, and ...
8398 (addsi3_internal_real): ...new.
8400 (subsi3_internal_virt): New.
8401 (subsi3_internal_real): New.
8402 (mulsi3): Add memory operand.
8403 (mulsi3_rl78): Likewise.
8404 (mulsi3_g13): Likewise.
8405 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8406 (cbranchqi4_real): Add more constraint options.
8407 (cbranchhi4_real): Expand pattern.
8408 (cbranchhi4_real_signed): New.
8409 (cbranchhi4_real_inverted): New.
8410 (cbranchsi4_real_lt): New.
8411 (cbranchsi4_real_ge): New.
8412 (cbranchsi4_real_signed): New.
8413 (cbranchsi4_real): New.
8415 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8417 (lshrsi3_virt): Likewise.
8418 (ashlsi3_virt): Likewise.
8419 (cbranchqi4_virt_signed): New.
8420 (cbranchhi4_virt_signed): New.
8421 (cbranchsi4_virt): New.
8422 * config/rl78/rl78.c: Whitespace fixes throughout.
8423 (move_elim_pass): New.
8424 (pass_data_rl78_move_elim): New.
8425 (pass_rl78_move_elim): New.
8426 (make_pass_rl78_move_elim): New.
8427 (rl78_devirt_info): Run devirt earlier.
8428 (rl78_move_elim_info): New.
8429 (rl78_asm_file_start): Register it.
8430 (rl78_split_movsi): New.
8431 (rl78_as_legitimate_address): Allow virtual base registers when
8433 (rl78_addr_space_convert): Remove spurious debug stuff.
8434 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8435 (rl78_print_operand): More cases for not printing '#'.
8436 (rl78_expand_compare): Remove most of the logic.
8437 (content_memory): New.
8438 (clear_content_memory): New.
8439 (get_content_index): New.
8440 (get_content_name): New.
8441 (display_content_memory): New.
8442 (update_content): New.
8443 (record_content): New.
8444 (already_contains): New.
8445 (insn_ok_now): Re-recog insns with virtual registers.
8446 (add_postponed_content_update): New.
8447 (process_postponed_content_update): New.
8448 (gen_and_emit_move): New.
8449 (transcode_memory_rtx): Record new location content.
8450 Use gen_and_emit_move.
8451 (force_into_acc): New.
8452 (move_to_acc): Use gen_and_emit_move.
8453 (move_from_acc): Likewise.
8454 (move_acc_to_reg): Likewise.
8455 (move_to_x): Likewise.
8456 (move_to_hl): Likewise.
8457 (move_to_de): Likewise.
8458 (rl78_alloc_physical_registers_op1): Record location content.
8459 (has_constraint): New.
8460 (rl78_alloc_physical_registers_op2): Record location content.
8462 (rl78_alloc_physical_registers_ro1): Likewise.
8463 (rl78_alloc_physical_registers_cmp): Likewise.
8464 (rl78_alloc_physical_registers_umul): Likewise.
8465 (rl78_alloc_address_registers_macax): New.
8466 (rl78_alloc_physical_registers): Initialize and set location
8467 content memory as needed.
8468 (rl78_reorg): Make sure split2 is called.
8469 (rl78_rtx_costs): New.
8471 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8473 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8474 for (not (neg ...)) and (neg (not ...)) cases.
8476 2013-09-11 Richard Biener <rguenther@suse.de>
8479 * passes.def: Split critical edges before late uninit warning passes.
8480 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8482 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8484 PR tree-optimization/58385
8485 * fold-const.c (build_range_check): If both low and high are NULL,
8486 use omit_one_operand_loc to preserve exp side-effects.
8488 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8490 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8492 2013-09-11 Richard Biener <rguenther@suse.de>
8494 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8495 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8496 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8497 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8498 stmts_from_loop, known_dependences_p, build_empty_rdg,
8499 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8500 * tree-loop-distribution.c: ... here.
8501 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8502 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8503 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8504 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8505 * tree-loop-distribution.c: ... here.
8506 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8507 (struct partition_s): Add loops member.
8508 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8509 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8510 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8512 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8513 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8514 Sergey Lega <sergey.s.lega@intel.com>
8515 Anna Tikhonova <anna.tikhonova@intel.com>
8516 Ilya Tocar <ilya.tocar@intel.com>
8517 Andrey Turetskiy <andrey.turetskiy@intel.com>
8518 Ilya Verbin <ilya.verbin@intel.com>
8519 Kirill Yukhin <kirill.yukhin@intel.com>
8520 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8522 * config/i386/constraints.md (k): New.
8524 * config/i386/i386.c (const regclass_map): Add new mask registers.
8525 (dbx_register_map): Ditto.
8526 (dbx64_register_map): Ditto.
8527 (svr4_dbx_register_map): Ditto.
8528 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8530 (ix86_preferred_reload_class): Disable constants for mask registers.
8531 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8532 (ix86_hard_regno_mode_ok): Support new mask registers.
8533 (x86_order_regs_for_local_alloc): Ditto.
8534 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8535 (FIXED_REGISTERS): Add new mask registers.
8536 (CALL_USED_REGISTERS): Ditto.
8537 (REG_ALLOC_ORDER): Ditto.
8538 (VALID_MASK_REG_MODE): New.
8539 (FIRST_MASK_REG): Ditto.
8540 (LAST_MASK_REG): Ditto.
8541 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8542 (MAYBE_MASK_CLASS_P): New.
8543 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8544 (REG_CLASS_CONTENTS): Ditto.
8545 (MASK_REGNO_P): New.
8546 (ANY_MASK_REG_P): Ditto.
8547 (HI_REGISTER_NAMES): Add new mask registers.
8548 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8549 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8551 (attribute "type"): Add mskmov, msklog.
8552 (attribute "length_immediate"): Support them.
8553 (attribute "memory"): Ditto.
8554 (attribute "prefix_0f"): Ditto.
8555 (*movhi_internal): Support new mask registers.
8556 (*movqi_internal): Ditto.
8557 (define_split): Split out clobber pattern is a logic
8558 insn on mask registers.
8559 (*k<logic><mode>): New.
8560 (*andhi_1): Extend to support mask regs.
8561 (*andqi_1): Extend to support mask regs.
8563 (define_split): Split and-not to and and not if operands
8565 (*<code><mode>_1): Separate HI mode to new pattern...
8566 (*<code>hi_1): This.
8567 (*<code>qi_1): Extend to support mask regs.
8569 (kortestzhi): Ditto.
8570 (kortestchi): Ditto.
8572 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8573 (*one_cmplhi2_1): ...Here, now with mask registers support.
8574 (*one_cmplqi2_1): Support new mask registers.
8575 (HI/QImode arithmetics splitter): Don't split if mask registers
8577 (HI/QImode not splitter): Ditto.
8578 * config/i386/predicated.md (mask_reg_operand): New.
8579 (general_reg_operand): Ditto.
8581 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8583 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8584 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8586 2013-09-10 Jeff Law <law@redhat.com>
8588 PR tree-optimization/58380
8589 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8590 of threading through a buried loop header.
8592 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8593 return value for single successor case.
8595 2013-09-10 Jan Hubicka <jh@suse.cz>
8597 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8599 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8602 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8603 support conditional execution.
8604 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8606 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8608 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8610 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8611 Check destination too to check move insn.
8612 (undo_optional_reloads): Add check that the original peudo did not
8613 changed its allocation and the optional reload was inherited on last
8614 inheritance pass. Break loop after deciding to keep optional reload.
8615 (lra_undo_inheritance): Add check that inherited pseudo still in
8618 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8620 * config/aarch64/aarch64.md (generic_sched): New.
8621 * config/aarch64/aarch64-generic.md (load): Make conditional
8622 on generic_sched attribute.
8623 (nonload): Likewise.
8625 2013-09-10 Jan Hubicka <jh@suse.cz>
8627 * lto-cgraph.c: Include ipa-utils.h.
8628 (compute_ltrans_boundary): Also add possible targets into the boundary.
8630 2013-09-10 Jan Hubicka <jh@suse.cz>
8632 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
8633 VAR_DECL of vtable rather than full expression.
8635 2013-09-10 Jan Hubicka <jh@suse.cz>
8636 Paolo Carlini <paolo.carlini@oracle.com>
8638 * cgraphunit.c (analyze_functions): Save input_location, set it
8639 to UNKNOWN_LOCATION and restore it at the end.
8641 2013-09-10 Martin Jambor <mjambor@suse.cz>
8643 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
8644 represented by a thunk.
8646 2013-09-10 Jeff Law <law@redhat.com>
8648 PR tree-optimization/58343
8649 * tree-ssa-threadupdate.c (thread_block): Identify and disable
8650 jump threading requests through loop headers buried in the middle
8651 of a jump threading path.
8653 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
8654 in return value/type.
8656 2013-09-10 Jakub Jelinek <jakub@redhat.com>
8658 PR rtl-optimization/58365
8659 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
8660 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
8663 2013-09-10 Richard Biener <rguenther@suse.de>
8665 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
8666 * tree-data-ref.c (create_rdg_vertices): Collect all data
8667 references, signal failure to the caller, use data-ref API.
8668 (build_rdg): Compute data references only once. Maintain lifetime
8669 of data references and data dependences from within RDG.
8670 (free_rdg): Free dependence relations.
8671 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
8672 inventing extra dependences.
8673 (distribute_loop): Update for RDG API changes.
8675 2013-09-10 Kai Tietz <ktietz@redhat.com>
8677 * doc/invoke.texi (fms-extensions): Document changed
8678 behavior for ms-abi targets.
8679 * config/i386/i386.c (ix86_option_override_internal):
8680 Set default value of option -fms-extension for ms-abi targets.
8682 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8684 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
8686 2013-09-10 Alan Modra <amodra@gmail.com>
8689 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
8691 2013-09-10 Alan Modra <amodra@gmail.com>
8693 * config/rs6000/predicates.md (add_cint_operand): New.
8694 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
8695 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
8696 using add_cint_operand.
8697 (largetoc_high_plus_aix): Likewise.
8699 2013-09-09 Jakub Jelinek <jakub@redhat.com>
8701 PR tree-optimization/58364
8702 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
8703 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
8704 the current range can't be an unconditional true or false.
8706 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
8708 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
8710 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
8712 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
8714 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
8717 * doc/invoke.texi (-Wdelete-incomplete): Document it.
8719 2013-09-09 Ian Bolton <ian.bolton@arm.com>
8721 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
8722 NO_REGS for immediate that can't be moved directly into FP_REGS.
8724 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8726 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
8727 comparison with negated operand.
8728 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
8731 2013-09-09 Richard Biener <rguenther@suse.de>
8734 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
8735 of a subloop record all its block as affecting loop-closed SSA form.
8737 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8739 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
8740 of an rtx/bitpos pair.
8741 (store_fixed_bit_field): Update accordingly.
8743 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8745 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
8747 * builtins.c (expand_errno_check): Likewise.
8748 * dwarf2cfi.c (init_return_column_size): Likewise.
8749 * except.c (sjlj_mark_call_sites): Likewise.
8750 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
8751 * lra-constraints.c (emit_inc): Likewise.
8752 * ree.c (combine_set_extension): Likewise.
8753 * regmove.c (fixup_match_2): Likewise.
8754 * reload1.c (inc_for_reload): Likewise.
8756 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8758 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
8759 (force_to_mode, simplify_shift_const_1, simplify_comparison):
8760 Use gen_int_mode with the mode of the associated simplify_* call.
8761 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
8762 * expmed.c (expand_shift_1): Likewise.
8763 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8764 * loop-iv.c (iv_number_of_iterations): Likewise.
8765 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8766 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
8768 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8770 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
8771 of the associated expand_* call.
8772 (asan_emit_stack_protection): Likewise.
8773 * builtins.c (round_trampoline_addr): Likewise.
8774 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
8775 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
8776 (emit_store_flag): Likewise.
8777 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
8778 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
8780 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8781 * ifcvt.c (noce_try_store_flag_constants): Likewise.
8782 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8783 * modulo-sched.c (generate_prolog_epilog): Likewise.
8784 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
8785 (expand_ctz, expand_ffs, expand_unop): Likewise.
8787 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8789 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
8790 of the associated gen_rtx_* call.
8791 * caller-save.c (init_caller_save): Likewise.
8792 * combine.c (find_split_point, make_extraction): Likewise.
8793 (make_compound_operation): Likewise.
8794 * dwarf2out.c (mem_loc_descriptor): Likewise.
8795 * explow.c (plus_constant, probe_stack_range): Likewise.
8796 * expmed.c (expand_mult_const): Likewise.
8797 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
8798 * reload1.c (init_reload): Likewise.
8799 * valtrack.c (cleanup_auto_inc_dec): Likewise.
8800 * var-tracking.c (adjust_mems): Likewise.
8801 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
8802 rather than gen_rtx_fmt_ee.
8804 2013-09-09 Jan Hubicka <jh@suse.cz>
8807 * value-prof.c (gimple_ic): Copy also abnormal edges.
8809 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8811 * asan.c (asan_shadow_cst): Use gen_int_mode.
8813 2013-09-08 Jan Hubicka <jh@suse.cz>
8815 * ipa-profile.c: Add toplevel comment.
8816 (ipa_propagate_frequency_1): Be more conservative when profile is read.
8817 (contains_hot_call_p): New function.
8818 (ipa_propagate_frequency): Set frequencies based on counts when
8820 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8821 profile; do not tamper with profile after inlining if it is read.
8823 2013-09-08 Jan Hubicka <jh@suse.cz>
8825 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8828 2013-09-08 Jan Hubicka <jh@suse.cz>
8830 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8833 2013-09-08 Jeff Law <law@redhat.com>
8836 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8839 2013-09-08 Andi Kleen <ak@linux.intel.com>
8841 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8843 2013-09-08 Jan Hubicka <jh@suse.cz>
8845 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8847 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8849 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8850 for non-debug insns.
8851 * lra.c (new_insn_reg): Take the containing insn as a parameter.
8852 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
8853 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
8856 2013-09-08 Jan Hubicka <jh@suse.cz>
8858 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
8859 targets and devirtualize to BUILT_IN_UNREACHABLE.
8860 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
8861 * ipa.c (walk_polymorphic_call_targets): New function.
8862 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
8863 functions; use the new timevar.
8864 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
8865 was removed from the program.
8866 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
8867 not consider it in the walk when its vtable is dead.
8868 (possible_polymorphic_call_targets_1): Pass anonymous flag to
8870 (devirt_variable_node_removal_hook): New function.
8871 (possible_polymorphic_call_targets): Also register
8872 devirt_variable_node_removal_hook.
8873 (ipa_devirt): Do not do non-speculative devirtualization.
8874 (gate_ipa_devirt): One execute if devirtualizing speculatively.
8876 2013-09-08 Jan Hubicka <jh@suse.cz>
8878 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
8879 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
8880 varpool_remove_variable_insertion_hook): Declare.
8881 * varpool.c (varpool_node_hook_list): New structure.
8882 (first_varpool_node_removal_hook,
8883 first_varpool_variable_insertion_hook): New variables.
8884 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
8885 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
8886 varpool_remove_variable_insertion_hook,
8887 varpool_call_variable_insertion_hooks): New functions.
8888 (varpool_remove_node): Use it.
8890 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
8893 * diagnostic.c (diagnostic_build_prefix): When s.file is
8894 "<built-in>" don't output line and column numbers.
8896 2013-09-06 Jan Hubicka <jh@suse.cz>
8898 * cgraphunit.c (expand_thunk): Get body before touching arguments.
8899 * lto-streamer-out.c: Stream thunks, too.
8900 * lto-streamer-in.c (input_function): Pop cfun here
8901 (lto_read_body): Instead of here.
8903 2013-09-06 Caroline Tice <cmtice@google.com>
8905 * doc/install.texi: Add documentation for the --enable-vtable-verify
8906 and the --disable-libvtv configure options.
8908 2013-09-06 Jeff Law <law@redhat.com>
8910 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8911 edge implied equivalences into successor phis.
8913 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
8915 * resource.c (mark_referenced_resources): Handle COND_EXEC.
8917 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
8919 * resource.c (mark_target_live_regs): Compute resources taking
8920 into account if a call is predicated or not.
8922 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
8924 * toplev.c (output_stack_usage): Be prepared for suffixes created by
8925 the compiler in the function names.
8927 2013-09-06 Jan Hubicka <jh@suse.cz>
8930 * ipa-inline.c (has_caller_p): New function.
8931 (want_inline_function_to_all_callers_p): Use it.
8932 (sum_callers, inline_to_all_callers): Break out from ...
8933 (ipa_inline): ... here.
8935 2013-09-06 Jan Hubicka <jh@suse.cz>
8937 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
8938 only when AVX is enabled.
8940 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8942 * config/aarch64/aarch64.md
8943 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
8944 is fpsimd_<load/store>2.
8945 (load_pair<mode>): Likewise.
8946 (store_pair<mode>): Likewise.
8948 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8950 * config/arm/types.md (type): Add "mrs" type.
8951 * config/aarch64/aarch64.md
8952 (aarch64_load_tp_hard): Make type "mrs".
8954 (load_tp_hard): Make type "mrs".
8955 * config/arm/cortex-a15.md: Update with new attributes.
8956 * config/arm/cortex-a5.md: Update with new attributes.
8957 * config/arm/cortex-a53.md: Update with new attributes.
8958 * config/arm/cortex-a7.md: Update with new attributes.
8959 * config/arm/cortex-a8.md: Update with new attributes.
8960 * config/arm/cortex-a9.md: Update with new attributes.
8961 * config/arm/cortex-m4.md: Update with new attributes.
8962 * config/arm/cortex-r4.md: Update with new attributes.
8963 * config/arm/fa526.md: Update with new attributes.
8964 * config/arm/fa606te.md: Update with new attributes.
8965 * config/arm/fa626te.md: Update with new attributes.
8966 * config/arm/fa726te.md: Update with new attributes.
8968 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8970 * config/aarch64/aarch64.md
8971 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
8972 (*movtf_aarch64): Likewise.
8974 (thumb1_movdi_insn): Use "multiple" for type where more than one
8975 instruction is used for a move.
8976 (*arm32_movhf): Likewise.
8977 (*thumb_movdf_insn): Likewise.
8979 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8981 * config/arm/types.md (type): Rename fcpys to fmov.
8983 (*arm_movsi_vfp): Rename type fcpys as fmov.
8984 (*thumb2_movsi_vfp): Likewise
8985 (*movhf_vfp_neon): Likewise
8986 (*movhf_vfp): Likewise
8987 (*movsf_vfp): Likewise
8988 (*thumb2_movsf_vfp): Likewise
8989 (*movsfcc_vfp): Likewise
8990 (*thumb2_movsfcc_vfp): Likewise
8991 * config/aarch64/aarch64-simd.md
8992 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
8993 * config/aarch64/aarch64.md
8994 (*movsi_aarch64): Replace type mov_reg with fmovs.
8995 (*movdi_aarch64): Likewise
8996 (*movsf_aarch64): Likewise
8997 (*movdf_aarch64): Likewise
8999 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
9000 * config/arm/iwmmxt.md
9001 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
9002 * config/arm/arm1020e.md: Update with new attributes.
9003 * config/arm/cortex-a15-neon.md: Update with new attributes.
9004 * config/arm/cortex-a5.md: Update with new attributes.
9005 * config/arm/cortex-a53.md: Update with new attributes.
9006 * config/arm/cortex-a7.md: Update with new attributes.
9007 * config/arm/cortex-a8-neon.md: Update with new attributes.
9008 * config/arm/cortex-a9.md: Update with new attributes.
9009 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9010 * config/arm/cortex-r4f.md: Update with new attributes.
9011 * config/arm/marvell-pj4.md: Update with new attributes.
9012 * config/arm/vfp11.md: Update with new attributes.
9014 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9016 * config/aarch64/aarch64.md
9017 (*madd<mode>): Fix type attribute.
9018 (*maddsi_uxtw): Likewise.
9019 (*msub<mode>): Likewise.
9020 (*msubsi_uxtw): Likewise.
9021 (<su_optab>maddsidi4): Likewise.
9022 (<su_optab>msubsidi4): Likewise.
9024 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9026 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9027 * config/arm/arm.md (core_cycles): Remove fdiv.
9028 * config/arm/vfp.md:
9029 (*sqrtsf2_vfp): Update for attribute changes.
9030 (*sqrtdf2_vfp): Likewise.
9031 * config/aarch64/aarch64.md:
9032 (sqrt<mode>2): Update for attribute changes.
9033 * config/arm/arm1020e.md: Update with new attributes.
9034 * config/arm/cortex-a15-neon.md: Update with new attributes.
9035 * config/arm/cortex-a5.md: Update with new attributes.
9036 * config/arm/cortex-a53.md: Update with new attributes.
9037 * config/arm/cortex-a7.md: Update with new attributes.
9038 * config/arm/cortex-a8-neon.md: Update with new attributes.
9039 * config/arm/cortex-a9.md: Update with new attributes.
9040 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9041 * config/arm/cortex-r4f.md: Update with new attributes.
9042 * config/arm/marvell-pj4.md: Update with new attributes.
9043 * config/arm/vfp11.md: Update with new attributes.
9045 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9047 * config/arm/types.md
9048 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9049 * config/aarch64/aarch64.md
9050 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9052 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9053 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9054 (float<GPI:mode><GPF:mode>2): Likewise.
9056 (*truncsisf2_vfp): Update with new attributes.
9057 (*truncsidf2_vfp): Likewise.
9058 (fixuns_truncsfsi2): Likewise.
9059 (fixuns_truncdfsi2): Likewise.
9060 (*floatsisf2_vfp): Likewise.
9061 (*floatsidf2_vfp): Likewise.
9062 (floatunssisf2): Likewise.
9063 (floatunssidf2): Likewise.
9064 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9065 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9066 * config/arm/arm1020e.md: Update with new attributes.
9067 * config/arm/cortex-a15-neon.md: Update with new attributes.
9068 * config/arm/cortex-a5.md: Update with new attributes.
9069 * config/arm/cortex-a53.md: Update with new attributes.
9070 * config/arm/cortex-a7.md: Update with new attributes.
9071 * config/arm/cortex-a8-neon.md: Update with new attributes.
9072 * config/arm/cortex-a9.md: Update with new attributes.
9073 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9074 * config/arm/cortex-r4f.md: Update with new attributes.
9075 * config/arm/marvell-pj4.md: Update with new attributes.
9076 * config/arm/vfp11.md: Update with new attributes.
9078 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9080 * config/aarch64/arm_neon.h
9081 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9082 (vqtbx<1,2,3,4><q>_s8): Likewise.
9084 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9086 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9087 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9088 (add<mode>3): Add type attribute.
9089 (add<mode>3): Likewise.
9090 (usadd<mode>3): Likewise.
9091 (ssadd<mode>3): Likewise.
9092 (sub<mode>3): Likewise.
9093 (sub<mode>3): Likewise.
9094 (ussub<mode>3): Likewise.
9095 (sssub<mode>3): Likewise.
9096 (ssmulsa3): Likewise.
9097 (usmulusa3): Likewise.
9098 (arm_usatsihi): Likewise.
9100 (*movdi_vfp): Add types for all instructions.
9101 (*movdi_vfp_cortexa8): Likewise.
9102 (*movhf_vfp_neon): Likewise.
9103 (*movhf_vfp): Likewise.
9104 (*movdf_vfp): Likewise.
9105 (*thumb2_movdf_vfp): Likewise.
9106 (*thumb2_movdfcc_vfp): Likewise.
9107 * config/arm/arm.md: Add type attribute to all insn patterns.
9108 (*thumb1_adddi3): Add type attribute.
9109 (*arm_adddi3): Likewise.
9110 (*adddi_sesidi_di): Likewise.
9111 (*adddi_zesidi_di): Likewise.
9112 (*thumb1_addsi3): Likewise.
9113 (addsi3_compare0): Likewise.
9114 (*addsi3_compare0_scratch): Likewise.
9115 (*compare_negsi_si): Likewise.
9116 (cmpsi2_addneg): Likewise.
9117 (*addsi3_carryin_<optab>): Likewise.
9118 (*addsi3_carryin_alt2_<optab>): Likewise.
9119 (*addsi3_carryin_clobercc_<optab>): Likewise.
9120 (*subsi3_carryin): Likewise.
9121 (*subsi3_carryin_const): Likewise.
9122 (*subsi3_carryin_compare): Likewise.
9123 (*subsi3_carryin_compare_const): Likewise.
9124 (*arm_subdi3): Likewise.
9125 (*thumb_subdi3): Likewise.
9126 (*subdi_di_zesidi): Likewise.
9127 (*subdi_di_sesidi): Likewise.
9128 (*subdi_zesidi_di): Likewise.
9129 (*subdi_sesidi_di): Likewise.
9130 (*subdi_zesidi_ze): Likewise.
9131 (thumb1_subsi3_insn): Likewise.
9132 (*arm_subsi3_insn): Likewise.
9133 (*anddi3_insn): Likewise.
9134 (*anddi_zesidi_di): Likewise.
9135 (*anddi_sesdi_di): Likewise.
9136 (*ne_zeroextracts): Likewise.
9137 (*ne_zeroextracts): Likewise.
9138 (*ite_ne_zeroextr): Likewise.
9139 (*ite_ne_zeroextr): Likewise.
9140 (*anddi_notdi_di): Likewise.
9141 (*anddi_notzesidi): Likewise.
9142 (*anddi_notsesidi): Likewise.
9143 (andsi_notsi_si): Likewise.
9144 (thumb1_bicsi3): Likewise.
9145 (*iordi3_insn): Likewise.
9146 (*iordi_zesidi_di): Likewise.
9147 (*iordi_sesidi_di): Likewise.
9148 (*thumb1_iorsi3_insn): Likewise.
9149 (*xordi3_insn): Likewise.
9150 (*xordi_zesidi_di): Likewise.
9151 (*xordi_sesidi_di): Likewise.
9152 (*arm_xorsi3): Likewise.
9153 (*andsi_iorsi3_no): Likewise.
9154 (*smax_0): Likewise.
9155 (*smax_m1): Likewise.
9156 (*arm_smax_insn): Likewise.
9157 (*smin_0): Likewise.
9158 (*arm_smin_insn): Likewise.
9159 (*arm_umaxsi3): Likewise.
9160 (*arm_uminsi3): Likewise.
9161 (*minmax_arithsi): Likewise.
9162 (*minmax_arithsi_): Likewise.
9163 (*satsi_<SAT:code>): Likewise.
9164 (arm_ashldi3_1bit): Likewise.
9165 (arm_ashrdi3_1bit): Likewise.
9166 (arm_lshrdi3_1bit): Likewise.
9167 (*arm_negdi2): Likewise.
9168 (*thumb1_negdi2): Likewise.
9169 (*arm_negsi2): Likewise.
9170 (*thumb1_negsi2): Likewise.
9171 (*negdi_extendsid): Likewise.
9172 (*negdi_zero_extend): Likewise.
9173 (*arm_abssi2): Likewise.
9174 (*thumb1_abssi2): Likewise.
9175 (*arm_neg_abssi2): Likewise.
9176 (*thumb1_neg_abss): Likewise.
9177 (one_cmpldi2): Likewise.
9178 (extend<mode>di2): Likewise.
9179 (*compareqi_eq0): Likewise.
9180 (*arm_extendhisi2addsi): Likewise.
9181 (*arm_movdi): Likewise.
9182 (*thumb1_movdi_insn): Likewise.
9183 (*arm_movt): Likewise.
9184 (*thumb1_movsi_insn): Likewise.
9185 (pic_add_dot_plus_four): Likewise.
9186 (pic_add_dot_plus_eight): Likewise.
9187 (tls_load_dot_plus_eight): Likewise.
9188 (*thumb1_movhi_insn): Likewise.
9189 (*thumb1_movsf_insn): Likewise.
9190 (*movdf_soft_insn): Likewise.
9191 (*thumb_movdf_insn): Likewise.
9192 (cbranchsi4_insn): Likewise.
9193 (cbranchsi4_scratch): Likewise.
9194 (*negated_cbranchsi4): Likewise.
9195 (*tbit_cbranch): Likewise.
9196 (*tlobits_cbranch): Likewise.
9197 (*tstsi3_cbranch): Likewise.
9198 (*cbranchne_decr1): Likewise.
9199 (*addsi3_cbranch): Likewise.
9200 (*addsi3_cbranch_scratch): Likewise.
9201 (*arm_cmpdi_insn): Likewise.
9202 (*arm_cmpdi_unsig): Likewise.
9203 (*arm_cmpdi_zero): Likewise.
9204 (*thumb_cmpdi_zero): Likewise.
9205 (*deleted_compare): Likewise.
9206 (*mov_scc): Likewise.
9207 (*mov_negscc): Likewise.
9208 (*mov_notscc): Likewise.
9209 (*cstoresi_eq0_thumb1_insn): Likewise.
9210 (cstoresi_nltu_thumb1): Likewise.
9211 (cstoresi_ltu_thu): Likewise.
9212 (thumb1_addsi3_addgeu): Likewise.
9213 (*arm_jump): Likewise.
9214 (*thumb_jump): Likewise.
9215 (*check_arch2): Likewise.
9216 (arm_casesi_internal): Likewise.
9217 (thumb1_casesi_dispatch): Likewise.
9218 (*arm_indirect_jump): Likewise.
9219 (*thumb1_indirect_jump): Likewise.
9221 (*and_scc): Likewise.
9222 (*ior_scc): Likewise.
9223 (*compare_scc): Likewise.
9224 (*cond_move): Likewise.
9225 (*cond_arith): Likewise.
9226 (*cond_sub): Likewise.
9227 (*cmp_ite0): Likewise.
9228 (*cmp_ite1): Likewise.
9229 (*cmp_and): Likewise.
9230 (*cmp_ior): Likewise.
9231 (*ior_scc_scc): Likewise.
9232 (*ior_scc_scc_cmp): Likewise.
9233 (*and_scc_scc): Likewise.
9234 (*and_scc_scc_cmp): Likewise.
9235 (*and_scc_scc_nod): Likewise.
9236 (*negscc): Likewise.
9237 (movcond_addsi): Likewise.
9238 (movcond): Likewise.
9239 (*ifcompare_plus_move): Likewise.
9240 (*if_plus_move): Likewise.
9241 (*ifcompare_move_plus): Likewise.
9242 (*if_move_plus): Likewise.
9243 (*ifcompare_arith_arith): Likewise.
9244 (*if_arith_arith): Likewise.
9245 (*ifcompare_arith_move): Likewise.
9246 (*if_arith_move): Likewise.
9247 (*ifcompare_move_arith): Likewise.
9248 (*if_move_arith): Likewise.
9249 (*ifcompare_move_not): Likewise.
9250 (*if_move_not): Likewise.
9251 (*ifcompare_not_move): Likewise.
9252 (*if_not_move): Likewise.
9253 (*ifcompare_shift_move): Likewise.
9254 (*if_shift_move): Likewise.
9255 (*ifcompare_move_shift): Likewise.
9256 (*if_move_shift): Likewise.
9257 (*ifcompare_shift_shift): Likewise.
9258 (*ifcompare_not_arith): Likewise.
9259 (*ifcompare_arith_not): Likewise.
9260 (*if_arith_not): Likewise.
9261 (*ifcompare_neg_move): Likewise.
9262 (*if_neg_move): Likewise.
9263 (*ifcompare_move_neg): Likewise.
9264 (*if_move_neg): Likewise.
9265 (prologue_thumb1_interwork): Likewise.
9266 (*cond_move_not): Likewise.
9267 (*sign_extract_onebit): Likewise.
9268 (*not_signextract_onebit): Likewise.
9269 (stack_tie): Likewise.
9270 (align_4): Likewise.
9271 (align_8): Likewise.
9272 (consttable_end): Likewise.
9273 (consttable_1): Likewise.
9274 (consttable_2): Likewise.
9275 (consttable_4): Likewise.
9276 (consttable_8): Likewise.
9277 (consttable_16): Likewise.
9278 (*thumb1_tablejump): Likewise.
9279 (prefetch): Likewise.
9280 (force_register_use): Likewise.
9281 (thumb_eh_return): Likewise.
9282 (load_tp_hard): Likewise.
9283 (load_tp_soft): Likewise.
9284 (tlscall): Likewise.
9285 (*arm_movtas_ze): Likewise.
9286 (*arm_rev): Likewise.
9287 (*arm_revsh): Likewise.
9288 (*arm_rev16): Likewise.
9289 * config/arm/thumb2.md
9290 (*thumb2_smaxsi3): Likewise.
9291 (*thumb2_sminsi3): Likewise.
9292 (*thumb32_umaxsi3): Likewise.
9293 (*thumb2_uminsi3): Likewise.
9294 (*thumb2_negdi2): Likewise.
9295 (*thumb2_abssi2): Likewise.
9296 (*thumb2_neg_abss): Likewise.
9297 (*thumb2_movsi_insn): Likewise.
9298 (tls_load_dot_plus_four): Likewise.
9299 (*thumb2_movhi_insn): Likewise.
9300 (*thumb2_mov_scc): Likewise.
9301 (*thumb2_mov_negs): Likewise.
9302 (*thumb2_mov_negs): Likewise.
9303 (*thumb2_mov_nots): Likewise.
9304 (*thumb2_mov_nots): Likewise.
9305 (*thumb2_movsicc_): Likewise.
9306 (*thumb2_movsfcc_soft_insn): Likewise.
9307 (*thumb2_indirect_jump): Likewise.
9308 (*thumb2_and_scc): Likewise.
9309 (*thumb2_ior_scc): Likewise.
9310 (*thumb2_ior_scc_strict_it): Likewise.
9311 (*thumb2_cond_move): Likewise.
9312 (*thumb2_cond_arith): Likewise.
9313 (*thumb2_cond_ari): Likewise.
9314 (*thumb2_cond_sub): Likewise.
9315 (*thumb2_negscc): Likewise.
9316 (*thumb2_movcond): Likewise.
9317 (thumb2_casesi_internal): Likewise.
9318 (thumb2_casesi_internal_pic): Likewise.
9319 (*thumb2_alusi3_short): Likewise.
9320 (*thumb2_mov<mode>_shortim): Likewise.
9321 (*thumb2_addsi_short): Likewise.
9322 (*thumb2_subsi_short): Likewise.
9323 (thumb2_addsi3_compare0): Likewise.
9324 (*thumb2_cbz): Likewise.
9325 (*thumb2_cbnz): Likewise.
9326 (*thumb2_one_cmplsi2_short): Likewise.
9327 (*thumb2_negsi2_short): Likewise.
9328 (*orsi_notsi_si): Likewise.
9329 * config/arm/arm1020e.md: Update with new attributes.
9330 * config/arm/arm1026ejs.md: Update with new attributes.
9331 * config/arm/arm1136jfs.md: Update with new attributes.
9332 * config/arm/arm926ejs.md: Update with new attributes.
9333 * config/arm/cortex-a15.md: Update with new attributes.
9334 * config/arm/cortex-a5.md: Update with new attributes.
9335 * config/arm/cortex-a53.md: Update with new attributes.
9336 * config/arm/cortex-a7.md: Update with new attributes.
9337 * config/arm/cortex-a8.md: Update with new attributes.
9338 * config/arm/cortex-a9.md: Update with new attributes.
9339 * config/arm/cortex-m4.md: Update with new attributes.
9340 * config/arm/cortex-r4.md: Update with new attributes.
9341 * config/arm/fa526.md: Update with new attributes.
9342 * config/arm/fa606te.md: Update with new attributes.
9343 * config/arm/fa626te.md: Update with new attributes.
9344 * config/arm/fa726te.md: Update with new attributes.
9346 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9348 * config/aarch64/aarch64-simd.md
9349 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9350 <vwx> iterator to ensure correct register choice.
9351 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9352 (aarch64_sqdmull_n<mode>): Likewise.
9353 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9354 * config/aarch64/arm_neon.h
9355 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9356 (vml<as><q>_n_<su>16): Likewise.
9357 (vml<as>l_high_lane<q>_<su>16): Likewise.
9358 (vml<as>l_high_n_<su>16): Likewise.
9359 (vml<as>l_lane<q>_<su>16): Likewise.
9360 (vml<as>l_n_<su>16): Likewise.
9361 (vmul<q>_lane<q>_<su>16): Likewise.
9362 (vmul<q>_n_<su>16): Likewise.
9363 (vmull_lane<q>_<su>16): Likewise.
9364 (vmull_n_<su>16): Likewise.
9365 (vmull_high_lane<q>_<su>16): Likewise.
9366 (vmull_high_n_<su>16): Likewise.
9367 (vqrdmulh<q>_n_s16): Likewise.
9369 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9371 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9372 have the correct lane parameter.
9374 2013-09-06 Richard Biener <rguenther@suse.de>
9376 * cfganal.c (control_dependences::~control_dependences):
9377 Properly free all of the vector.
9379 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9382 * config/i386/i386.c (ix86_conditional_register_usage):
9383 Proper initialize extended SSE registers.
9385 2013-09-06 Jan Hubicka <jh@suse.cz>
9387 PR tree-optimization/58311
9388 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9390 2013-09-06 Jan Hubicka <jh@suse.cz>
9392 * Makefile.in (tree-sra.o): Update dependencies.
9393 * tree-sra.c: Include ipa-utils.h
9394 (scan_function): Use recursive_call_p.
9395 (has_caller_p): New function.
9396 (cgraph_for_node_and_aliases): Count also callers of aliases.
9398 2013-09-06 Jan Hubicka <jh@suse.cz>
9401 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9402 * tree-tailcall.c: Include ipa-utils.h.
9403 (find_tail_calls): Use it.
9404 * ipa-pure-const.c (check_call): Likewise.
9405 * ipa-utils.c (recursive_call_p): New function.
9406 * ipa-utils.h (recursive_call_p): Dclare.
9407 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9408 (symtab_semantically_equivalent_p): New function.
9409 * Makefile.in (tree-tailcall.o): Update dependencies.
9411 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9413 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9416 2013-09-06 Richard Biener <rguenther@suse.de>
9418 * lto-streamer.h (lto_global_var_decls): Remove.
9419 * Makefile.in (OBJS): Remove lto-symtab.o.
9420 (lto-symtab.o): Remove.
9421 (GTFILES): Remove lto-symtab.c
9422 * lto-symtab.c: Move to lto/
9424 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9426 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9427 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9428 (UNSPEC_FPINT_RINT): New constant definitions.
9429 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9430 definition with 2 attributes.
9431 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9432 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9435 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9437 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9438 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9439 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9440 attribute to "z196_alone".
9441 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9444 2013-09-06 Richard Biener <rguenther@suse.de>
9446 * basic-block.h (class control_dependences): New.
9447 * tree-ssa-dce.c (control_dependence_map): Remove.
9449 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9450 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9451 find_pdom, find_control_dependence, find_all_control_dependences):
9453 (mark_control_dependent_edges_necessary,
9454 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9455 tree_dce_done, perform_tree_ssa_dce): Adjust.
9456 * cfganal.c (set_control_dependence_map_bit,
9457 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9458 find_all_control_dependences): Move from tree-ssa-dce.c and
9459 implement as methods of control_dependences class.
9460 (control_dependences::control_dependences): New.
9461 (control_dependences::~control_dependences): Likewise.
9462 (control_dependences::get_edges_dependent_on): Likewise.
9463 (control_dependences::get_edge): Likewise.
9465 2013-09-04 Jan Hubicka <jh@suse.cz>
9467 * tree.c (types_same_for_odr): Drop overactive check.
9468 * ipa-devirt.c (hash_type_name): Likewise.
9470 2013-09-04 Jan Hubicka <jh@suse.cz>
9472 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9473 (analyze_functions): ... here.
9475 2013-09-04 Jan Hubicka <jh@suse.cz>
9478 * cgraphunit.c (analyze_functions): Clear AUX fields
9479 after processing; initialize assembler name has.
9481 2013-09-05 Jeff Law <law@redhat.com>
9483 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9484 from thread_around_empty_block. Record threading path into PATH.
9485 Recurse if threading through the initial block is successful.
9486 (thread_across_edge): Corresponding changes to slightly simplify.
9488 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9490 * config/aarch64/aarch64.md
9491 (type): Remove frecpe, frecps, frecpx.
9492 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9493 fix to be a TARGET_SIMD instruction.
9494 (aarch64_frecps): Remove.
9495 * config/aarch64/aarch64-simd.md
9496 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9497 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9499 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9500 Sofiane Naci <sofiane.naci@arm.com>
9502 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9503 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9504 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9505 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9506 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9507 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9508 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9509 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9510 into "clz, "rbit". Rename "shift" to "shift_imm".
9511 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9512 changes. Update for attribute changes all occurrences of arlo_* and
9514 * config/arm/arm-fixed.md: Update for attribute changes
9515 all occurrences of arlo_* types.
9516 * config/arm/thumb2.md: Update for attribute changes all occurrences
9518 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9519 (cortexa7_older_only): Likewise.
9520 (cortexa7_younger): Likewise.
9521 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9522 (1020alu_shift_op): Likewise.
9523 (1020alu_shift_reg_op): Likewise.
9524 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9525 (alu_shift_op): Likewise.
9526 (alu_shift_reg_op): Likewise.
9527 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9528 (11_alu_shift_op): Likewise.
9529 (11_alu_shift_reg_op): Likewise.
9530 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9531 (9_alu_shift_reg_op): Likewise.
9532 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9534 (cortex_a15_alu_shift): Likewise.
9535 (cortex_a15_alu_shift_reg): Likewise.
9536 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9538 (cortex_a5_alu_shift): Likewise.
9539 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9541 (cortex_a53_alu_shift): Likewise.
9542 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9544 (cortex_a7_alu_reg): Likewise.
9545 (cortex_a7_alu_shift): Likewise.
9546 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9548 (cortex_a8_alu_shift): Likewise.
9549 (cortex_a8_alu_shift_reg): Likewise.
9550 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9551 (cortex_a9_dp_shift): Likewise.
9552 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9554 * config/arm/cortex-r4.md
9555 (cortex_r4_alu): Update for attribute changes.
9556 (cortex_r4_mov): Likewise.
9557 (cortex_r4_alu_shift_reg): Likewise.
9558 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9559 (526_alu_shift_op): Likewise.
9560 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9561 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9562 (626te_alu_shift_op): Likewise.
9563 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9564 (726te_alu_shift_op): Likewise.
9565 (726te_alu_shift_reg_op): Likewise.
9566 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9567 (mp626_alu_shift_op): Likewise.
9568 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9569 (pj4_alu_conds): Likewise.
9570 (pj4_shift): Likewise.
9571 (pj4_shift_conds): Likewise.
9572 (pj4_alu_shift): Likewise.
9573 (pj4_alu_shift_conds): Likewise.
9574 * config/aarch64/aarch64.md: Update for attribute change
9575 all occurrences of arlo_* and shift* types.
9577 2013-09-05 Mike Stump <mikestump@comcast.net>
9579 * tree.h: Move documentation for tree_function_decl to tree-core.h
9580 with the declaration.
9582 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9585 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9586 looking for widest mode.
9588 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9590 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9592 2013-09-05 Richard Biener <rguenther@suse.de>
9594 PR tree-optimization/58137
9595 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9596 Do not create vectors of pointers.
9597 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9598 types for the components of the vector initializer.
9599 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9600 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9602 2013-09-05 Martin Jambor <mjambor@suse.cz>
9604 * ipa-prop.c (remove_described_reference): Accept missing references,
9605 return false if that hppens, otherwise return true.
9606 (cgraph_node_for_jfunc): New function.
9607 (try_decrement_rdesc_refcount): Likewise.
9608 (try_make_edge_direct_simple_call): Use them.
9609 (ipa_edge_removal_hook): Remove references from rdescs.
9610 (ipa_edge_duplication_hook): Clone rdescs and their references
9611 when the new edge has the same caller as the old one.
9612 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9613 reference before removing any edges.
9615 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9617 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9619 * thumb2.md (thumb2_storewb_parisi): New pattern.
9621 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9623 * config/aarch64/aarch64-option-extensions.def: Add
9624 AARCH64_OPT_EXTENSION of 'crc'.
9625 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9626 (AARCH64_ISA_CRC): Ditto.
9627 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9628 description of the CRC extension.
9630 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
9632 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
9633 * config/rs6000/linux.h: Ditto.
9634 * alpha/linux.h: Ditto.
9635 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
9636 no_c99_libc_has_function.
9637 * config/c6x/uclinux-elf.h: Ditto.
9638 * config/lm32/uclinux-elf.h: Ditto.
9639 * config/m68k/uclinux.h: Ditto.
9640 * config/moxie/uclinux.h: Ditto.
9641 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
9642 (crisv32-*-linux*, cris-*-linux*): Ditto.
9643 * config/bfin/bfin.c: Include "tm_p.h".
9645 2013-09-05 Richard Biener <rguenther@suse.de>
9647 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
9648 check for a definition without a basic-block.
9650 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9651 Sofiane Naci <sofiane.naci@arm.com>
9653 * config/aarch64/aarch64.md
9654 (*movti_aarch64): Rename r_2_f and f_2_r.
9655 (*movsf_aarch64): Likewise.
9656 (*movdf_aarch64): Likewise.
9657 (*movtf_aarch64): Likewise.
9658 (aarch64_movdi_<mode>low): Likewise.
9659 (aarch64_movdi_<mode>high): Likewise.
9660 (aarch64_mov<mode>high_di): Likewise.
9661 (aarch64_mov<mode>low_di): Likewise.
9662 (aarch64_movtilow_tilow): Likewise.
9663 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
9664 values to config/arm/types.md
9665 (attribute "conds"): Update for attribute change.
9666 (anddi3_insn): Likewise.
9667 (iordi3_insn): Likewise.
9668 (xordi3_insn): Likewise.
9669 (one_cmpldi2): Likewise.
9670 * config/arm/types.md (type): Add Neon types.
9671 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
9672 use "type" attribute.
9673 (movmisalign<mode>_neon_store): Likewise.
9674 (movmisalign<mode>_neon_load): Likewise.
9675 (vec_set<mode>_internal): Likewise.
9676 (vec_setv2di_internal): Likewise.
9677 (vec_extract<mode>): Likewise.
9678 (vec_extractv2di): Likewise.
9679 (add<mode>3_neon): Likewise.
9680 (adddi3_neon): Likewise.
9681 (sub<mode>3_neon): Likewise.
9682 (subdi3_neon): Likewise.
9683 (mul<mode>3_neon): Likewise.
9684 (mul<mode>3add<mode>_neon): Likewise.
9685 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
9686 (fma<VCVTF:mode>4)): Likewise.
9687 (fma<VCVTF:mode>4_intrinsic): Likewise.
9688 (fmsub<VCVTF:mode>4)): Likewise.
9689 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
9690 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
9691 (ior<mode>3): Likewise.
9692 (and<mode>3): Likewise.
9693 (anddi3_neon): Likewise.
9694 (orn<mode>3_neon): Likewise.
9695 (orndi3_neon): Likewise.
9696 (bic<mode>3_neon): Likewise.
9697 (bicdi3_neon): Likewise.
9698 (xor<mode>3): Likewise.
9699 (one_cmpl<mode>2): Likewise.
9700 (abs<mode>2): Likewise.
9701 (neg<mode>2): Likewise.
9702 (umin<mode>3_neon): Likewise.
9703 (umax<mode>3_neon): Likewise.
9704 (smin<mode>3_neon): Likewise.
9705 (smax<mode>3_neon): Likewise.
9706 (vashl<mode>3): Likewise.
9707 (vashr<mode>3_imm): Likewise.
9708 (vlshr<mode>3_imm): Likewise.
9709 (ashl<mode>3_signed): Likewise.
9710 (ashl<mode>3_unsigned): Likewise.
9711 (neon_load_count): Likewise.
9712 (ashldi3_neon_noclobber): Likewise.
9713 (signed_shift_di3_neon): Likewise.
9714 (unsigned_shift_di3_neon): Likewise.
9715 (ashrdi3_neon_imm_noclobber): Likewise.
9716 (lshrdi3_neon_imm_noclobber): Likewise.
9717 (widen_ssum<mode>3): Likewise.
9718 (widen_usum<mode>3): Likewise.
9719 (quad_halves_<code>v4si): Likewise.
9720 (quad_halves_<code>v4sf): Likewise.
9721 (quad_halves_<code>v8hi): Likewise.
9722 (quad_halves_<code>v16qi): Likewise.
9723 (reduc_splus_v2di): Likewise.
9724 (neon_vpadd_internal<mode>): Likewise.
9725 (neon_vpsmin<mode>): Likewise.
9726 (neon_vpsmax<mode>): Likewise.
9727 (neon_vpumin<mode>): Likewise.
9728 (neon_vpumax<mode>): Likewise.
9729 (ss_add<mode>_neon): Likewise.
9730 (us_add<mode>_neon): Likewise.
9731 (ss_sub<mode>_neon): Likewise.
9732 (us_sub<mode>_neon): Likewise.
9733 (neon_vadd<mode>_unspec): Likewise.
9734 (neon_vaddl<mode>): Likewise.
9735 (neon_vaddw<mode>): Likewise.
9736 (neon_vhadd<mode>): Likewise.
9737 (neon_vqadd<mode>): Likewise.
9738 (neon_vaddhn<mode>): Likewise.
9739 (neon_vmul<mode>): Likewise.
9740 (neon_vmla<mode>): Likewise.
9741 (neon_vmlal<mode>): Likewise.
9742 (neon_vmls<mode>): Likewise.
9743 (neon_vmlsl<mode>): Likewise.
9744 (neon_vqdmulh<mode>): Likewise.
9745 (neon_vqdmlal<mode>): Likewise.
9746 (neon_vqdmlsl<mode>): Likewise.
9747 (neon_vmull<mode>): Likewise.
9748 (neon_vqdmull<mode>): Likewise.
9749 (neon_vsub<mode>_unspec): Likewise.
9750 (neon_vsubl<mode>): Likewise.
9751 (neon_vsubw<mode>): Likewise.
9752 (neon_vqsub<mode>): Likewise.
9753 (neon_vhsub<mode>): Likewise.
9754 (neon_vsubhn<mode>): Likewise.
9755 (neon_vceq<mode>): Likewise.
9756 (neon_vcge<mode>): Likewise.
9757 (neon_vcgeu<mode>): Likewise.
9758 (neon_vcgt<mode>): Likewise.
9759 (neon_vcgtu<mode>): Likewise.
9760 (neon_vcle<mode>): Likewise.
9761 (neon_vclt<mode>): Likewise.
9762 (neon_vcage<mode>): Likewise.
9763 (neon_vcagt<mode>): Likewise.
9764 (neon_vtst<mode>): Likewise.
9765 (neon_vabd<mode>): Likewise.
9766 (neon_vabdl<mode>): Likewise.
9767 (neon_vaba<mode>): Likewise.
9768 (neon_vabal<mode>): Likewise.
9769 (neon_vmax<mode>): Likewise.
9770 (neon_vmin<mode>): Likewise.
9771 (neon_vpaddl<mode>): Likewise.
9772 (neon_vpadal<mode>): Likewise.
9773 (neon_vpmax<mode>): Likewise.
9774 (neon_vpmin<mode>): Likewise.
9775 (neon_vrecps<mode>): Likewise.
9776 (neon_vrsqrts<mode>): Likewise.
9777 (neon_vqabs<mode>): Likewise.
9778 (neon_vqneg<mode>): Likewise.
9779 (neon_vcls<mode>): Likewise.
9780 (clz<mode>2): Likewise.
9781 (popcount<mode>2): Likewise.
9782 (neon_vrecpe): Likewise.
9783 (neon_vrsqrte): Likewise.
9784 (neon_vget_lane<mode>_sext_internal): Likewise.
9785 (neon_vget_lane<mode>_zext_internal): Likewise.
9786 (neon_vdup_n<mode>): Likewise.
9787 (neon_vdup_nv2di): Likewise.
9788 (neon_vdpu_lane<mode>_internal): Likewise.
9789 (neon_vswp<mode>): Likewise.
9790 (float<mode><V_cvtto>2): Likewise.
9791 (floatuns<mode><V_cvtto>2): Likewise.
9792 (fix_trunc<mode><V_cvtto>)2): Likewise
9793 (fixuns_trunc<mode><V_cvtto)2): Likewise.
9794 (neon_vcvt<mode>): Likewise.
9795 (neon_vcvtv4sfv4hf): Likewise.
9796 (neon_vcvtv4hfv4sf): Likewise.
9797 (neon_vcvt_n<mode>): Likewise.
9798 (neon_vmovn<mode>): Likewise.
9799 (neon_vqmovn<mode>): Likewise.
9800 (neon_vqmovun<mode>): Likewise.
9801 (neon_vmovl<mode>): Likewise.
9802 (neon_vmul_lane<mode>): Likewise.
9803 (neon_vmull_lane<mode>): Likewise.
9804 (neon_vqdmull_lane<mode>): Likewise.
9805 (neon_vqdmulh_lane<mode>): Likewise.
9806 (neon_vmla_lane<mode>): Likewise.
9807 (neon_vmlal_lane<mode>): Likewise.
9808 (neon_vqdmlal_lane<mode>): Likewise.
9809 (neon_vmls_lane<mode>): Likewise.
9810 (neon_vmlsl_lane<mode>): Likewise.
9811 (neon_vqdmlsl_lane<mode>): Likewise.
9812 (neon_vext<mode>): Likewise.
9813 (neon_vrev64<mode>): Likewise.
9814 (neon_vrev32<mode>): Likewise.
9815 (neon_vrev16<mode>): Likewise.
9816 (neon_vbsl<mode>_internal): Likewise.
9817 (neon_vshl<mode>): Likewise.
9818 (neon_vqshl<mode>): Likewise.
9819 (neon_vshr_n<mode>): Likewise.
9820 (neon_vshrn_n<mode>): Likewise.
9821 (neon_vqshrn_n<mode>): Likewise.
9822 (neon_vqshrun_n<mode>): Likewise.
9823 (neon_vshl_n<mode>): Likewise.
9824 (neon_vqshl_n<mode>): Likewise.
9825 (neon_vqshlu_n<mode>): Likewise.
9826 (neon_vshll_n<mode>): Likewise.
9827 (neon_vsra_n<mode>): Likewise.
9828 (neon_vsri_n<mode>): Likewise.
9829 (neon_vsli_n<mode>): Likewise.
9830 (neon_vtbl1v8qi): Likewise.
9831 (neon_vtbl2v8qi): Likewise.
9832 (neon_vtbl3v8qi): Likewise.
9833 (neon_vtbl4v8qi): Likewise.
9834 (neon_vtbx1v8qi): Likewise.
9835 (neon_vtbx2v8qi): Likewise.
9836 (neon_vtbx3v8qi): Likewise.
9837 (neon_vtbx4v8qi): Likewise.
9838 (neon_vtrn<mode>_internal): Likewise.
9839 (neon_vzip<mode>_internal): Likewise.
9840 (neon_vuzp<mode>_internal): Likewise.
9841 (neon_vld1<mode>): Likewise.
9842 (neon_vld1_lane<mode>): Likewise.
9843 (neon_vld1_dup<mode>): Likewise.
9844 (neon_vld1_dupv2di): Likewise.
9845 (neon_vst1<mode>): Likewise.
9846 (neon_vst1_lane<mode>): Likewise.
9847 (neon_vld2<mode>): Likewise.
9848 (neon_vld2_lane<mode>): Likewise.
9849 (neon_vld2_dup<mode>): Likewise.
9850 (neon_vst2<mode>): Likewise.
9851 (neon_vst2_lane<mode>): Likewise.
9852 (neon_vld3<mode>): Likewise.
9853 (neon_vld3qa<mode>): Likewise.
9854 (neon_vld3qb<mode>): Likewise.
9855 (neon_vld3_lane<mode>): Likewise.
9856 (neon_vld3_dup<mode>): Likewise.
9857 (neon_vst3<mode>): Likewise.
9858 (neon_vst3qa<mode>): Likewise.
9859 (neon_vst3qb<mode>): Likewise.
9860 (neon_vst3_lane<mode>): Likewise.
9861 (neon_vld4<mode>): Likewise.
9862 (neon_vld4qa<mode>): Likewise.
9863 (neon_vld4qb<mode>): Likewise.
9864 (neon_vld4_lane<mode>): Likewise.
9865 (neon_vld4_dup<mode>): Likewise.
9866 (neon_vst4<mode>): Likewise.
9867 (neon_vst4qa<mode>): Likewise.
9868 (neon_vst4qb<mode>): Likewise.
9869 (neon_vst4_lane<mode>): Likewise.
9870 (neon_vec_unpack<US>_lo_<mode>): Likewise.
9871 (neon_vec_unpack<US>_hi_<mode>): Likewise.
9872 (neon_vec_<US>mult_lo_<mode>): Likewise.
9873 (neon_vec_<US>mult_hi_<mode>): Likewise.
9874 (neon_vec_<US>shiftl_<mode>): Likewise.
9875 (neon_unpack<US>_<mode>): Likewise.
9876 (neon_vec_<US>mult_<mode>): Likewise.
9877 (vec_pack_trunc_<mode>): Likewise.
9878 (neon_vec_pack_trunk_<mode>): Likewise.
9879 (neon_vabd<mode>_2): Likewise.
9880 (neon_vabd<mode>_3): Likewise.
9881 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9882 (thumb2_movsi_vfp): Likewise.
9883 (movdi_vfp): Likewise.
9884 (movdi_vfp_cortexa8): Likewise.
9885 (movhf_vfp_neon): Likewise.
9886 (movhf_vfp): Likewiwse.
9887 (movsf_vfp): Likewiwse.
9888 (thumb2_movsf_vfp): Likewiwse.
9889 (movdf_vfp): Likewise.
9890 (thumb2_movdf_vfp): Likewise.
9891 (movsfcc_vfp): Likewise.
9892 (thumb2_movsfcc_vfp): Likewise.
9893 (movdfcc_vfp): Likewise.
9894 (thumb2_movdfcc_vfp): Likewise.
9895 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
9896 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
9897 (v10_v2c): Likewise.
9898 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
9900 (cortex_a15_neon_int_2): Likewise.
9901 (cortex_a15_neon_int_3): Likewise.
9902 (cortex_a15_neon_int_4): Likewise.
9903 (cortex_a15_neon_int_5): Likewise.
9904 (cortex_a15_neon_vqneg_vqabs): Likewise.
9905 (cortex_a15_neon_vmov): Likewise.
9906 (cortex_a15_neon_vaba): Likewise.
9907 (cortex_a15_neon_vaba_qqq): Likewise.
9908 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9909 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9910 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
9911 scalar_64_32_long_scalar): Likewise.
9912 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9913 (cortex_a15_neon_mla_qqq_8_16): Likewise.
9914 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9915 lotype_qdd_64_32_long): Likewise.
9916 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9917 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9918 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
9919 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9920 (cortex_a15_neon_shift_1): Likewise.
9921 (cortex_a15_neon_shift_2): Likewise.
9922 (cortex_a15_neon_shift_3): Likewise.
9923 (cortex_a15_neon_vshl_ddd): Likewise.
9924 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9925 (cortex_a15_neon_vsra_vrsra): Likewise.
9926 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
9927 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
9928 (cortex_a15_neon_fp_vmul_ddd): Likewise.
9929 (cortex_a15_neon_fp_vmul_qqd): Likewise.
9930 (cortex_a15_neon_fp_vmla_ddd): Likewise.
9931 (cortex_a15_neon_fp_vmla_qqq): Likewise.
9932 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
9933 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
9934 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9935 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9936 (cortex_a15_neon_bp_simple): Likewise.
9937 (cortex_a15_neon_bp_2cycle): Likewise.
9938 (cortex_a15_neon_bp_3cycle): Likewise.
9939 (cortex_a15_neon_vld1_1_2_regs): Likewise.
9940 (cortex_a15_neon_vld1_3_4_regs): Likewise.
9941 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9942 (cortex_a15_neon_vld2_4_regs): Likewise.
9943 (cortex_a15_neon_vld3_vld4): Likewise.
9944 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9945 (cortex_a15_neon_vst1_3_4_regs): Likewise.
9946 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
9947 (cortex_a15_neon_vst3_vst4): Likewise.
9948 (cortex_a15_neon_vld1_vld2_lane): Likewise.
9949 (cortex_a15_neon_vld3_vld4_lane" 10
9950 (cortex_a15_neon_vst1_vst2_lane): Likewise.
9951 (cortex_a15_neon_vst3_vst4_lane): Likewise.
9952 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
9953 (cortex_a15_neon_ldm_2): Likewise.0
9954 (cortex_a15_neon_stm_2): Likewise.
9955 (cortex_a15_neon_mcr): Likewise.
9956 (cortex_a15_neon_mcr_2_mcrr): Likewise.
9957 (cortex_a15_neon_mrc): Likewise.
9958 (cortex_a15_neon_mrrc): Likewise.
9959 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9961 (cortex_a15_alu_shift): Likewise.
9962 (cortex_a15_alu_shift_reg): Likewise.
9963 (cortex_a15_mult32): Likewise.
9964 (cortex_a15_mult64): Likewise.
9965 (cortex_a15_block): Likewise.
9966 (cortex_a15_branch): Likewise.
9967 (cortex_a15_load1): Likewise.
9968 (cortex_a15_load3): Likewise.
9969 (cortex_a15_store1): Likewise.
9970 (cortex_a15_store3): Likewise.
9971 (cortex_a15_call): Likewise.
9972 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
9973 (cortex_a5_f2r): Likewise.
9974 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
9976 (cortex_a53_f2r): Likewise.
9977 * config/arm/cortex-a7.md
9978 (cortex_a7_branch): Update for attribute change.
9979 (cortex_a7_call): Likewise.
9980 (cortex_a7_alu_imm): Likewise.
9981 (cortex_a7_alu_reg): Likewise.
9982 (cortex_a7_alu_shift): Likewise.
9983 (cortex_a7_mul): Likewise.
9984 (cortex_a7_load1): Likewise.
9985 (cortex_a7_store1): Likewise.
9986 (cortex_a7_load2): Likewise.
9987 (cortex_a7_store2): Likewise.
9988 (cortex_a7_load3): Likewise.
9989 (cortex_a7_store3): Likewise.
9990 (cortex_a7_load4): Likewise.
9991 (cortex_a7_store4): Likewise.
9992 (cortex_a7_fpalu): Likewise.
9993 (cortex_a7_fconst): Likewise.
9994 (cortex_a7_fpmuls): Likewise.
9995 (cortex_a7_neon_mul): Likewise.
9996 (cortex_a7_fpmacs): Likewise.
9997 (cortex_a7_neon_mla: Likewise.
9998 (cortex_a7_fpmuld: Likewise.
9999 (cortex_a7_fpmacd: Likewise.
10000 (cortex_a7_fpfmad: Likewise.
10001 (cortex_a7_fdivs: Likewise.
10002 (cortex_a7_fdivd: Likewise.
10003 (cortex_a7_r2f: Likewise.
10004 (cortex_a7_f2r: Likewise.
10005 (cortex_a7_f_flags: Likewise.
10006 (cortex_a7_f_loads: Likewise.
10007 (cortex_a7_f_loadd: Likewise.
10008 (cortex_a7_f_stores: Likewise.
10009 (cortex_a7_f_stored: Likewise.
10010 (cortex_a7_neon): Likewise.
10011 * config/arm/cortex-a8-neon.md
10012 (cortex_a8_neon_mrc): Update for attribute change.
10013 (cortex_a8_neon_mrrc): Likewise.
10014 (cortex_a8_neon_int_1): Likewise.
10015 (cortex_a8_neon_int_2): Likewise.
10016 (cortex_a8_neon_int_3): Likewise.
10017 (cortex_a8_neon_int_4): Likewise.
10018 (cortex_a8_neon_int_5): Likewise.
10019 (cortex_a8_neon_vqneg_vqabs): Likewise.
10020 (cortex_a8_neon_vmov): Likewise.
10021 (cortex_a8_neon_vaba): Likewise.
10022 (cortex_a8_neon_vaba_qqq): Likewise.
10023 (cortex_a8_neon_vsma): Likewise.
10024 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10025 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10026 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10027 long_scalar): Likewise.
10028 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10029 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10030 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10031 long_scalar_qdd_64_32_long): Likewise.
10032 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10033 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10034 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10035 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10036 (cortex_a8_neon_shift_1): Likewise.
10037 (cortex_a8_neon_shift_2): Likewise.
10038 (cortex_a8_neon_shift_3): Likewise.
10039 (cortex_a8_neon_vshl_ddd): Likewise.
10040 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10041 (cortex_a8_neon_vsra_vrsra): Likewise.
10042 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10043 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10044 (cortex_a8_neon_fp_vsum): Likewise.
10045 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10046 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10047 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10048 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10049 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10050 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10051 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10052 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10053 (cortex_a8_neon_bp_simple): Likewise.
10054 (cortex_a8_neon_bp_2cycle): Likewise.
10055 (cortex_a8_neon_bp_3cycle): Likewise.
10056 (cortex_a8_neon_ldr): Likewise.
10057 (cortex_a8_neon_str): Likewise.
10058 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10059 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10060 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10061 (cortex_a8_neon_vld2_4_regs): Likewise.
10062 (cortex_a8_neon_vld3_vld4): Likewise.
10063 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10064 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10065 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10066 (cortex_a8_neon_vst3_vst4): Likewise.
10067 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10068 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10069 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10070 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10071 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10072 (cortex_a8_neon_mcr): Likewise.
10073 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10074 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10075 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10077 (ca9_neon_mrrc): Likewise.
10078 (cortex_a9_neon_int_1): Likewise.
10079 (cortex_a9_neon_int_2): Likewise.
10080 (cortex_a9_neon_int_3): Likewise.
10081 (cortex_a9_neon_int_4): Likewise.
10082 (cortex_a9_neon_int_5): Likewise.
10083 (cortex_a9_neon_vqneg_vqabs): Likewise.
10084 (cortex_a9_neon_vmov): Likewise.
10085 (cortex_a9_neon_vaba): Likewise.
10086 (cortex_a9_neon_vaba_qqq): Likewise.
10087 (cortex_a9_neon_vsma): Likewise.
10088 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10089 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10090 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10091 long_scalar): Likewise.
10092 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10093 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10094 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10095 long_scalar_qdd_64_32_long): Likewise.
10096 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10097 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10098 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10099 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10100 (cortex_a9_neon_shift_1): Likewise.
10101 (cortex_a9_neon_shift_2): Likewise.
10102 (cortex_a9_neon_shift_3): Likewise.
10103 (cortex_a9_neon_vshl_ddd): Likewise.
10104 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10105 (cortex_a9_neon_vsra_vrsra): Likewise.
10106 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10107 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10108 (cortex_a9_neon_fp_vsum): Likewise.
10109 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10110 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10111 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10112 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10113 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10114 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10115 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10116 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10117 (cortex_a9_neon_bp_simple): Likewise.
10118 (cortex_a9_neon_bp_2cycle): Likewise.
10119 (cortex_a9_neon_bp_3cycle): Likewise.
10120 (cortex_a9_neon_ldr): Likewise.
10121 (cortex_a9_neon_str): Likewise.
10122 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10123 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10124 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10125 (cortex_a9_neon_vld2_4_regs): Likewise.
10126 (cortex_a9_neon_vld3_vld4): Likewise.
10127 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10128 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10129 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10130 (cortex_a9_neon_vst3_vst4): Likewise.
10131 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10132 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10133 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10134 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10135 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10136 (cortex_a9_neon_mcr): Likewise.
10137 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10138 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10139 (cortex_a9_fps): Likewise.
10140 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10142 (cortex_m4_fmuls): Likewise.
10143 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10145 (cortex_r4_mrc): Likewise.
10146 * config/arm/iterators.md: Update comment referring to neon_type.
10147 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10148 (iwmmxt_movsi_insn): Likewise.
10149 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10151 (pj4_core_to_vfp): Likewise.
10152 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10154 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10155 (vfp_fstore): Likewise.
10156 * doc/md.texi: Change references to neon_type to refer to type.
10158 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10160 * tree.h (DECL_BUILT_IN): Fix typo in comment.
10162 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
10164 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
10165 lglobl if not weak.
10167 2013-09-04 Easwaran Raman <eraman@google.com>
10169 PR middle-end/57370
10170 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
10171 (build_and_add_sum): Use it.
10172 (appears_later_in_bb): Simplify code.
10174 2013-09-04 Teresa Johnson <tejohnson@google.com>
10176 * dumpfile.c (dump_finish): Don't close stderr/stdout.
10178 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
10180 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
10182 2013-09-04 Jan Hubicka <jh@suse.cz>
10184 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
10185 * ipa-devirt.c: Include diganostic.h
10186 (odr_type_d): Add types and types_set.
10187 (hash_type_name): Work for types with vtables during LTO.
10188 (odr_hasher::remove): Fix comment; destroy types_set.
10189 (add_type_duplicate): New function,
10190 (get_odr_type): Use it.
10191 (dump_type_inheritance_graph): Dump type duplicates.
10192 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
10194 * tree.c (types_same_for_odr): Give exact answers on types with
10197 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10199 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
10200 explaining their differences.
10202 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
10204 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
10206 2013-09-03 Jeff Law <law@redhat.com>
10208 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
10209 when not threading through a joiner block. Pass joiner/no joiner
10210 state to register_jump_thread.
10211 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10212 state from argument rather than implying on path length.
10213 Dump the entire jump thread path into debugging dump.
10214 * tree-flow.h (register_jump_thread): Update prototype.
10216 2013-08-29 Xinliang David Li <davidxl@google.com>
10218 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10219 Remove a trivial gcc_assert.
10221 2013-08-29 Xinliang David Li <davidxl@google.com>
10223 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10224 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10225 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10226 Delay base decl alignment adjustment.
10227 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10228 * tree-vectorizer.h: New data structure.
10229 (set_dr_misalignment): New function.
10230 (dr_misalignment): Ditto.
10231 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10232 (vectorizable_load): Ditto.
10233 (ensure_base_align): New function.
10234 (vectorize_loops): Add dbg_cnt support.
10235 (execute_vect_slp): Ditto.
10236 * dbgcnt.def: New debug counter.
10237 * Makefile: New dependency.
10239 2013-09-03 Meador Inge <meadori@codesourcery.com>
10243 2013-08-30 Meador Inge <meadori@codesourcery.com>
10245 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10247 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10249 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10250 function descriptor.
10252 2013-09-03 Richard Biener <rguenther@suse.de>
10254 * tree-affine.c (add_elt_to_tree): Fix association issue,
10255 avoid useless converts and make sure to always return a
10256 properly typed result.
10258 2013-09-03 Richard Biener <rguenther@suse.de>
10260 PR middle-end/57656
10261 * fold-const.c (negate_expr_p): Fix division case.
10262 (negate_expr): Likewise.
10264 2013-09-03 Richard Biener <rguenther@suse.de>
10267 * tree-streamer-out.c: Include tm.h.
10268 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10270 2013-09-03 Jan Hubicka <jh@suse.cz>
10272 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10274 2013-09-03 Alan Modra <amodra@gmail.com>
10276 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10277 multilibs to big-endian target and vice versa.
10278 * config/rs6000/t-linux64: Use := assignment on all vars.
10279 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10280 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10281 * config/rs6000/t-linux64le: New file.
10282 * config/rs6000/t-linux64bele: New file.
10283 * config/rs6000/t-linux64lebe: New file.
10285 2013-09-02 Jan Hubicka <jh@suse.cz>
10287 * ipa-inline-transform.c (inline_transform): Do not
10288 optimize_inline_calls when not optimizing.
10290 2013-09-02 Jan Hubicka <jh@suse.cz>
10292 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10293 duplicated nodes for assembler names.
10294 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10295 hard registers from assembler name hash.
10297 2013-09-02 Jan Hubicka <jh@suse.cz>
10299 * ipa-split.c (execute_split_functions): Split externally visible
10300 functions called once.
10302 2013-09-02 Martin Jambor <mjambor@suse.cz>
10305 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10306 linked list. When finding the correct duplicate, also consider also
10307 the caller in additon to its inlined_to node.
10309 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10311 * config/aarch64/aarch64-simd-builtins.def
10312 (dup_lane_scalar): Remove.
10313 * config/aarch64/aarch64-simd.md
10314 (aarch64_simd_dup): Add 'w->w' alternative.
10315 (aarch64_dup_lane<mode>): Allow for VALL.
10316 (aarch64_dup_lane_scalar<mode>): Remove.
10317 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10318 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10319 (aarch64_get_lane_unsigned<mode>): Likewise.
10320 (aarch64_get_lane<mode>): Likewise.
10321 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10322 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10323 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10324 (VCON): Change container of V2SF.
10325 (vswap_width_name): Likewise.
10326 * config/aarch64/arm_neon.h
10327 (__aarch64_vdup_lane_any): New.
10328 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10329 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10330 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10332 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10334 PR middle-end/56382
10335 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10336 the source or the destination is a single hard register.
10338 2013-09-02 Richard Biener <rguenther@suse.de>
10340 PR middle-end/57511
10341 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10344 2013-09-02 Richard Biener <rguenther@suse.de>
10346 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10347 arithmetic to sizetype.
10349 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10351 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10352 Find auto-increment use both before and after candidate.
10354 2013-09-02 Marek Polacek <polacek@redhat.com>
10356 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10358 2013-09-01 Jan Hubicka <jh@suse.cz>
10360 * Makefile.in: Add ipa-profile.o
10361 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10362 * cgraph.c (struct cgraph_propagate_frequency_data,
10363 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10364 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10365 * cgraph.h (cgraph_propagate_frequency): Remove.
10366 * ipa-inline-analysis.c: Include ipa-utils.h;
10367 drop duplicated cfgloop.h.
10368 (inline_update_callee_summaries): Update.
10369 * ipa-profile.c: New file.
10370 * ipa-utils.h (ipa_propagate_frequency): Declare.
10371 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10372 data-streamer.h, value-prof.h.
10373 (symtab_remove_unreachable_nodes): Update profile.
10374 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10375 account_time_size, cmp_counts, dump_histogram,
10376 ipa_profile_generate_summary, ipa_profile_write_summary,
10377 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10378 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10379 Move to ipa-profile.c.
10381 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10383 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10385 2013-09-01 Jan Hubicka <jh@suse.cz>
10387 * common.opt (fdevirtualize-speculatively): New function.
10388 * invoke.texi (fdevirtualize-speculatively): Document.
10389 * ipa-devirt.c: Include ipa-inline.h
10390 (likely_target_p): New function.
10391 (ipa_devirt): New function.
10392 (gate_ipa_devirt): New function.
10393 (pass_data_ipa_devirt): New static var.
10394 (pass_ipa_devirt): Likewise.
10395 (make_pass_ipa_devirt): New function.
10396 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10397 (common_handle_option): Disable devirtualization when
10398 value range profiling is available.
10399 * passes.def (pass_ipa_devirt): Add.
10400 * timever.def (TV_IPA_DEVIRT): New timevar.
10401 * tree-pass.h (make_pass_ipa_devirt):
10403 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10405 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10406 include sanitize(undefined).
10408 2013-08-31 Diego Novillo <dnovillo@google.com>
10410 * Makefile.in (TREE_CORE_H): Define.
10412 (GTFILES): Add tree-core.h.
10413 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10415 * tree-core.h: New file.
10416 Move all data structures, enum, typedefs, global
10417 declarations and constants from ...
10418 * tree.h: ... here.
10420 2013-08-31 Jan Hubicka <jh@suse.cz>
10422 * bulitins.c (expand_builtin): Do not early exit for gcov
10423 instrumented functions.
10425 2013-08-31 Marek Polacek <polacek@redhat.com>
10427 * ubsan.c: Include tm_p.h.
10429 2013-08-31 Jan Hubicka <jh@suse.cz>
10431 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10434 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10435 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10436 * tree-cfg.c (verify_gimple_label): ... here.
10437 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10438 (ipa_merge_profiles): New function.
10439 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10440 (lto_input_function_body): Likewise.
10441 * ipa-utils.h (ipa_merge_profiles): Declare.
10442 * lto-streamer.h (lto_input_function_body): Update prototype.
10443 (emit_label_in_global_context_p): Remove.
10444 * lto-symtab.c: Include ipa-utils.h
10445 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10447 2013-08-31 Jan Hubicka <jh@suse.cz>
10449 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10451 2013-08-31 Jan Hubicka <jh@suse.cz>
10453 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10455 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10457 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10458 "cmp" RTX before signed_comparison_operator check to account
10459 for "code" changes.
10461 2013-08-30 Jan Hubicka <jh@suse.cz>
10463 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10464 (detect_type_change_1): Rename to ...
10465 (detect_type_change): ... this one; early return on non-polymorphic
10467 (detect_type_change_ssa): Add comp_type parameter; update
10468 use of detect_type_change.
10469 (compute_complex_assign_jump_func): Add param_type parameter;
10470 update use of detect_type_change_ssa.
10471 (compute_complex_ancestor_jump_func): Likewise.
10472 (ipa_get_callee_param_type): New function.
10473 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10474 update calls to the jump function computation functions.
10476 2013-08-30 Teresa Johnson <tejohnson@google.com>
10477 Steven Bosscher <steven@gcc.gnu.org>
10479 * cfgrtl.c (fixup_new_cold_bb): New routine.
10480 (commit_edge_insertions): Invoke fixup_partitions.
10481 (find_partition_fixes): New routine.
10482 (fixup_partitions): Ditto.
10483 (verify_hot_cold_block_grouping): Update comments.
10484 (rtl_verify_edges): Invoke find_partition_fixes.
10485 (rtl_verify_bb_pointers): Update comments.
10486 (rtl_verify_bb_layout): Ditto.
10487 * basic-block.h (probably_never_executed_edge_p): Declare.
10488 (fixup_partitions): Ditto.
10489 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10490 * bb-reorder.c (sanitize_hot_paths): New function.
10491 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10492 sanitize_hot_paths.
10493 * predict.c (probably_never_executed_edge_p): New routine.
10494 * cfg.c (check_bb_profile): Add partition insanity warnings.
10496 2013-08-30 Meador Inge <meadori@codesourcery.com>
10498 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10500 2013-08-30 Marek Polacek <polacek@redhat.com>
10502 * Makefile.in (ubsan.o): Add.
10503 (c-family/c-ubsan.o): Add.
10504 (builtins.o): Add ubsan.h dependency.
10505 * ubsan.h: New file.
10506 * ubsan.c: New file.
10507 * common.opt: Add -fsanitize=undefined option.
10508 (flag_sanitize): Add variable.
10509 (fsanitize=): Add option. Add Driver.
10510 (fsanitize=thread): Remove option.
10511 (fsanitize=address): Likewise.
10512 (static-libubsan): New option.
10513 * doc/invoke.texi: Document the new flag and -static-libubsan.
10514 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10515 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10516 * builtin-attrs.def (ATTR_COLD): Define.
10517 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10518 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10519 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10520 * flag-types.h (sanitize_code): New enum.
10521 * opts.c (common_handle_option): Parse command line arguments
10522 of -fsanitize=. Add -fsanitize=unreachable option.
10523 * varasm.c (get_variable_section): Adjust.
10524 (assemble_noswitch_variable): Likewise.
10525 (assemble_variable): Likewise.
10526 (output_constant_def_contents): Likewise.
10527 (categorize_decl_for_section): Likewise.
10528 (place_block_symbol): Likewise.
10529 (output_object_block): Likewise.
10530 * builtins.def: Likewise.
10531 * toplev.c (compile_file): Likewise.
10532 (process_options): Likewise.
10533 * cppbuiltin.c: Likewise.
10534 * tsan.c (tsan_pass): Likewise.
10535 (tsan_gate): Likewise.
10536 (tsan_gate_O0): Likewise.
10537 * cfgexpand.c (partition_stack_vars): Likewise.
10538 (expand_stack_vars): Likewise.
10539 (defer_stack_allocation): Likewise.
10540 (expand_used_vars): Likewise.
10541 * cfgcleanup.c (old_insns_match_p): Likewise.
10542 * asan.c (asan_finish_file): Likewise.
10543 (asan_instrument): Likewise.
10544 (gate_asan): Likewise.
10545 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10546 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10547 (asan_global_struct): Use pointer_sized_int_node instead
10548 calling build_nonstandard_integer_type.
10549 (initialize_sanitizer_builtins): Likewise.
10550 (asan_finish_file): Likewise.
10551 * gcc.c: Document %{%:function(args):X}.
10552 (static_spec_functions): Add sanitize.
10553 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10554 store funcval != NULL there.
10555 (do_spec_1): Adjust handle_spec_function caller.
10556 (handle_braces): Allow %:function(args) as condition.
10557 (sanitize_spec_function): New function.
10558 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10559 (LIBUBSAN_SPEC): Likewise.
10560 (LIBUBSAN_EARLY_SPEC): Likewise.
10561 (SANITIZER_SPEC): Handle libubsan.
10562 (SANITIZER_EARLY_SPEC): Likewise.
10563 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10564 instead of fsanitize=address.
10565 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10566 instead of fsanitize=address*.
10567 * builtins.c: Include ubsan.h.
10568 (fold_builtin_0): Instrument __builtin_unreachable.
10569 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10570 instead of flag_asan.
10571 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10572 (pointer_sized_int_node): Define.
10573 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10575 2013-08-30 Mike Stump <mikestump@comcast.net>
10577 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10580 2013-08-29 Jan Hubicka <jh@suse.cz>
10582 * cgraph.c (cgraph_function_body_availability): Handle weakref
10584 * passes.def: Remove pass_fixup_cfg.
10585 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10586 track when we need to remove functions.
10587 (gate_ipa_inline): Execute inlining always; add comment why.
10588 (pass_data_ipa_inline): Remove TODO_remove_functions.
10589 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10590 do not produce summaries.
10591 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10592 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10593 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10594 constant pool and vtable.
10596 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10598 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10599 New arm_neon.h's internal macros to specify 64-bit constants.
10600 Avoid using stdint.h's macros.
10602 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10604 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10605 before applying REG_P.
10607 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10609 PR tree-optimization/58277
10610 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10611 after seeing too many stmts with vdef in between dombb and current
10612 bb, invalidate everything.
10614 2013-08-30 Richard Biener <rguenther@suse.de>
10616 * fold-const.c (fold_single_bit_test): Fix overflow test.
10618 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10620 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10621 and which can live in a register, always retrieve the value on entry.
10622 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10623 passed by invisible reference specially.
10624 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10625 (vt_add_function_parameter): Correctly deal with a parameter passed by
10626 invisible reference.
10628 2013-08-30 Jan Hubicka <jh@suse.cz>
10630 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
10632 2013-08-30 Richard Biener <rguenther@suse.de>
10634 PR tree-optimization/58228
10635 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
10636 allow invariant loads in nested loop vectorization.
10638 2013-08-30 Richard Biener <rguenther@suse.de>
10640 PR tree-optimization/58223
10641 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
10642 (has_anti_or_output_dependence): ... this and adjust to also
10643 look for output dependences.
10644 (mark_nodes_having_upstream_mem_writes): Adjust.
10645 (rdg_flag_uses): Likewise.
10647 2013-08-30 Richard Biener <rguenther@suse.de>
10649 PR tree-optimization/58010
10650 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
10651 assert that we have a loop-closed PHI.
10653 2013-08-29 Jan Hubicka <jh@suse.cz>
10655 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
10656 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
10657 * lto-section-in.c (lto_free_function_in_decl_state): New function.
10658 (lto_free_function_in_decl_state_for_node): New function.
10660 2013-08-29 Xinliang David Li <davidxl@google.com>
10662 * loop-unroll.c (report_unroll_peel): Minor message change.
10663 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
10664 Emit alignment peeling message with default -fopt-info.
10665 (vect_loop_versioning): Emit loop version info message.
10666 * tree-vectorizer.c (vectorize_loops): Minor message change.
10667 (execute_vect_slp): Ditto.
10669 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
10671 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
10672 of the clone from the DECL_NAME of the original function.
10674 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
10676 * passes.c (register_pass): Add overload.
10677 * tree-pass.h (register_pass): Forward declare it. Add comment.
10679 2013-08-29 Jan Hubicka <jh@suse.cz>
10681 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
10682 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
10683 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
10684 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10685 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
10686 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
10687 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10688 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
10690 2013-08-29 Teresa Johnson <tejohnson@google.com>
10692 * dumpfile.c (dump_loc): Output column number.
10693 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
10694 * doc/invoke.texi: Document optall -fopt-info flag.
10695 * profile.c (read_profile_edge_counts): Use new dump framework.
10696 (compute_branch_probabilities): Ditto.
10697 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
10698 when pass not in any opt group.
10699 * pass_manager.h (pass_manager::get_pass_profile): New method.
10700 * value-prof.c (check_counter): Use new dump framework.
10701 (check_ic_target): Ditto.
10702 * coverage.c (get_coverage_counts): Ditto.
10703 (coverage_init): Setup new dump framework.
10705 2013-08-29 Richard Biener <rguenther@suse.de>
10707 PR tree-optimization/58246
10708 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
10709 handle the dominance check inside a basic-block.
10711 2013-08-29 Richard Biener <rguenther@suse.de>
10713 PR middle-end/57287
10714 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
10715 of default defs that appear in abnormal PHI nodes.
10717 2013-08-29 Richard Biener <rguenther@suse.de>
10719 PR tree-optimization/57685
10720 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
10721 single-use operands to avoid exponential complexity.
10723 2013-08-28 Dehao Chen <dehao@google.com>
10725 * ipa-inline.c (edge_badness): Fix integer underflow.
10727 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
10729 * gtm-builtins.def (_ITM_free): Declare leaf.
10731 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10734 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
10735 (*tls_local_dynamic_base_64_largepic): Likewise.
10736 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
10737 Remove predicate from call operand.
10738 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
10739 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
10741 2013-08-28 Jeff Law <law@redhat.com>
10743 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
10744 checks for the number of predecessors and successors allowed.
10745 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
10746 which require copying a joiner block if there is a request which
10747 is a subpath that requires no joiner block copying.
10749 2013-08-28 Jan Hubicka <jh@suse.cz>
10751 * lto-streamer-out.c (DFS_write_tree_body): Drop
10752 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
10753 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
10754 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
10755 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
10756 Do not read DECL_ERROR_ISSUED.
10757 (unpack_ts_decl_with_vis_value_fields): Do not read
10759 (lto_input_ts_binfo_tree_pointers): Do not read
10760 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
10761 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
10762 write DECL_ERROR_ISSUED..
10763 (pack_ts_decl_with_vis_value_fields): Do not write
10765 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
10766 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
10767 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
10768 * tree.h (tree_decl_common): Update comment.
10769 (DECL_ERROR_ISSUED): Remove.
10771 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10773 PR middle-end/58257
10774 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
10776 2013-08-28 Jan Hubicka <jh@suse.cz>
10778 * builtins.def (free): Declare leaf.
10780 2013-08-27 David Malcolm <dmalcolm@redhat.com>
10782 * gdbhooks.py: New.
10783 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
10784 * configure: Regenerate.
10786 2013-08-27 Martin Jambor <mjambor@suse.cz>
10788 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
10789 (ipa_ancestor_jf_data): Likewise.
10790 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
10791 (ipa_get_jf_pass_through_type_preserved): New function.
10792 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
10793 (ipa_get_jf_ancestor_type_preserved): New function.
10794 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
10795 (ipa_get_jf_ancestor_result): Likewise.
10796 (propagate_vals_accross_pass_through): Use
10797 ipa_get_jf_pass_through_result to do all the value mappings.
10798 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
10799 type_preserved flag.
10800 (ipa_set_jf_cst_copy): New function.
10801 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
10802 (ipa_set_jf_arith_pass_through): Likewise.
10803 (ipa_set_ancestor_jf): Likewise.
10804 (compute_complex_assign_jump_func): Set type_preserved instead of
10806 (ipa_compute_jump_functions_for_edge): Likewise.
10807 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
10808 (update_jump_functions_after_inlining): Update type_preserved.
10809 Explicitely create jump functions when combining one with pass_through.
10810 (ipa_write_jump_function): Stream the type_preserved flags.
10811 (ipa_read_jump_function): Likewise.
10813 2013-08-27 Jakub Jelinek <jakub@redhat.com>
10814 Aldy Hernandez <aldyh@redhat.com>
10816 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
10817 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10818 * function.h (struct function): Add has_force_vect_loops and
10820 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10821 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10823 * gimple.def: Update CLAUSES comments.
10824 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10825 (gimple_build_omp_for): Add argument to prototype.
10826 (gimple_omp_for_kind): New.
10827 (gimple_omp_for_set_kind): New.
10828 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10829 GOVD_DATA_SHARE_CLASS.
10830 (enum omp_region_type): Add ORT_SIMD.
10831 (gimple_add_tmp_var): Handle ORT_SIMD.
10832 (gimplify_var_or_parm_decl): Same.
10833 (is_gimple_stmt): Same.
10834 (omp_firstprivatize_variable): Same.
10835 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10836 (omp_notice_variable): Handle ORT_SIMD.
10837 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10838 (omp_check_private): Handle ORT_SIMD.
10839 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10840 OMP_CLAUSE_SAFELEN.
10841 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10842 Handle OMP_CLAUSE_LASTPRIVATE.
10843 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10844 OMP_CLAUSE_SAFELEN.
10845 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10846 (gimplify_expr): Handle OMP_SIMD.
10847 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10848 (expand_GOMP_SIMD_VF): New.
10849 (expand_GOMP_SIMD_LAST_LANE): New.
10850 * internal-fn.def (GOMP_SIMD_LANE): New.
10851 (GOMP_SIMD_VF): New.
10852 (GOMP_SIMD_LAST_LANE): New.
10853 * omp-low.c: Include target.h.
10854 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
10855 OMP_CLAUSE_SAFELEN.
10856 (check_omp_nesting_restrictions): Same.
10858 (lower_rec_simd_input_clauses): New.
10859 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
10861 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
10862 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
10863 (expand_omp_build_assign): New.
10864 (expand_omp_for_init_counts): New.
10865 (expand_omp_for_init_vars): New.
10866 (extract_omp_for_update_vars): New.
10867 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
10868 and rewrite accordingly.
10869 (expand_omp_simd): New.
10870 (expand_omp_for): Use expand_omp_simd.
10871 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
10872 (lower_omp_for): Do not lower the body.
10873 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
10874 in their own loops.
10875 * tree-flow.h (find_omp_clause): Remove prototype.
10876 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
10877 forcing vectorization of the loop, or if flag_tree_vectorize.
10878 (gate_tree_if_conversion): Similarly.
10879 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
10880 gimple_build_omp_for.
10881 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
10882 * tree-parloops (create_parallel_loop): Pass kind argument to
10883 gimple_build_omp_for.
10884 * tree-pretty-print.c (dump_omp_clause): Add cases for
10885 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
10886 OMP_CLAUSE__SIMDUID_.
10887 (dump_generic_node): Handle OMP_SIMD.
10888 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
10889 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
10890 unroll OMP_SIMD loops here.
10891 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
10892 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
10894 (vect_analyze_data_refs): Handle simd loops.
10895 * tree-vect-loop.c (vectorizable_live_operation): Handle
10897 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
10898 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
10899 (vectorizable_load): Same.
10900 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
10901 (struct simduid_to_vf): New.
10902 (simduid_to_vf::hash): New.
10903 (simduid_to-vf::equal): New.
10904 (struct simd_array_to_simduid): New.
10905 (simd_array_to_simduid::hash): New.
10906 (simd_array_to_simduid::equal): New.
10907 (adjust_simduid_builtins): New.
10908 (struct note_simd_array_uses_struct): New.
10909 (note_simd_array_uses_cb): New.
10910 (note_simd_array_uses): New.
10911 (vectorize_loops): Handle simd hints and adjust simd builtins
10913 * tree-vectorizer.h (struct _stmt_vec_info): Add
10914 simd_lane_access_p field.
10915 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
10916 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
10917 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
10918 (omp_clause_code_name): Same.
10919 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
10920 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
10921 * tree.def (OMP_SIMD): New entry.
10922 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
10923 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
10924 (OMP_CLAUSE_DECL): Adjust range for new clauses.
10925 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
10926 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
10927 (OMP_CLAUSE_LINEAR_STEP): New.
10928 (OMP_CLAUSE_SAFELEN_EXPR): New.
10929 (OMP_CLAUSE__SIMDUID__DECL): New.
10930 (find_omp_clause): New prototype.
10932 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
10934 * config/i386/driver-i386.c (host_detect_local_cpu): Update
10935 Haswell processor detection.
10937 2013-08-27 Christian Widmer <shadow@umbrox.de>
10940 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
10941 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
10942 AVX2 capable processors.
10944 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
10946 * config/aarch64/arm_neon.h: Replace all inline asm implementations
10947 of vget_low_* with implementations in terms of other intrinsics.
10949 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
10951 PR middle-end/57219
10952 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
10953 values to -1, 0 and 1.
10955 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
10957 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
10958 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
10959 (<optab><mode>3_insn): Remove.
10960 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
10961 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
10962 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
10963 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
10964 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
10965 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
10966 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
10967 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
10968 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
10969 (ror<mode>3_insn): Likewise.
10970 * config/aarch64/predicates.md (aarch64_simd_register): New.
10972 2013-08-27 Richard Biener <rguenther@suse.de>
10974 PR tree-optimization/57521
10975 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
10976 one edge is non-critical.
10977 (find_phi_replacement_condition): Make sure to use a non-critical
10978 edge. Cleanup and remove old bug workarounds.
10979 (bb_postdominates_preds): Remove.
10980 (if_convertible_loop_p_1): Do not compute post-dominators.
10981 (combine_blocks): Do not free post-dominators.
10982 (main_tree_if_conversion): Likewise.
10983 (pass_data_if_conversion): Add TODO_verify_ssa.
10985 2013-08-27 DJ Delorie <dj@redhat.com>
10987 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
10989 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
10991 * function.c (assign_parm_find_data_types): Set passed_mode and
10992 nominal_mode to the TYPE_MODE of nominal_type for the built
10993 pointer type in case of the struct-pass-by-reference.
10995 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10997 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
10998 (UINT16_TYPE): Change default to "unsigned int".
11000 * config/avr/avr.opt (mfract-convert-truncate): New option.
11001 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
11002 is set, round negative fractional integers according to n1169
11003 when converting to integer types.
11005 2013-08-26 Jan Hubicka <jh@suse.cz>
11007 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
11008 methods can not be called indirectly when their address is not taken.
11010 2013-08-26 Jan Hubicka <jh@suse.cz>
11012 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
11015 2013-08-26 Jan Hubicka <jh@suse.cz>
11017 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11020 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11022 * reload.c (find_valid_class): Allow classes that do not include
11023 FIRST_PSEUDO_REGISTER - 1.
11025 2013-08-26 Jan Hubicka <jh@suse.cz>
11027 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11028 fix edge count/frequency when speculation failed; fix type check
11029 for the direct call.
11031 2013-08-26 Jan Hubicka <jh@suse.cz>
11033 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11035 2013-08-26 Jan Hubicka <jh@suse.cz>
11037 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11038 to be changed by edge redirection.
11040 2013-08-26 Jan Hubicka <jh@suse.cz>
11042 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11043 formating; add sanity check.
11044 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11045 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11046 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11047 (resolve_noninline_speculation): Update callee keys, too.
11049 2013-08-26 Jan Hubicka <jh@suse.cz>
11051 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11052 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11054 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11056 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11059 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11061 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11062 (x86_64_elf_select_section): Ditto.
11063 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11065 (ix86_pass_by_reference): Ditto.
11066 (output_set_got): Ditto.
11067 (ix86_unary_operator_ok): Ditto.
11068 (ix86_expand_builtin): Ditto.
11070 2013-08-23 Jan Hubicka <jh@suse.cz>
11072 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11074 2013-08-23 Jan Hubicka <jh@suse.cz>
11076 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11077 (tree_decl_with_vis): Add FINAL field.
11079 2013-08-23 Jeff Law <law@redhat.com>
11081 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11082 the debugging dump when the expression is fully redundant.
11084 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11086 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11087 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11088 * pretty-print.c (pp_formatted_text_data): Likewise.
11089 (pp_write_text_to_stream): Likewise.
11090 (pp_write_text_as_dot_label_to_stream): Likewise.
11091 (pp_append_r): Likewise.
11092 (pp_format): Likewise.
11093 (pp_flush): Likewise.
11094 (pp_clear_output_area): Likewise.
11095 (pp_append_text): Likewise.
11096 (pp_formatted_text): Likewise.
11097 (pp_remaining_character_count_for_line): Likewise.
11098 (pp_newline): Likewise.
11099 (pp_character): Likewise.
11100 (output_buffer::~output_buffer): Define.
11101 (pretty_printer::~pretty_printer): Destruct output buffer.
11102 * pretty-print.h (output_buffer::~output_buffer): Declare.
11103 (pretty_printer::~pretty_printer): Declare virtual.
11105 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11108 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11109 HOST_WIDE_INT_M1U): New macros.
11110 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11111 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11112 unsigned -1 for lshift.
11113 * cse.c (cse_insn): Likewise.
11114 * double-int.c (rshift_double, lshift_double): Likewise.
11115 * builtins.c (fold_builtin_bitop): Likewise.
11116 * combine.c (force_to_mode): Likewise.
11117 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11118 * simplify-rtx.c (simplify_const_unary_operation,
11119 simplify_const_binary_operation): Likewise.
11120 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11121 check_va_list_escapes): Likewise.
11122 * rtlanal.c (nonzero_bits1): Likewise.
11123 * expmed.c (expand_smod_pow2): Likewise.
11124 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11126 2013-08-23 Jan Hubicka <jh@suse.cz>
11128 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11129 as having address taken.
11131 2013-08-23 Jan Hubicka <jh@suse.cz>
11133 * ipa-utils.h (method_class_type): Declare.
11134 * ipa-devirt.c (method_class_type): Export.
11136 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11137 do not walk base classes of anonymous types.
11139 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11141 PR rtl-optimization/58220
11142 PR regression/58221
11143 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11144 handle SEQUENCE insns properly.
11146 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11148 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11150 (pp_newline_and_indent): Likewise.
11151 (pp_separate_with): Likewise.
11152 * pretty-print.c (pp_newline_and_flush): Define.
11153 (pp_newline_and_indent): Likewise.
11154 (pp_separate_with): Likewise.
11156 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11159 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11160 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
11162 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
11164 * config/i386/predicates.md (ext_sse_reg_operand): New.
11165 * config/i386/i386.md (*movti_internal): Use
11166 predicate to determine if EVEX is needed.
11167 (*movsi_internal): Ditto.
11168 (*movdf_internal): Ditto.
11169 (*movsf_internal): Ditto.
11170 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
11172 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11174 PR tree-optimization/58209
11175 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
11176 (find_tail_calls): Give up for pointer result types if m is non-NULL.
11177 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
11178 emit POINTER_PLUS_EXPR.
11179 (create_tailcall_accumulator): For pointer result type accumulate in
11182 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
11184 * configure.ac: Add backslashes missing from the last change.
11185 * configure: Regenerate.
11187 2013-08-22 Jan Hubicka <jh@susue.cz>
11189 * ipa.c (function_and_variable_visibility): First remember function
11190 was global and then make it local.
11192 2013-08-22 Julian Brown <julian@codesourcery.com>
11194 * configure.ac: Add aarch64 to list of arches which use "nop" in
11196 * configure: Regenerate.
11198 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11200 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
11201 gnu_libc_has_function.
11202 * config/s390/tpf.h: Likewise.
11204 2013-08-22 Jan Hubicka <jh@susue.cz>
11206 * timevar.c (validate_phases): Add cast.
11208 2013-08-22 Jan Hubicka <jh@susue.cz>
11210 * timevar.c (validate_phases): Use size_t for memory.
11211 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11213 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11215 * pretty-print.h (output_buffer::output_buffer): Declare.
11216 (pretty_printer::pretty_printer): Likewise.
11217 (pp_construct): Remove.
11218 * pretty-print.c (output_buffer::output_buffer): Define.
11219 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11220 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11221 (print_gimple_expr): Likewise.
11222 (print_gimple_seq): Likewise.
11223 (gimple_dump_bb): Likewise.
11224 * sched-vis.c (dump_value_slim): Likewise.
11225 (dump_insn_slim): Likewise.
11226 (dump_rtl_slim): Likewise.
11227 (str_pattern_slim): Likewise.
11228 * tree-mudflap.c (mf_varname_tree): Likewise.
11229 * graph.c (print_graph_cfg): Likewise.
11230 (start_graph_dump): Likewise.
11231 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11233 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11234 pretty printer initialization.
11235 * coretypes.h (diagnostic_context): Remove superflous type alias
11237 (pretty_printer): Likewise. Declare directly as a class.
11238 (pretty_print_info): Remove declaration as class.
11239 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11240 and pp_clear_output_area.
11241 (asan_add_global): Likewise.
11243 2013-08-22 Jan Hubicka <jh@suse.cz>
11245 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11246 * ipa-utils.h (update_type_inheritance_graph): Declare.
11247 (possible_polymorphic_call_target_p): Declare.
11248 (possible_polymorphic_call_target_p): New.
11249 * ipa-devirt.c: Update toplevel comments.
11250 (cached_polymorphic_call_targets): Move up.
11251 (odr_type_d): Move ID down.
11252 (polymorphic_type_binfo_p): Update comment.
11253 (odr_hasher::remove): Likewise;
11254 (get_odr_type): Set anonymous_namespace.
11255 (dump_odr_type): Dump it.
11256 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11257 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11258 (record_binfo): Add comment.
11259 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11261 (devirt_node_removal_hook): Do not iCE when cache is freed.
11262 (possible_polymorphic_call_target_p): New predicate.
11263 (update_type_inheritance_graph): New function.
11265 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11266 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11267 Sergey Lega <sergey.s.lega@intel.com>
11268 Anna Tikhonova <anna.tikhonova@intel.com>
11269 Ilya Tocar <ilya.tocar@intel.com>
11270 Andrey Turetskiy <andrey.turetskiy@intel.com>
11271 Ilya Verbin <ilya.verbin@intel.com>
11272 Kirill Yukhin <kirill.yukhin@intel.com>
11273 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11275 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11276 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11277 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11278 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11279 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11280 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11281 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11282 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11283 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11284 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11285 OPT_mavx512pf, OPT_mavx512er cases.
11286 * config/i386/constraints.md (v): New constraint.
11287 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11288 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11289 (bit_AVX512CD): New.
11290 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11291 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11292 * config/i386/i386-c.c (ix86_target_macros_internal):
11293 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11295 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11296 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11297 * config/i386/i386.c (regclass_map, dbx_register_map)
11298 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11299 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11300 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11301 -mavx512pf options.
11302 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11303 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11304 -mavx512pf options. Fix formatting.
11305 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11306 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11307 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11308 -mavx512cd, -mavx512pf options.
11309 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11310 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11311 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11313 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11314 (ix86_set_reg_reg_cost): Ditto.
11315 (x86_order_regs_for_local_alloc): Ditto.
11316 (MAX_VECT_LEN): Extend to 64-byte.
11317 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11318 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11319 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11320 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11321 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11322 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11323 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11324 (SSE_REG_MODE_P): Support new modes.
11325 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11326 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11327 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11328 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11329 (REG_CLASS_CONTENTS): Add new registers.
11330 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11331 (EXT_REX_SSE_REGNO_P): New.
11332 (HI_REGISTER_NAMES): Add new registers.
11333 * config/i386/i386.md: Define constants for new registers.
11334 (mode): Add new 512-bit modes.
11335 (prefix): Support evex prefix.
11336 (isa): Support avx512f, noavx512f, fma_avx512f.
11337 (ssemodesuffix): Add new 512-bit modes.
11339 (*movxi_internal_avx512f): Ditto.
11340 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11342 (*movsi_internal): Likewise.
11343 (*movdf_internal): Likewise.
11344 (*movsf_internal): Likewise.
11345 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11346 (<code><mode>3): Likewise.
11347 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11349 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11350 with the new constraint "v".
11351 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11353 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11354 with the new constraint "v".
11355 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11356 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11357 (*<plusminus_insn><mode>3): Likewise.
11358 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11359 (*mul<mode>3): Likewise.
11360 (<sse>_vmmul<mode>3): Likewise.
11361 (<sse>_div<mode>3): Likewise.
11362 (<sse>_vmdiv<mode>3): Likewise.
11363 (<sse>_sqrt<mode>2): Likewise.
11364 (<sse>_vmsqrt<mode>2): Likewise.
11365 (*<code><mode>3_finite): Likewise.
11366 (*<code><mode>3) <smaxmin>: Likewise.
11367 (<sse>_vm<code><mode>3): Likewise.
11368 (*<code><mode>3) <any_logic>: Likewise.
11369 (*fma_fmadd_<mode>): Likewise.
11370 (*fma_fmsub_<mode>): Likewise.
11371 (*fma_fnmadd_<mode>): Likewise.
11372 (*fma_fnmsub_<mode>): Likewise.
11373 (*fma_fmaddsub_<mode>): Likewise.
11374 (*fma_fmsubadd_<mode>): Likewise.
11375 (*fmai_fmadd_<mode>): Likewise.
11376 (*fmai_fmsub_<mode>): Likewise.
11377 (*fmai_fnmadd_<mode>): Likewise.
11378 (*fmai_fnmsub_<mode>): Likewise.
11379 (sse_cvtsi2ss): Likewise.
11380 (sse_cvtsi2ssq): Likewise.
11381 (sse_cvtss2si): Likewise.
11382 (sse_cvtss2si_2): Likewise.
11383 (sse_cvtss2siq): Likewise.
11384 (sse_cvtss2siq_2): Likewise.
11385 (sse_cvttss2si): Likewise.
11386 (sse_cvtss2siq_2): Likewise.
11387 (float<sseintvecmodelower><mode>2): Likewise.
11388 (sse2_cvtsd2si_2): Likewise.
11389 (sse2_cvtsd2siq_2): Likewise.
11390 (*<plusminus_insn><mode>3): Likewise.
11391 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11392 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11393 (ashr<mode>3): Likewise.
11394 (<shift_insn><mode>3): Likewise.
11395 (avx2_<code><mode>3): Likewise.
11396 (*avx2_<code><mode>3): Likewise.
11397 (*andnot<mode>3): Likewise.
11398 (*<code><mode>3) <any_logic>: Likewise.
11399 (abs<mode>2): Likewise.
11400 (avx2_permvar<mode>): Likewise.
11401 (avx2_perm<mode>_1): Likewise.
11402 (*avx_vpermilp<mode>): Likewise.
11403 (avx_vpermilvar<mode>3): Likewise.
11404 (avx2_ashrv<mode>): Likewise.
11405 (avx2_<shift_insn>v<mode>): Likewise.
11406 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11408 * doc/rtl.texi: Document XImode.
11410 2013-08-21 Jeff Law <law@redhat.com>
11412 * tree-flow.h (register_jump_thread): Pass vector of edges
11413 instead of each important edge.
11414 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11415 thread path into a vector and pass that to register_jump_thread.
11416 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11417 passed in edge vector to the current 3-edge form.
11420 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11422 * dce.c (fini_dce): Call df_analyze again just in case
11423 delete_unmarked_insns removed anything.
11425 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11427 * reload.h (struct reg_equivs): Rename to ..
11428 (struct reg_equivs_s): .. this.
11430 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11432 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11434 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11436 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11438 2013-08-21 Jeff Law <law@redhat.com>
11440 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11441 simplify assignments too. If the RHS collapses to a singleton
11442 range, then return the value for the range.
11444 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11446 * config/i386/sse.md (V16): Rename to...
11448 (mov<mode>): Update iterator name.
11449 (*mov<mode>_internal): Ditto.
11450 (push<mode>1): Ditto.
11451 (movmisalign<mode>): Ditto.
11453 2013-08-20 Jan Hubicka <jh@suse.cz>
11456 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11457 entry for direct edges.
11458 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11460 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11462 Revert my last two changes, r201865 and r201864:
11465 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11467 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11468 instances can own GC refs.
11470 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11471 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11472 (gcc::context::gt_pch_nx): Likewise.
11473 (gcc::context::gt_pch_nx): Likewise.
11474 * ggc.h (gt_ggc_mx <T>): New.
11475 (gt_pch_nx_with_op <T>): New.
11476 (gt_pch_nx <T>): New.
11477 * passes.c (opt_pass::gt_ggc_mx): New.
11478 (opt_pass::gt_pch_nx): New.
11479 (opt_pass::gt_pch_nx_with_op): New.
11480 (pass_manager::gt_ggc_mx): New.
11481 (pass_manager::gt_pch_nx): New.
11482 (pass_manager::gt_pch_nx_with_op): New.
11483 (pass_manager::operator new): Use
11484 ggc_internal_cleared_alloc_stat rather than xcalloc.
11485 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11486 (pass_manager::gt_ggc_mx): New.
11487 (pass_manager::gt_pch_nx): New.
11488 (pass_manager::gt_pch_nx_with_op): New.
11489 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11490 (opt_pass::operator new): New.
11491 (opt_pass::gt_ggc_mx): New.
11492 (opt_pass::gt_pch_nx): New.
11493 (opt_pass::gt_pch_nx_with_op): New.
11496 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11498 * Makefile.in (GTFILES): Add context.h.
11499 * context.c (gcc::context::operator new): New.
11500 (gcc::context::gt_ggc_mx): New.
11501 (gcc::context::gt_pch_nx): New.
11502 (gcc::context::gt_pch_nx): New.
11503 * context.h (gcc::context): Add GTY((user)) marking.
11504 (gcc::context::operator new): New.
11505 (gcc::context::gt_ggc_mx): New.
11506 (gcc::context::gt_pch_nx): New.
11507 (gcc::context::gt_pch_nx): New.
11508 (g): Add GTY marking.
11509 (gt_ggc_mx (gcc::context *)): New.
11510 (gt_pch_nx (gcc::context *)): New.
11511 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11512 void *cookie)): New.
11513 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11515 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11517 * dce.c (fini_dce): Call df_analyze again just in case
11518 delete_unmarked_insns removed anything.
11520 2013-08-20 Teresa Johnson <tejohnson@google.com>
11522 PR rtl-optimizations/57451
11523 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11524 from crossing split section boundaries.
11526 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11528 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11529 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11530 with MULTLIB_DEFAULTS.
11532 2013-08-20 Nick Clifton <nickc@redhat.com>
11534 * target.def (narrow_volatile_bitfield): Note that the default
11535 value is false, not !TARGET_STRICT_ALIGN.
11536 * doc/tm.texi: Regenerate.
11538 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11540 Fix LIB_SPEC for systems without libpthread.
11542 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11543 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11545 * config/i386/linux-common.h: Likewise.
11546 * config/mips/linux-common.h: Likewise.
11548 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11550 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11553 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11555 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11556 instances can own GC refs.
11558 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11559 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11560 (gcc::context::gt_pch_nx): Likewise.
11561 (gcc::context::gt_pch_nx): Likewise.
11562 * ggc.h (gt_ggc_mx <T>): New.
11563 (gt_pch_nx_with_op <T>): New.
11564 (gt_pch_nx <T>): New.
11565 * passes.c (opt_pass::gt_ggc_mx): New.
11566 (opt_pass::gt_pch_nx): New.
11567 (opt_pass::gt_pch_nx_with_op): New.
11568 (pass_manager::gt_ggc_mx): New.
11569 (pass_manager::gt_pch_nx): New.
11570 (pass_manager::gt_pch_nx_with_op): New.
11571 (pass_manager::operator new): Use
11572 ggc_internal_cleared_alloc_stat rather than xcalloc.
11573 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11574 (pass_manager::gt_ggc_mx): New.
11575 (pass_manager::gt_pch_nx): New.
11576 (pass_manager::gt_pch_nx_with_op): New.
11577 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11578 (opt_pass::operator new): New.
11579 (opt_pass::gt_ggc_mx): New.
11580 (opt_pass::gt_pch_nx): New.
11581 (opt_pass::gt_pch_nx_with_op): New.
11583 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11585 * Makefile.in (GTFILES): Add context.h.
11586 * context.c (gcc::context::operator new): New.
11587 (gcc::context::gt_ggc_mx): New.
11588 (gcc::context::gt_pch_nx): New.
11589 (gcc::context::gt_pch_nx): New.
11590 * context.h (gcc::context): Add GTY((user)) marking.
11591 (gcc::context::operator new): New.
11592 (gcc::context::gt_ggc_mx): New.
11593 (gcc::context::gt_pch_nx): New.
11594 (gcc::context::gt_pch_nx): New.
11595 (g): Add GTY marking.
11596 (gt_ggc_mx (gcc::context *)): New.
11597 (gt_pch_nx (gcc::context *)): New.
11598 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11599 void *cookie)): New.
11600 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11602 2013-08-20 Alan Modra <amodra@gmail.com>
11605 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11606 (rs6000_emit_epilogue): Likewise.
11608 2013-08-19 Dehao Chen <dehao@google.com>
11610 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11612 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11613 Jakub Jelinek <jakub@redhat.com>
11615 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11616 (BUILT_IN_FABSD64): Likewise.
11617 (BUILT_IN_FABSD128): Likewise.
11618 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11619 (fold_builtin_1): Likewise.
11620 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11621 destination and source operands.
11622 (*abstd2_fpr): Likewise.
11623 (*nabstd2_fpr): Likewise.
11625 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11627 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
11630 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
11632 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
11635 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
11637 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
11638 * builtins.c (default_libc_has_function): New.
11639 (gnu_libc_has_function): Ditto.
11640 (no_c99_libc_has_function): Ditto.
11641 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
11642 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
11643 (fold_builtin_sincos): Likewise.
11644 (fold_builtin_cexp): Likewise.
11645 * builtins.def (DEF_C94_BUILTIN): Likewise.
11646 (DEF_C99_BUILTIN): Likewise.
11647 (DEF_C99_C90RES_BUILTIN): Likewise.
11648 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
11649 definitions to using this define.
11650 * config/darwin-protos.h (darwin_libc_has_function): New.
11651 * config/darwin.c (darwin_libc_has_function): Ditto.
11652 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
11653 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
11654 * config/darwin.h: Ditto.
11655 * config/elfos.h: Ditto.
11656 * config/freebsd.h: Ditto.
11657 * config/i386/cygming.h: Ditto.
11658 * config/i386/djgpp.h: Ditto.
11659 * config/i386/i386-interix.h: Ditto.
11660 * config/microblaze/microblaze.h: Ditto.
11661 * config/mmix/mmix.h: Ditto.
11662 * config/gnu-user.h: Ditto.
11663 * config/ia64/hpux.h: Ditto.
11664 * config/pa/pa-hpux.h: Ditto.
11665 * config/pdp11/pdp11.h: Ditto.
11666 * config/picochip/picochip.h: Ditto.
11667 * config/linux.h: Ditto.
11668 * config/netbsd.h: Ditto.
11669 * config/openbsd.h: Ditto.
11670 * config/rs6000/aix43.h: Ditto.
11671 * config/rs6000/aix51.h: Ditto.
11672 * config/rs6000/aix52.h: Ditto.
11673 * config/rs6000/aix53.h: Ditto.
11674 * config/rs6000/aix61.h: Ditto.
11675 * config/rs6000/darwin.h: Ditto.
11676 * config/rs6000/linux.h: Ditto.
11677 * config/rs6000/linux64.h: Ditto.
11678 * config/s390/tpf.h: Ditto.
11679 * config/sol2-10.h: Ditto.
11680 * config/sol2.h: Ditto.
11681 * config/vms/vms.h: Ditto.
11682 * config/vxworks.h: Ditto.
11683 * config/linux-android.c (linux_android_libc_has_function):
11684 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
11685 * config/linux-protos.h (linux_android_libc_has_function):
11687 * config/i386/i386.c (ix86_libc_has_function): New.
11688 * config/i386/i386-protos.h
11689 (ix86_libc_has_function): New declaration.
11690 * config/i386/i386.md
11691 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
11692 ("isinf<mode>2): Likewise.
11693 * convert.c (convert_to_integer): Using new target hook
11694 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
11695 TARGET_C99_FUNCTIONS.
11696 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
11697 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
11698 * coretypes.h (function_class): New enum for different
11699 classes of functions.
11700 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
11701 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
11702 (TARGET_HAS_SINCOS): Likewise.
11703 (TARGET_LIBC_HAS_FUNCTION): New.
11704 * doc/tm.texi: Regenerated.
11705 * targhooks.h (default_libc_has_function): New declaration.
11706 (no_c99_libc_has_function): Ditto.
11707 (gnu_libc_has_function): Ditto.
11708 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
11709 and TARGET_HAS_SINCOS.
11711 2013-08-18 Jan Hubicka <jh@suse.cz>
11713 * Makeifle-in (ipa-devirt.o): New.
11714 (GTFILES): Add ipa-utils.h and ipa-devirt.c
11715 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
11716 (analyze_functions): Look into possible targets of polymorphic call.
11717 * dumpfile.c (dump_files): Add type-inheritance dump.
11718 * dumpfile.h (TDI_inheritance): New.
11719 * ipa-devirt.c: New file.
11720 * ipa-utils.h (odr_type_d): Forward declare.
11721 (odr_type): New type.
11722 (build_type_inheritance_graph): Declare.
11723 (possible_polymorphic_call_targets): Declare and introduce inline
11724 variant when only edge is pased.
11725 (dump_possible_polymorphic_call_targets): Likewise.
11726 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
11727 * tree.c (type_in_anonymous_namespace_p): Break out from ...
11728 (types_same_for_odr): ... here.
11729 * tree.h (type_in_anonymous_namespace_p): Declare.
11731 2013-08-18 Jakub Jelinek <jakub@redhat.com>
11733 PR tree-optimization/58006
11734 * tree-parloops.c (take_address_of): Don't ICE if get_name
11736 (eliminate_local_variables_stmt): Remove clobber stmts.
11738 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
11740 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
11741 error message is issued for an alias to undefined symbol.
11743 2013-08-18 Jan Hubicka <jh@suse.cz>
11745 * cgraph.c (cgraph_create_indirect_edge): Discover
11746 polymorphic calls and record basic info into indirect_info.
11747 * gimple-fold.c (gimple_fold_call): When doing BINFO based
11748 devirtualization, ignore objc function calls.
11749 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
11750 call with no parm index info.
11751 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
11752 * tree.c (virtual_method_call_p): New function.
11753 * tree.h (virtual_method_call_p): Declare.
11755 2013-08-16 Jan Hubicka <jh@suse.cz>
11757 PR middle-end/58179
11758 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
11760 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
11762 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
11765 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11767 * gengtype.c (type_for_name): Add special-case support for
11768 locating types within the "gcc::" namespace.
11769 (open_base_files): Emit a "using namespace gcc" directive.
11771 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11774 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
11775 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
11777 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
11778 array instead of each individual operand as a separate argument.
11779 (emit_fusion_gpr_load): Likewise.
11780 (expand_fusion_gpr_load): Add new function declaration.
11782 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
11783 signature to have the operands passed as an array, instead of as
11784 separate arguments. Allow ZERO_EXTEND to be in the memory
11785 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
11786 depend on the register live/dead flags when peepholes are run.
11787 (expand_fusion_gpr_load): New function to be called from the
11788 peephole2 pass, to change the register that addis sets to be the
11790 (emit_fusion_gpr_load): Change the calling signature to have the
11791 operands passed as an array, instead of as separate arguments.
11792 Allow ZERO_EXTEND to be in the memory address, and also
11793 SIGN_EXTEND if -mpower8-fusion-sign.
11795 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
11796 unspec enumeration.
11797 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
11798 adjust the register addis loads up in the peephole2 pass. Do not
11799 depend on the register live/dead state when the peephole pass is done.
11801 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11803 * gengtype.c (create_user_defined_type): Ensure that the kind
11804 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
11805 declaration is seen before the GTY((user)) marking.
11807 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
11810 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
11812 2013-08-16 Jan Hubicka <jh@suse.cz>
11814 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
11815 arugment expected_type.
11816 (gimple_fold_call): Use it.
11817 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11818 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11819 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11820 (try_make_edge_direct_virtual_call): Likewise.
11821 * tree.c (obj_type_ref_class): New.
11822 * tree.h (obj_type_ref_class): Use it.
11824 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11826 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11827 (rtl_slim_pp): Likewise.
11828 (init_rtl_slim_pretty_print): Likewise.
11829 (dump_value_slim): Don't call it. Use local pretty printer.
11830 (dump_insn_slim): Likewise.
11831 (dump_rtl_slim): Likewise.
11832 (str_pattern_slim): Likewise.
11833 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11836 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11838 PR tree-optimization/58164
11839 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11840 walk gimple_goto_dest of GIMPLE_GOTO.
11842 PR tree-optimization/58165
11843 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11844 bi_call must be the last stmt in a bb, don't split_block, instead
11845 use fallthru edge from it and give up if there is none.
11846 Release conds vector when returning early.
11848 2013-08-14 Xinliang David Li <davidxl@google.com>
11850 * config/i386/i386.c (ix86_option_override_internal):
11851 Remove unused variable and field.
11853 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11856 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
11857 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
11858 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
11859 and Linux, correct BLKmode alignment when 128-bit alignment is
11860 required and compatibility flag is not set.
11861 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
11862 for zero-size arguments when compatibility flag is not set.
11864 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11866 PR tree-optimization/58145
11867 * tree-sra.c (build_ref_for_offset): If prev_base has
11868 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
11870 2013-08-14 Xinliang David Li <davidxl@google.com>
11872 * config/i386/i386.c (ix86_option_override_internal):
11873 Fix uninitialized variable error.
11875 2013-08-14 Xinliang David Li <davidxl@google.com>
11877 * config/i386/i386.opt: Define two new options.
11878 * config/i386/x86-tune.def: Add arch selector field in macros.
11879 * config/i386/i386.h: Adjust macro definition.
11880 * config/i386/i386.c (ix86_option_override_internal):
11882 (parse_mtune_ctrl_str): New function.
11883 (set_ix86_tune_features): New function.
11884 (ix86_function_specific_restore): Call the new helper function.
11886 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
11888 PR rtl-optimization/57662
11889 * sel-sched.c (code_motion_process_successors): When the current insn
11890 is removed after the recursive traversal, break from the loop.
11891 Add comments and debug printouts.
11893 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11894 Alexandre Oliva <aoliva@redhat.com>
11897 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
11898 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
11899 there also UNSPEC_PLTOFF.
11901 2013-08-14 Marek Polacek <polacek@redhat.com>
11903 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
11904 NUM_CONDITIONS bit positions.
11906 2013-08-13 Cary Coutant <ccoutant@google.com>
11908 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
11909 (attr_checksum): Hash vector contents instead of pointer.
11910 (attr_checksum_ordered): Likewise.
11912 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
11914 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
11915 when Pmode != word_mode. Add length_address attribute.
11916 (sse3_monitor_<mode>): Merge from sse3_monitor and
11917 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
11918 Pmode != word_mode. Update insn length attribute.
11919 * config/i386/i386.c (ix86_option_override_internal): Update
11920 ix86_gen_monitor selection for merged sse3_monitor insn.
11922 2013-08-13 Julian Brown <julian@codesourcery.com>
11924 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
11925 perform invalid legitimization on greater-than-word-size modes for
11926 TARGET_E500_DOUBLE.
11928 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
11930 * ira.c (setup_class_translate_array): Use aclass instead of cl
11931 for classes not fully covered by allocno classes.
11933 2013-08-13 Jakub Jelinek <jakub@redhat.com>
11935 PR tree-optimization/57661
11936 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
11937 * tree-inline.c (tree_function_versioning): Initialize it.
11938 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
11939 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
11940 that is not being copied.
11943 * asan.c (instrument_strlen_call): Fix typo in comment.
11944 Use char * type even for the lhs of POINTER_PLUS_EXPR.
11946 2013-08-13 Steve Ellcey <sellcey@mips.com>
11948 * config/mips/mips.md (prefetch): Use lw instead of ld on
11949 loongson in 32bit mode.
11951 2013-08-13 Nick Clifton <nickc@redhat.com>
11953 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
11955 2013-08-13 Jan Hubicka <jh@suse.cz>
11957 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
11958 introduced edge; fix typo in sanity check.
11959 (cgraph_resolve_speculation): Export; improve diagnostic.
11960 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
11961 speculation at type mismatch.
11962 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
11963 (cgraph_resolve_speculation): Declare.
11964 (symtab_can_be_discarded): New function.
11965 * value-prof.c (gimple_ic_transform): Remove actual transform code.
11966 * ipa-inline-transform.c (speculation_removed): New global var.
11967 (clone_inlined_nodes): See if speculation can be removed.
11968 (inline_call): If speculations was removed, we growths may not match.
11969 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
11970 (speculation_useful_p): New function.
11971 (resolve_noninline_speculation): New function.
11972 (inline_small_functions): Resolve useless speculations.
11973 * ipa-inline.h (speculation_useful_p): Declare
11974 * ipa.c (can_replace_by_local_alias): Simplify.
11975 (ipa_profile): Produce speculative calls in non-lto, too;
11976 add simple cost model; produce local aliases.
11978 2013-08-13 David Malcolm <dmalcolm@redhat.com>
11980 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
11983 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
11985 * config/i386/i386.c (ix86_function_versions): Use error + inform.
11987 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
11989 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
11990 iterator instead of X87MODEF.
11992 2013-08-12 Perez Read <netfirewall@gmail.com>
11995 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
11996 operand 0 for intel asm alternative.
11997 (*movabs<mode>_2): Ditto for operand 1.
11999 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
12001 * config/aarch64/arm_none.h
12002 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
12004 2013-08-12 Nick Clifton <nickc@redhat.com>
12006 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
12008 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12010 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
12011 expand for QI/HImode operand to produce more effictive code for
12012 unsigned char(short) --> float(double) conversion.
12014 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
12016 * doc/invoke.texi: Mention that -ftls-model does not force the final
12019 2013-08-12 Marek Polacek <polacek@redhat.com>
12020 Marc Glisse <marc.glisse@inria.fr>
12022 PR tree-optimization/57980
12023 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12024 when creating -1 constant.
12026 2013-08-10 Jan Hubicka <jh@suse.cz>
12028 Workaround binutils PR14342.
12029 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12030 (gimple_init_edge_profiler): Likewise.
12031 (gimple_gen_ic_func_profiler): Likewise.
12033 2013-08-09 Jan Hubicka <jh@suse.cz>
12035 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12037 2013-08-09 Xinliang David Li <davidxl@google.com>
12039 * config/i386/stringop.def: New file.
12040 * config/i386/stringop.opt: New file.
12041 * config/i386/i386-opts.h: Include stringopt.def.
12042 * config/i386/i386.opt: Include stringopt.opt.
12043 * config/i386/i386.c (ix86_option_override_internal):
12044 Override default size based stringop inline strategies with options.
12045 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12048 2013-08-09 Jan Hubicka <jh@suse.cz>
12050 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12052 2013-08-09 Jan Hubicka <jh@suse.cz>
12054 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12056 (dump_cgraph_node): Dump profile-id.
12057 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12058 and common_target_probability.
12059 * lto-cgraph.c (lto_output_edge): Stream common targets.
12060 (lto_output_node): Stream profile ids.
12061 (input_node): Stream profile ids.
12062 (input_edge): Stream common targets.
12063 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12064 * ipa.c: Include value-prof.h
12065 (ipa_profile_generate_summary): Turn indirect call statement histograms
12066 into common targets.
12067 (ipa_profile): Turn common targets into speculative edges.
12069 2013-08-09 Jan Hubicka <jh@suse.cz>
12071 * cgraph.h (cgraph_node): Add profile_id.
12072 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12073 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12074 (del_node_map): Update.
12075 (find_func_by_funcdef_no): Replace by ...
12076 (find_func_by_profile_id): ... this one.
12077 (gimple_ic_transform): Do not remove useful histograms when
12078 speculation is not done; dump info when indirect call removal
12080 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12081 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12082 (__gcov_indirect_call_profiler_v2): .. this one.
12083 * profile.h (init_node_map): Update.
12084 * coverage.c (coverage_compute_profile_id): New function.
12085 * coverage.h (coverage_compute_profile_id): Declare.
12086 * tree-profile.c (init_ic_make_global_vars): Make
12087 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12088 (gimple_init_edge_profiler): Update prototype of
12089 __gcov_indirect_call_profiler.
12090 (gimple_gen_ic_func_profiler): Simplify.
12091 (tree_profiling): Use init_node_map
12093 2013-08-09 Jan Hubicka <jh@suse.cz>
12095 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12096 non-speculative refs.
12097 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12098 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12099 (cgraph_set_call_stmt): Likewise.
12100 (cgraph_create_edge_1): Fix release checking compilatoin;
12101 clear lto_stmt_uid.
12102 (cgraph_free_edge): Free indirect info.
12103 (cgraph_turn_edge_to_speculative): New function.
12104 (cgraph_speculative_call_info): New function.
12105 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12106 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12107 (dump_cgraph_node): Dump speculation.
12108 (verify_edge_count_and_frequency): Accept speculative edges.
12109 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12110 (verify_cgraph_node): Handle speculation.
12111 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12112 (cgraph_set_call_stmt): Update prototype.
12113 (cgraph_make_edge_direct): Update prototype.
12114 (cgraph_speculative_call_info): Declare.
12115 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12116 to change; update call of ipa_find_references.
12117 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12118 lto_stmt_uid and speculative flags.
12119 (ipa_dump_references): Dump speculation.
12120 (ipa_clone_references): Clone speculative flag.
12121 (ipa_clone_referring): Likewise.
12122 (ipa_clone_ref): New function.
12123 (ipa_find_reference): Look into lto_stmt_uids
12124 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12125 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12126 (ipa_find_reference): Update declaration.
12127 (ipa_clone_ref): Declare.
12128 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12129 stream speculative flag.
12130 (lto_output_ref): Stream statements uids and speculation.
12131 (input_ref): Likewise.
12132 (input_edge): Stream speuclation.
12133 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12134 (cgraph_set_call_stmt_including_clones): Handle speculation.
12135 * ipa-inline.c (heap_edge_removal_hook): New function.
12136 (inline_small_functions): Register it.
12137 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12138 also initialize refs.
12139 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12141 (try_make_edge_direct_simple_call): Likewise.
12142 (try_make_edge_direct_simple_call): Likewise.
12143 (update_indirect_edges_after_inlining): Likewise.
12144 (remove_described_reference): Look proper lto_stmt_uid.
12145 (propagate_controlled_uses): Likewise.
12146 (propagate_controlled_uses): Liekwise.
12147 * tree-inline.c (copy_bb): Copy speculative edges.
12148 (redirect_all_calls): New function.
12149 (copy_cfg_body): Do redirection after loop info is updated.
12150 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12152 2013-08-09 Jan Hubicka <jh@suse.cz>
12154 * lto-streamer-out.c (output_function): Renumber PHIs.
12155 * lto-streamer-in.c (input_function): Likewise.
12157 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12159 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12160 (get_lane_unsigned): Likewise.
12161 (dup_lane_scalar): Likewise.
12162 (get_lane): enable for VALL.
12163 * config/aarch64/aarch64-simd.md
12164 (aarch64_dup_lane_scalar<mode>): Remove.
12165 (aarch64_get_lane_signed<mode>): Likewise.
12166 (aarch64_get_lane_unsigned<mode>): Likewise.
12167 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
12168 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
12169 (aarch64_get_lane<mode>): Enable for all vector modes.
12170 (aarch64_get_lanedi): Remove misleading constraints.
12171 * config/aarch64/arm_neon.h
12172 (__aarch64_vget_lane_any): Define.
12173 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
12174 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
12175 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
12176 * config/aarch64/iterators.md (VDQQH): New.
12177 (VDQQHS): Likewise.
12178 (vwcore): Likewise.
12180 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
12182 * configure.ac: Add GAS check for LEON instructions on SPARC.
12183 * configure: Regenerate.
12184 * config.in: Likewise.
12185 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
12187 * config/sparc/sparc.opt (LEON, LEON3): New masks.
12188 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
12190 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
12191 (AS_LEON_FLAG): New macro.
12192 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
12193 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
12194 Deal with LEON and LEON3 for the memory model.
12195 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
12196 (atomic_compare_and_swap<mode>_1): Likewise.
12197 (*atomic_compare_and_swap<mode>_1): Likewise.
12199 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12201 * config/arm/neon.md (vcond): Fix floating-point vector
12202 comparisons against 0.
12204 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
12206 * lra-constraints.c (emit_spill_move): Remove assert.
12207 (process_alt_operands): Add more debugging
12208 output. Increase reject for spilling into memory. Decrease
12209 reject for reloading scratch.
12210 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12212 2013-08-08 Steve Ellcey <sellcey@mips.com>
12214 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12215 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12216 micromips incompatible. Add nan2008.
12217 (MULTILIB_DIRNAMES): Add nan2008.
12218 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12219 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12220 and micromips incompatible. Add nan2008.
12221 (MULTILIB_DIRNAMES): Add nan2008.
12222 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12224 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12226 PR rtl-optimization/58079
12227 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12228 simplify_comparison has widened a comparison with an integer.
12230 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12232 * config/arm/neon.md (movmisalign<mode>): Disable when we
12233 don't allow unaligned accesses.
12234 (*movmisalign<mode>_neon_store): Likewise.
12235 (*movmisalign<mode>_neon_load): Likewise.
12236 (*movmisalign<mode>_neon_store): Likewise.
12237 (*movmisalign<mode>_neon_load): Likewise.
12239 2013-08-08 Jan Hubicka <jh@suse.cz>
12241 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12242 (make_pass_rebuild_cgraph_edges): Also clear references.
12243 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12244 * ipa-inline-transform.c (inline_transform): Remove all references
12246 * cgraphunit.c (expand_function): Remove all references after
12248 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12249 (ipa_find_reference): Rewrite to iterator.
12250 (remove_stmt_references): Likewise.
12251 (ipa_clear_stmts_in_references): New function.
12252 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12253 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12255 * ipa-split.c (split_function): Remove references in split function.
12257 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12260 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12261 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12263 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12266 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12267 suggested mode for the assignment isn't compatible with the
12268 registers required.
12270 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12273 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12275 2013-08-07 Xinliang David Li <davidxl@google.com>
12277 * config/i386/i386.opt: New option -mtune-ctrl=.
12278 * config/i386/x86-tune.def: New file.
12279 * config/i386/i386.h: include x86-tune.def.
12280 * config/i386/i386.c (ix86_option_override_internal):
12281 Parsing -mtune-ctrl= option and set tune features.
12283 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12286 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12287 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12290 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12292 * rtl.h (update_alignments): Declare.
12293 * final.c (grow_label_align): New function extracted from...
12294 (shorten_branches): ...here. Call it.
12295 (update_alignments): New function.
12296 * reorg.c (sibling_labels): New variable.
12297 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12298 the new label along with it onto the sibling_labels vector.
12299 (fill_simple_delay_slots): Adjust call to get_label_before.
12300 (fill_slots_from_thread): Likewise.
12301 (relax_delay_slots): Likewise.
12302 (make_return_insns): Likewise.
12303 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12305 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12307 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12308 document its semantics.
12309 (diagnostic_report_diagnostic): Adjust accordingly.
12311 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12313 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12314 (sparc_option_override): ...and port to new C++ pass API.
12315 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12317 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12319 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12321 2013-08-06 Caroline Tice <cmtice@google.com>
12323 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12324 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12325 * tree-pass.h: Add pass_vtable_verify.
12326 * varasm.c (assemble_variable): Add code to properly set the comdat
12327 section and name for the .vtable_map_vars section.
12328 (assemble_vtyv_preinit_initializer): New function.
12329 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12331 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12332 * vtable-verify.c: New file.
12333 * vtable-verify.h: New file.
12334 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12335 initialiation levels.
12336 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12337 * passes.def: Insert pass_vtable_verify.
12338 * aclocal.m4: Reorder includes.
12339 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12340 -fvtv-counts options.
12341 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12342 as appropriate, if -fvtable-verify=... is used.
12343 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12344 -fvtable-verify=... is used.
12345 * Makefile.in (OBJS): Add vtable-verify.o to list.
12346 (vtable-verify.o): Add new build rule.
12347 (GTFILES): Add vtable-verify.c to list.
12348 * common.opt (fvtable-verify=): New flag.
12349 (vtv_priority): Values for fvtable-verify= flag.
12350 (fvtv-counts): New flag.
12351 (fvtv-debug): New flag.
12352 * tree.h (save_vtable_map_decl): New extern function decl.
12354 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12356 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12357 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12358 (pass_data_rl78_devirt): ...new pass_data instance and...
12359 (make_pass_rl78_devirt): ...new function.
12360 (rl78_asm_file_start): Port pass registration to new C++ API.
12362 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12364 * coretypes.h (rtl_opt_pass): Add.
12365 (gcc::context): Add.
12366 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12367 (epiphany_init): Port to new C++ pass API.
12368 (epiphany_optimize_mode_switching): Likewise.
12369 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12370 (pass_manager::get_pass_mode_switching): New.
12371 (pass_manager::get_pass_peephole2): New.
12372 * mode-switching.c (pass_mode_switching): Add clone method.
12373 * recog.c (pass_peephole2): Add clone method.
12374 (pass_split_all_insns): Add clone method.
12376 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12378 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12379 (mips_option_override): ...here, porting to new C++ API for passes.
12381 2013-08-06 Jan Hubicka <jh@suse.cz>
12383 * cgraph.c (cgraph_get_body): New function based on lto.c
12385 * cgraph.h (cgraph_get_body): Declare.
12386 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12388 * cgraphunit.c (expand_function): Get body prior expanding.
12389 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12390 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12392 * passes.c (do_per_function_toporder): Get body.
12393 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12394 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12396 2013-08-06 Martin Jambor <mjambor@suse.cz>
12399 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12400 is not re-finalized. Rename second parameter to no_collect.
12402 2013-08-06 Martin Jambor <mjambor@suse.cz>
12404 PR middle-end/58041
12405 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12406 MEM_REF has proper alignment information.
12408 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12411 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12413 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12414 rtx (*) (rtx, ...) with insn_gen_fn.
12415 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12416 insn_gen_fn::stored_funcptr. Add initializer braces.
12418 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12420 Rewrite how instances of passes are cloned to remove assumptions
12421 about their sizes (thus allowing pass subclasses to have
12422 additional data fields, albeit non-GC-managed ones at this point).
12424 * passes.c (make_pass_instance): Now that passes have clone
12425 methods, rewrite this function to eliminate XNEW and memcpy
12426 calls that used hardcoded sizes. Since this function no longer
12427 creates pass instances, rename it to...
12428 (add_pass_instance): ...this. Document the old way that passes were
12429 numbered and flagged, and rework this function to continue using it.
12430 (next_pass_1): Add an initial_pass argument for use by
12432 (position_pass): When adding multiple instances of a pass, use
12433 the pass's clone method, rather than relying on the XNEW/memcpy
12434 within the former make_pass_instance (now add_pass_instance).
12435 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12436 the initial instance of the current pass within the pass manager.
12438 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12440 This is the automated part of the conversion of passes from C
12441 structs to C++ classes.
12443 Patch autogenerated by refactor_passes.py from
12444 https://github.com/davidmalcolm/gcc-refactoring-scripts
12445 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12447 * asan.c (pass_asan): Convert from a global struct to a subclass of
12448 gimple_opt_pass along with...
12449 (pass_data_asan): ...new pass_data instance and...
12450 (make_pass_asan): ...new function.
12451 (pass_asan_O0): Convert from a global struct to a subclass of
12452 gimple_opt_pass along with...
12453 (pass_data_asan_O0): ...new pass_data instance and...
12454 (make_pass_asan_O0): ...new function.
12455 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12456 subclass of rtl_opt_pass along with...
12457 (pass_data_inc_dec): ...new pass_data instance and...
12458 (make_pass_inc_dec): ...new function.
12459 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12460 a subclass of rtl_opt_pass along with...
12461 (pass_data_reorder_blocks): ...new pass_data instance and...
12462 (make_pass_reorder_blocks): ...new function.
12463 (pass_duplicate_computed_gotos): Convert from a global struct to a
12464 subclass of rtl_opt_pass along with...
12465 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12466 (make_pass_duplicate_computed_gotos): ...new function.
12467 (pass_partition_blocks): Convert from a global struct to a subclass of
12468 rtl_opt_pass along with...
12469 (pass_data_partition_blocks): ...new pass_data instance and...
12470 (make_pass_partition_blocks): ...new function.
12471 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12472 struct to a subclass of rtl_opt_pass along with...
12473 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12475 (make_pass_branch_target_load_optimize1): ...new function.
12476 (pass_branch_target_load_optimize2): Convert from a global struct to a
12477 subclass of rtl_opt_pass along with...
12478 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12480 (make_pass_branch_target_load_optimize2): ...new function.
12481 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12482 of rtl_opt_pass along with...
12483 (pass_data_jump): ...new pass_data instance and...
12484 (make_pass_jump): ...new function.
12485 (pass_jump2): Convert from a global struct to a subclass of
12486 rtl_opt_pass along with...
12487 (pass_data_jump2): ...new pass_data instance and...
12488 (make_pass_jump2): ...new function.
12489 * cfgexpand.c (pass_expand): Convert from a global struct to a
12490 subclass of rtl_opt_pass along with...
12491 (pass_data_expand): ...new pass_data instance and...
12492 (make_pass_expand): ...new function.
12493 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12494 of rtl_opt_pass along with...
12495 (pass_data_free_cfg): ...new pass_data instance and...
12496 (make_pass_free_cfg): ...new function.
12497 (pass_into_cfg_layout_mode): Convert from a global struct to a
12498 subclass of rtl_opt_pass along with...
12499 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12500 (make_pass_into_cfg_layout_mode): ...new function.
12501 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12502 subclass of rtl_opt_pass along with...
12503 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12504 (make_pass_outof_cfg_layout_mode): ...new function.
12505 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12506 struct to a subclass of gimple_opt_pass along with...
12507 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12508 (make_pass_build_cgraph_edges): ...new function.
12509 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12510 subclass of gimple_opt_pass along with...
12511 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12512 (make_pass_rebuild_cgraph_edges): ...new function.
12513 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12514 subclass of gimple_opt_pass along with...
12515 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12517 (make_pass_remove_cgraph_callee_edges): ...new function.
12518 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12519 struct to a subclass of rtl_opt_pass along with...
12520 (pass_data_stack_adjustments): ...new pass_data instance and...
12521 (make_pass_stack_adjustments): ...new function.
12522 * combine.c (pass_combine): Convert from a global struct to a subclass
12523 of rtl_opt_pass along with...
12524 (pass_data_combine): ...new pass_data instance and...
12525 (make_pass_combine): ...new function.
12526 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12527 global struct to a subclass of rtl_opt_pass along with...
12528 (pass_data_compare_elim_after_reload): ...new pass_data instance
12530 (make_pass_compare_elim_after_reload): ...new function.
12531 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12532 of rtl_opt_pass along with...
12533 (pass_data_rtl_cprop): ...new pass_data instance and...
12534 (make_pass_rtl_cprop): ...new function.
12535 * cse.c (pass_cse): Convert from a global struct to a subclass of
12536 rtl_opt_pass along with...
12537 (pass_data_cse): ...new pass_data instance and...
12538 (make_pass_cse): ...new function.
12539 (pass_cse2): Convert from a global struct to a subclass of
12540 rtl_opt_pass along with...
12541 (pass_data_cse2): ...new pass_data instance and...
12542 (make_pass_cse2): ...new function.
12543 (pass_cse_after_global_opts): Convert from a global struct to a
12544 subclass of rtl_opt_pass along with...
12545 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12546 (make_pass_cse_after_global_opts): ...new function.
12547 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12548 of rtl_opt_pass along with...
12549 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12550 (make_pass_ud_rtl_dce): ...new function.
12551 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12552 rtl_opt_pass along with...
12553 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12554 (make_pass_fast_rtl_dce): ...new function.
12555 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12556 a subclass of rtl_opt_pass along with...
12557 (pass_data_df_initialize_opt): ...new pass_data instance and...
12558 (make_pass_df_initialize_opt): ...new function.
12559 (pass_df_initialize_no_opt): Convert from a global struct to a
12560 subclass of rtl_opt_pass along with...
12561 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12562 (make_pass_df_initialize_no_opt): ...new function.
12563 (pass_df_finish): Convert from a global struct to a subclass of
12564 rtl_opt_pass along with...
12565 (pass_data_df_finish): ...new pass_data instance and...
12566 (make_pass_df_finish): ...new function.
12567 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12568 rtl_opt_pass along with...
12569 (pass_data_rtl_dse1): ...new pass_data instance and...
12570 (make_pass_rtl_dse1): ...new function.
12571 (pass_rtl_dse2): Convert from a global struct to a subclass of
12572 rtl_opt_pass along with...
12573 (pass_data_rtl_dse2): ...new pass_data instance and...
12574 (make_pass_rtl_dse2): ...new function.
12575 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12576 subclass of rtl_opt_pass along with...
12577 (pass_data_dwarf2_frame): ...new pass_data instance and...
12578 (make_pass_dwarf2_frame): ...new function.
12579 * except.c (pass_set_nothrow_function_flags): Convert from a global
12580 struct to a subclass of rtl_opt_pass along with...
12581 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12583 (make_pass_set_nothrow_function_flags): ...new function.
12584 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12585 subclass of rtl_opt_pass along with...
12586 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12588 (make_pass_convert_to_eh_region_ranges): ...new function.
12589 * final.c (pass_compute_alignments): Convert from a global struct to a
12590 subclass of rtl_opt_pass along with...
12591 (pass_data_compute_alignments): ...new pass_data instance and...
12592 (make_pass_compute_alignments): ...new function.
12593 (pass_final): Convert from a global struct to a subclass of
12594 rtl_opt_pass along with...
12595 (pass_data_final): ...new pass_data instance and...
12596 (make_pass_final): ...new function.
12597 (pass_shorten_branches): Convert from a global struct to a subclass of
12598 rtl_opt_pass along with...
12599 (pass_data_shorten_branches): ...new pass_data instance and...
12600 (make_pass_shorten_branches): ...new function.
12601 (pass_clean_state): Convert from a global struct to a subclass of
12602 rtl_opt_pass along with...
12603 (pass_data_clean_state): ...new pass_data instance and...
12604 (make_pass_clean_state): ...new function.
12605 * function.c (pass_instantiate_virtual_regs): Convert from a global
12606 struct to a subclass of rtl_opt_pass along with...
12607 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12608 (make_pass_instantiate_virtual_regs): ...new function.
12609 (pass_leaf_regs): Convert from a global struct to a subclass of
12610 rtl_opt_pass along with...
12611 (pass_data_leaf_regs): ...new pass_data instance and...
12612 (make_pass_leaf_regs): ...new function.
12613 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12614 subclass of rtl_opt_pass along with...
12615 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12617 (make_pass_thread_prologue_and_epilogue): ...new function.
12618 (pass_match_asm_constraints): Convert from a global struct to a
12619 subclass of rtl_opt_pass along with...
12620 (pass_data_match_asm_constraints): ...new pass_data instance and...
12621 (make_pass_match_asm_constraints): ...new function.
12622 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12623 subclass of rtl_opt_pass along with...
12624 (pass_data_rtl_fwprop): ...new pass_data instance and...
12625 (make_pass_rtl_fwprop): ...new function.
12626 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12627 rtl_opt_pass along with...
12628 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12629 (make_pass_rtl_fwprop_addr): ...new function.
12630 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
12631 rtl_opt_pass along with...
12632 (pass_data_rtl_pre): ...new pass_data instance and...
12633 (make_pass_rtl_pre): ...new function.
12634 (pass_rtl_hoist): Convert from a global struct to a subclass of
12635 rtl_opt_pass along with...
12636 (pass_data_rtl_hoist): ...new pass_data instance and...
12637 (make_pass_rtl_hoist): ...new function.
12638 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
12639 subclass of gimple_opt_pass along with...
12640 (pass_data_lower_cf): ...new pass_data instance and...
12641 (make_pass_lower_cf): ...new function.
12642 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
12643 from a global struct to a subclass of gimple_opt_pass along with...
12644 (pass_data_strength_reduction): ...new pass_data instance and...
12645 (make_pass_strength_reduction): ...new function.
12646 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
12647 of rtl_opt_pass along with...
12648 (pass_data_rtl_ifcvt): ...new pass_data instance and...
12649 (make_pass_rtl_ifcvt): ...new function.
12650 (pass_if_after_combine): Convert from a global struct to a subclass of
12651 rtl_opt_pass along with...
12652 (pass_data_if_after_combine): ...new pass_data instance and...
12653 (make_pass_if_after_combine): ...new function.
12654 (pass_if_after_reload): Convert from a global struct to a subclass of
12655 rtl_opt_pass along with...
12656 (pass_data_if_after_reload): ...new pass_data instance and...
12657 (make_pass_if_after_reload): ...new function.
12658 * init-regs.c (pass_initialize_regs): Convert from a global struct to
12659 a subclass of rtl_opt_pass along with...
12660 (pass_data_initialize_regs): ...new pass_data instance and...
12661 (make_pass_initialize_regs): ...new function.
12662 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
12663 of ipa_opt_pass_d along with...
12664 (pass_data_ipa_cp): ...new pass_data instance and...
12665 (make_pass_ipa_cp): ...new function.
12666 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
12667 global struct to a subclass of gimple_opt_pass along with...
12668 (pass_data_inline_parameters): ...new pass_data instance and...
12669 (make_pass_inline_parameters): ...new function.
12670 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
12671 subclass of gimple_opt_pass along with...
12672 (pass_data_early_inline): ...new pass_data instance and...
12673 (make_pass_early_inline): ...new function.
12674 (pass_ipa_inline): Convert from a global struct to a subclass of
12675 ipa_opt_pass_d along with...
12676 (pass_data_ipa_inline): ...new pass_data instance and...
12677 (make_pass_ipa_inline): ...new function.
12678 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
12679 struct to a subclass of gimple_opt_pass along with...
12680 (pass_data_local_pure_const): ...new pass_data instance and...
12681 (make_pass_local_pure_const): ...new function.
12682 (pass_ipa_pure_const): Convert from a global struct to a subclass of
12683 ipa_opt_pass_d along with...
12684 (pass_data_ipa_pure_const): ...new pass_data instance and...
12685 (make_pass_ipa_pure_const): ...new function.
12686 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
12687 to a subclass of ipa_opt_pass_d along with...
12688 (pass_data_ipa_reference): ...new pass_data instance and...
12689 (make_pass_ipa_reference): ...new function.
12690 * ipa-split.c (pass_split_functions): Convert from a global struct to
12691 a subclass of gimple_opt_pass along with...
12692 (pass_data_split_functions): ...new pass_data instance and...
12693 (make_pass_split_functions): ...new function.
12694 (pass_feedback_split_functions): Convert from a global struct to a
12695 subclass of gimple_opt_pass along with...
12696 (pass_data_feedback_split_functions): ...new pass_data instance and...
12697 (make_pass_feedback_split_functions): ...new function.
12698 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
12699 global struct to a subclass of simple_ipa_opt_pass along with...
12700 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
12702 (make_pass_ipa_function_and_variable_visibility): ...new function.
12703 (pass_ipa_free_inline_summary): Convert from a global struct to a
12704 subclass of simple_ipa_opt_pass along with...
12705 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
12706 (make_pass_ipa_free_inline_summary): ...new function.
12707 (pass_ipa_whole_program_visibility): Convert from a global struct to a
12708 subclass of ipa_opt_pass_d along with...
12709 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
12711 (make_pass_ipa_whole_program_visibility): ...new function.
12712 (pass_ipa_profile): Convert from a global struct to a subclass of
12713 ipa_opt_pass_d along with...
12714 (pass_data_ipa_profile): ...new pass_data instance and...
12715 (make_pass_ipa_profile): ...new function.
12716 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
12717 ipa_opt_pass_d along with...
12718 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
12719 (make_pass_ipa_cdtor_merge): ...new function.
12720 * ira.c (pass_ira): Convert from a global struct to a subclass of
12721 rtl_opt_pass along with...
12722 (pass_data_ira): ...new pass_data instance and...
12723 (make_pass_ira): ...new function.
12724 (pass_reload): Convert from a global struct to a subclass of
12725 rtl_opt_pass along with...
12726 (pass_data_reload): ...new pass_data instance and...
12727 (make_pass_reload): ...new function.
12728 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
12729 subclass of rtl_opt_pass along with...
12730 (pass_data_cleanup_barriers): ...new pass_data instance and...
12731 (make_pass_cleanup_barriers): ...new function.
12732 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
12733 of rtl_opt_pass along with...
12734 (pass_data_loop2): ...new pass_data instance and...
12735 (make_pass_loop2): ...new function.
12736 (pass_rtl_loop_init): Convert from a global struct to a subclass of
12737 rtl_opt_pass along with...
12738 (pass_data_rtl_loop_init): ...new pass_data instance and...
12739 (make_pass_rtl_loop_init): ...new function.
12740 (pass_rtl_loop_done): Convert from a global struct to a subclass of
12741 rtl_opt_pass along with...
12742 (pass_data_rtl_loop_done): ...new pass_data instance and...
12743 (make_pass_rtl_loop_done): ...new function.
12744 (pass_rtl_move_loop_invariants): Convert from a global struct to a
12745 subclass of rtl_opt_pass along with...
12746 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
12747 (make_pass_rtl_move_loop_invariants): ...new function.
12748 (pass_rtl_unswitch): Convert from a global struct to a subclass of
12749 rtl_opt_pass along with...
12750 (pass_data_rtl_unswitch): ...new pass_data instance and...
12751 (make_pass_rtl_unswitch): ...new function.
12752 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
12753 subclass of rtl_opt_pass along with...
12754 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
12756 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12757 (pass_rtl_doloop): Convert from a global struct to a subclass of
12758 rtl_opt_pass along with...
12759 (pass_data_rtl_doloop): ...new pass_data instance and...
12760 (make_pass_rtl_doloop): ...new function.
12761 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
12762 a subclass of rtl_opt_pass along with...
12763 (pass_data_lower_subreg): ...new pass_data instance and...
12764 (make_pass_lower_subreg): ...new function.
12765 (pass_lower_subreg2): Convert from a global struct to a subclass of
12766 rtl_opt_pass along with...
12767 (pass_data_lower_subreg2): ...new pass_data instance and...
12768 (make_pass_lower_subreg2): ...new function.
12769 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
12770 struct to a subclass of ipa_opt_pass_d along with...
12771 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
12772 (make_pass_ipa_lto_gimple_out): ...new function.
12773 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
12774 of ipa_opt_pass_d along with...
12775 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
12776 (make_pass_ipa_lto_finish_out): ...new function.
12777 * mode-switching.c (pass_mode_switching): Convert from a global struct
12778 to a subclass of rtl_opt_pass along with...
12779 (pass_data_mode_switching): ...new pass_data instance and...
12780 (make_pass_mode_switching): ...new function.
12781 * modulo-sched.c (pass_sms): Convert from a global struct to a
12782 subclass of rtl_opt_pass along with...
12783 (pass_data_sms): ...new pass_data instance and...
12784 (make_pass_sms): ...new function.
12785 * omp-low.c (pass_expand_omp): Convert from a global struct to a
12786 subclass of gimple_opt_pass along with...
12787 (pass_data_expand_omp): ...new pass_data instance and...
12788 (make_pass_expand_omp): ...new function.
12789 (pass_lower_omp): Convert from a global struct to a subclass of
12790 gimple_opt_pass along with...
12791 (pass_data_lower_omp): ...new pass_data instance and...
12792 (make_pass_lower_omp): ...new function.
12793 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
12794 of gimple_opt_pass along with...
12795 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
12796 (make_pass_diagnose_omp_blocks): ...new function.
12797 * passes.c (pass_early_local_passes): Convert from a global struct to
12798 a subclass of simple_ipa_opt_pass along with...
12799 (pass_data_early_local_passes): ...new pass_data instance and...
12800 (make_pass_early_local_passes): ...new function.
12801 (pass_all_early_optimizations): Convert from a global struct to a
12802 subclass of gimple_opt_pass along with...
12803 (pass_data_all_early_optimizations): ...new pass_data instance and...
12804 (make_pass_all_early_optimizations): ...new function.
12805 (pass_all_optimizations): Convert from a global struct to a subclass
12806 of gimple_opt_pass along with...
12807 (pass_data_all_optimizations): ...new pass_data instance and...
12808 (make_pass_all_optimizations): ...new function.
12809 (pass_all_optimizations_g): Convert from a global struct to a subclass
12810 of gimple_opt_pass along with...
12811 (pass_data_all_optimizations_g): ...new pass_data instance and...
12812 (make_pass_all_optimizations_g): ...new function.
12813 (pass_rest_of_compilation): Convert from a global struct to a subclass
12814 of rtl_opt_pass along with...
12815 (pass_data_rest_of_compilation): ...new pass_data instance and...
12816 (make_pass_rest_of_compilation): ...new function.
12817 (pass_postreload): Convert from a global struct to a subclass of
12818 rtl_opt_pass along with...
12819 (pass_data_postreload): ...new pass_data instance and...
12820 (make_pass_postreload): ...new function.
12821 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12822 subclass of rtl_opt_pass along with...
12823 (pass_data_gcse2): ...new pass_data instance and...
12824 (make_pass_gcse2): ...new function.
12825 * postreload.c (pass_postreload_cse): Convert from a global struct to
12826 a subclass of rtl_opt_pass along with...
12827 (pass_data_postreload_cse): ...new pass_data instance and...
12828 (make_pass_postreload_cse): ...new function.
12829 * predict.c (pass_profile): Convert from a global struct to a subclass
12830 of gimple_opt_pass along with...
12831 (pass_data_profile): ...new pass_data instance and...
12832 (make_pass_profile): ...new function.
12833 (pass_strip_predict_hints): Convert from a global struct to a subclass
12834 of gimple_opt_pass along with...
12835 (pass_data_strip_predict_hints): ...new pass_data instance and...
12836 (make_pass_strip_predict_hints): ...new function.
12837 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12838 of rtl_opt_pass along with...
12839 (pass_data_peephole2): ...new pass_data instance and...
12840 (make_pass_peephole2): ...new function.
12841 (pass_split_all_insns): Convert from a global struct to a subclass of
12842 rtl_opt_pass along with...
12843 (pass_data_split_all_insns): ...new pass_data instance and...
12844 (make_pass_split_all_insns): ...new function.
12845 (pass_split_after_reload): Convert from a global struct to a subclass
12846 of rtl_opt_pass along with...
12847 (pass_data_split_after_reload): ...new pass_data instance and...
12848 (make_pass_split_after_reload): ...new function.
12849 (pass_split_before_regstack): Convert from a global struct to a
12850 subclass of rtl_opt_pass along with...
12851 (pass_data_split_before_regstack): ...new pass_data instance and...
12852 (make_pass_split_before_regstack): ...new function.
12853 (pass_split_before_sched2): Convert from a global struct to a subclass
12854 of rtl_opt_pass along with...
12855 (pass_data_split_before_sched2): ...new pass_data instance and...
12856 (make_pass_split_before_sched2): ...new function.
12857 (pass_split_for_shorten_branches): Convert from a global struct to a
12858 subclass of rtl_opt_pass along with...
12859 (pass_data_split_for_shorten_branches): ...new pass_data instance
12861 (make_pass_split_for_shorten_branches): ...new function.
12862 * ree.c (pass_ree): Convert from a global struct to a subclass of
12863 rtl_opt_pass along with...
12864 (pass_data_ree): ...new pass_data instance and...
12865 (make_pass_ree): ...new function.
12866 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
12867 subclass of rtl_opt_pass along with...
12868 (pass_data_stack_regs): ...new pass_data instance and...
12869 (make_pass_stack_regs): ...new function.
12870 (pass_stack_regs_run): Convert from a global struct to a subclass of
12871 rtl_opt_pass along with...
12872 (pass_data_stack_regs_run): ...new pass_data instance and...
12873 (make_pass_stack_regs_run): ...new function.
12874 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
12875 subclass of rtl_opt_pass along with...
12876 (pass_data_cprop_hardreg): ...new pass_data instance and...
12877 (make_pass_cprop_hardreg): ...new function.
12878 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
12879 subclass of rtl_opt_pass along with...
12880 (pass_data_reginfo_init): ...new pass_data instance and...
12881 (make_pass_reginfo_init): ...new function.
12882 * regmove.c (pass_regmove): Convert from a global struct to a subclass
12883 of rtl_opt_pass along with...
12884 (pass_data_regmove): ...new pass_data instance and...
12885 (make_pass_regmove): ...new function.
12886 * regrename.c (pass_regrename): Convert from a global struct to a
12887 subclass of rtl_opt_pass along with...
12888 (pass_data_regrename): ...new pass_data instance and...
12889 (make_pass_regrename): ...new function.
12890 * reorg.c (pass_delay_slots): Convert from a global struct to a
12891 subclass of rtl_opt_pass along with...
12892 (pass_data_delay_slots): ...new pass_data instance and...
12893 (make_pass_delay_slots): ...new function.
12894 (pass_machine_reorg): Convert from a global struct to a subclass of
12895 rtl_opt_pass along with...
12896 (pass_data_machine_reorg): ...new pass_data instance and...
12897 (make_pass_machine_reorg): ...new function.
12898 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
12899 of rtl_opt_pass along with...
12900 (pass_data_sched): ...new pass_data instance and...
12901 (make_pass_sched): ...new function.
12902 (pass_sched2): Convert from a global struct to a subclass of
12903 rtl_opt_pass along with...
12904 (pass_data_sched2): ...new pass_data instance and...
12905 (make_pass_sched2): ...new function.
12906 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
12907 to a subclass of rtl_opt_pass along with...
12908 (pass_data_stack_ptr_mod): ...new pass_data instance and...
12909 (make_pass_stack_ptr_mod): ...new function.
12910 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
12911 to a subclass of rtl_opt_pass along with...
12912 (pass_data_rtl_store_motion): ...new pass_data instance and...
12913 (make_pass_rtl_store_motion): ...new function.
12914 * tracer.c (pass_tracer): Convert from a global struct to a subclass
12915 of gimple_opt_pass along with...
12916 (pass_data_tracer): ...new pass_data instance and...
12917 (make_pass_tracer): ...new function.
12918 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
12919 to a subclass of gimple_opt_pass along with...
12920 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
12921 (make_pass_diagnose_tm_blocks): ...new function.
12922 (pass_lower_tm): Convert from a global struct to a subclass of
12923 gimple_opt_pass along with...
12924 (pass_data_lower_tm): ...new pass_data instance and...
12925 (make_pass_lower_tm): ...new function.
12926 (pass_tm_init): Convert from a global struct to a subclass of
12927 gimple_opt_pass along with...
12928 (pass_data_tm_init): ...new pass_data instance and...
12929 (make_pass_tm_init): ...new function.
12930 (pass_tm_mark): Convert from a global struct to a subclass of
12931 gimple_opt_pass along with...
12932 (pass_data_tm_mark): ...new pass_data instance and...
12933 (make_pass_tm_mark): ...new function.
12934 (pass_tm_edges): Convert from a global struct to a subclass of
12935 gimple_opt_pass along with...
12936 (pass_data_tm_edges): ...new pass_data instance and...
12937 (make_pass_tm_edges): ...new function.
12938 (pass_tm_memopt): Convert from a global struct to a subclass of
12939 gimple_opt_pass along with...
12940 (pass_data_tm_memopt): ...new pass_data instance and...
12941 (make_pass_tm_memopt): ...new function.
12942 (pass_ipa_tm): Convert from a global struct to a subclass of
12943 simple_ipa_opt_pass along with...
12944 (pass_data_ipa_tm): ...new pass_data instance and...
12945 (make_pass_ipa_tm): ...new function.
12946 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
12947 subclass of gimple_opt_pass along with...
12948 (pass_data_call_cdce): ...new pass_data instance and...
12949 (make_pass_call_cdce): ...new function.
12950 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
12951 subclass of gimple_opt_pass along with...
12952 (pass_data_build_cfg): ...new pass_data instance and...
12953 (make_pass_build_cfg): ...new function.
12954 (pass_split_crit_edges): Convert from a global struct to a subclass of
12955 gimple_opt_pass along with...
12956 (pass_data_split_crit_edges): ...new pass_data instance and...
12957 (make_pass_split_crit_edges): ...new function.
12958 (pass_warn_function_return): Convert from a global struct to a
12959 subclass of gimple_opt_pass along with...
12960 (pass_data_warn_function_return): ...new pass_data instance and...
12961 (make_pass_warn_function_return): ...new function.
12962 (pass_warn_function_noreturn): Convert from a global struct to a
12963 subclass of gimple_opt_pass along with...
12964 (pass_data_warn_function_noreturn): ...new pass_data instance and...
12965 (make_pass_warn_function_noreturn): ...new function.
12966 (pass_warn_unused_result): Convert from a global struct to a subclass
12967 of gimple_opt_pass along with...
12968 (pass_data_warn_unused_result): ...new pass_data instance and...
12969 (make_pass_warn_unused_result): ...new function.
12970 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
12971 a subclass of gimple_opt_pass along with...
12972 (pass_data_merge_phi): ...new pass_data instance and...
12973 (make_pass_merge_phi): ...new function.
12974 * tree-complex.c (pass_lower_complex): Convert from a global struct to
12975 a subclass of gimple_opt_pass along with...
12976 (pass_data_lower_complex): ...new pass_data instance and...
12977 (make_pass_lower_complex): ...new function.
12978 (pass_lower_complex_O0): Convert from a global struct to a subclass of
12979 gimple_opt_pass along with...
12980 (pass_data_lower_complex_O0): ...new pass_data instance and...
12981 (make_pass_lower_complex_O0): ...new function.
12982 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
12983 subclass of gimple_opt_pass along with...
12984 (pass_data_lower_eh): ...new pass_data instance and...
12985 (make_pass_lower_eh): ...new function.
12986 (pass_refactor_eh): Convert from a global struct to a subclass of
12987 gimple_opt_pass along with...
12988 (pass_data_refactor_eh): ...new pass_data instance and...
12989 (make_pass_refactor_eh): ...new function.
12990 (pass_lower_resx): Convert from a global struct to a subclass of
12991 gimple_opt_pass along with...
12992 (pass_data_lower_resx): ...new pass_data instance and...
12993 (make_pass_lower_resx): ...new function.
12994 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
12995 of gimple_opt_pass along with...
12996 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
12997 (make_pass_lower_eh_dispatch): ...new function.
12998 (pass_cleanup_eh): Convert from a global struct to a subclass of
12999 gimple_opt_pass along with...
13000 (pass_data_cleanup_eh): ...new pass_data instance and...
13001 (make_pass_cleanup_eh): ...new function.
13002 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
13003 to a subclass of simple_ipa_opt_pass along with...
13004 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
13005 (make_pass_ipa_lower_emutls): ...new function.
13006 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
13007 a subclass of gimple_opt_pass along with...
13008 (pass_data_if_conversion): ...new pass_data instance and...
13009 (make_pass_if_conversion): ...new function.
13010 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
13011 subclass of gimple_opt_pass along with...
13012 (pass_data_build_ssa): ...new pass_data instance and...
13013 (make_pass_build_ssa): ...new function.
13014 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
13015 global struct to a subclass of gimple_opt_pass along with...
13016 (pass_data_loop_distribution): ...new pass_data instance and...
13017 (make_pass_loop_distribution): ...new function.
13018 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13019 subclass of gimple_opt_pass along with...
13020 (pass_data_mudflap_1): ...new pass_data instance and...
13021 (make_pass_mudflap_1): ...new function.
13022 (pass_mudflap_2): Convert from a global struct to a subclass of
13023 gimple_opt_pass along with...
13024 (pass_data_mudflap_2): ...new pass_data instance and...
13025 (make_pass_mudflap_2): ...new function.
13026 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13027 subclass of gimple_opt_pass along with...
13028 (pass_data_mudflap_1): ...new pass_data instance and...
13029 (make_pass_mudflap_1): ...new function.
13030 (pass_mudflap_2): Convert from a global struct to a subclass of
13031 gimple_opt_pass along with...
13032 (pass_data_mudflap_2): ...new pass_data instance and...
13033 (make_pass_mudflap_2): ...new function.
13034 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13035 gimple_opt_pass along with...
13036 (pass_data_nrv): ...new pass_data instance and...
13037 (make_pass_nrv): ...new function.
13038 (pass_return_slot): Convert from a global struct to a subclass of
13039 gimple_opt_pass along with...
13040 (pass_data_return_slot): ...new pass_data instance and...
13041 (make_pass_return_slot): ...new function.
13042 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13043 to a subclass of gimple_opt_pass along with...
13044 (pass_data_object_sizes): ...new pass_data instance and...
13045 (make_pass_object_sizes): ...new function.
13046 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13047 global struct to a subclass of gimple_opt_pass along with...
13048 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13050 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13051 (pass_fixup_cfg): Convert from a global struct to a subclass of
13052 gimple_opt_pass along with...
13053 (pass_data_fixup_cfg): ...new pass_data instance and...
13054 (make_pass_fixup_cfg): ...new function.
13055 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13056 (make_pass_mudflap_1): ...new function.
13057 (pass_mudflap_2): Replace declaration with that of...
13058 (make_pass_mudflap_2): ...new function.
13059 (pass_asan): Replace declaration with that of...
13060 (make_pass_asan): ...new function.
13061 (pass_asan_O0): Replace declaration with that of...
13062 (make_pass_asan_O0): ...new function.
13063 (pass_tsan): Replace declaration with that of...
13064 (make_pass_tsan): ...new function.
13065 (pass_tsan_O0): Replace declaration with that of...
13066 (make_pass_tsan_O0): ...new function.
13067 (pass_lower_cf): Replace declaration with that of...
13068 (make_pass_lower_cf): ...new function.
13069 (pass_refactor_eh): Replace declaration with that of...
13070 (make_pass_refactor_eh): ...new function.
13071 (pass_lower_eh): Replace declaration with that of...
13072 (make_pass_lower_eh): ...new function.
13073 (pass_lower_eh_dispatch): Replace declaration with that of...
13074 (make_pass_lower_eh_dispatch): ...new function.
13075 (pass_lower_resx): Replace declaration with that of...
13076 (make_pass_lower_resx): ...new function.
13077 (pass_build_cfg): Replace declaration with that of...
13078 (make_pass_build_cfg): ...new function.
13079 (pass_early_tree_profile): Replace declaration with that of...
13080 (make_pass_early_tree_profile): ...new function.
13081 (pass_cleanup_eh): Replace declaration with that of...
13082 (make_pass_cleanup_eh): ...new function.
13083 (pass_sra): Replace declaration with that of...
13084 (make_pass_sra): ...new function.
13085 (pass_sra_early): Replace declaration with that of...
13086 (make_pass_sra_early): ...new function.
13087 (pass_early_ipa_sra): Replace declaration with that of...
13088 (make_pass_early_ipa_sra): ...new function.
13089 (pass_tail_recursion): Replace declaration with that of...
13090 (make_pass_tail_recursion): ...new function.
13091 (pass_tail_calls): Replace declaration with that of...
13092 (make_pass_tail_calls): ...new function.
13093 (pass_tree_loop): Replace declaration with that of...
13094 (make_pass_tree_loop): ...new function.
13095 (pass_tree_loop_init): Replace declaration with that of...
13096 (make_pass_tree_loop_init): ...new function.
13097 (pass_lim): Replace declaration with that of...
13098 (make_pass_lim): ...new function.
13099 (pass_tree_unswitch): Replace declaration with that of...
13100 (make_pass_tree_unswitch): ...new function.
13101 (pass_predcom): Replace declaration with that of...
13102 (make_pass_predcom): ...new function.
13103 (pass_iv_canon): Replace declaration with that of...
13104 (make_pass_iv_canon): ...new function.
13105 (pass_scev_cprop): Replace declaration with that of...
13106 (make_pass_scev_cprop): ...new function.
13107 (pass_empty_loop): Replace declaration with that of...
13108 (make_pass_empty_loop): ...new function.
13109 (pass_record_bounds): Replace declaration with that of...
13110 (make_pass_record_bounds): ...new function.
13111 (pass_graphite): Replace declaration with that of...
13112 (make_pass_graphite): ...new function.
13113 (pass_graphite_transforms): Replace declaration with that of...
13114 (make_pass_graphite_transforms): ...new function.
13115 (pass_if_conversion): Replace declaration with that of...
13116 (make_pass_if_conversion): ...new function.
13117 (pass_loop_distribution): Replace declaration with that of...
13118 (make_pass_loop_distribution): ...new function.
13119 (pass_vectorize): Replace declaration with that of...
13120 (make_pass_vectorize): ...new function.
13121 (pass_slp_vectorize): Replace declaration with that of...
13122 (make_pass_slp_vectorize): ...new function.
13123 (pass_complete_unroll): Replace declaration with that of...
13124 (make_pass_complete_unroll): ...new function.
13125 (pass_complete_unrolli): Replace declaration with that of...
13126 (make_pass_complete_unrolli): ...new function.
13127 (pass_parallelize_loops): Replace declaration with that of...
13128 (make_pass_parallelize_loops): ...new function.
13129 (pass_loop_prefetch): Replace declaration with that of...
13130 (make_pass_loop_prefetch): ...new function.
13131 (pass_iv_optimize): Replace declaration with that of...
13132 (make_pass_iv_optimize): ...new function.
13133 (pass_tree_loop_done): Replace declaration with that of...
13134 (make_pass_tree_loop_done): ...new function.
13135 (pass_ch): Replace declaration with that of...
13136 (make_pass_ch): ...new function.
13137 (pass_ccp): Replace declaration with that of...
13138 (make_pass_ccp): ...new function.
13139 (pass_phi_only_cprop): Replace declaration with that of...
13140 (make_pass_phi_only_cprop): ...new function.
13141 (pass_build_ssa): Replace declaration with that of...
13142 (make_pass_build_ssa): ...new function.
13143 (pass_build_alias): Replace declaration with that of...
13144 (make_pass_build_alias): ...new function.
13145 (pass_build_ealias): Replace declaration with that of...
13146 (make_pass_build_ealias): ...new function.
13147 (pass_dominator): Replace declaration with that of...
13148 (make_pass_dominator): ...new function.
13149 (pass_dce): Replace declaration with that of...
13150 (make_pass_dce): ...new function.
13151 (pass_dce_loop): Replace declaration with that of...
13152 (make_pass_dce_loop): ...new function.
13153 (pass_cd_dce): Replace declaration with that of...
13154 (make_pass_cd_dce): ...new function.
13155 (pass_call_cdce): Replace declaration with that of...
13156 (make_pass_call_cdce): ...new function.
13157 (pass_merge_phi): Replace declaration with that of...
13158 (make_pass_merge_phi): ...new function.
13159 (pass_split_crit_edges): Replace declaration with that of...
13160 (make_pass_split_crit_edges): ...new function.
13161 (pass_pre): Replace declaration with that of...
13162 (make_pass_pre): ...new function.
13163 (pass_profile): Replace declaration with that of...
13164 (make_pass_profile): ...new function.
13165 (pass_strip_predict_hints): Replace declaration with that of...
13166 (make_pass_strip_predict_hints): ...new function.
13167 (pass_lower_complex_O0): Replace declaration with that of...
13168 (make_pass_lower_complex_O0): ...new function.
13169 (pass_lower_complex): Replace declaration with that of...
13170 (make_pass_lower_complex): ...new function.
13171 (pass_lower_vector): Replace declaration with that of...
13172 (make_pass_lower_vector): ...new function.
13173 (pass_lower_vector_ssa): Replace declaration with that of...
13174 (make_pass_lower_vector_ssa): ...new function.
13175 (pass_lower_omp): Replace declaration with that of...
13176 (make_pass_lower_omp): ...new function.
13177 (pass_diagnose_omp_blocks): Replace declaration with that of...
13178 (make_pass_diagnose_omp_blocks): ...new function.
13179 (pass_expand_omp): Replace declaration with that of...
13180 (make_pass_expand_omp): ...new function.
13181 (pass_expand_omp_ssa): Replace declaration with that of...
13182 (make_pass_expand_omp_ssa): ...new function.
13183 (pass_object_sizes): Replace declaration with that of...
13184 (make_pass_object_sizes): ...new function.
13185 (pass_strlen): Replace declaration with that of...
13186 (make_pass_strlen): ...new function.
13187 (pass_fold_builtins): Replace declaration with that of...
13188 (make_pass_fold_builtins): ...new function.
13189 (pass_stdarg): Replace declaration with that of...
13190 (make_pass_stdarg): ...new function.
13191 (pass_early_warn_uninitialized): Replace declaration with that of...
13192 (make_pass_early_warn_uninitialized): ...new function.
13193 (pass_late_warn_uninitialized): Replace declaration with that of...
13194 (make_pass_late_warn_uninitialized): ...new function.
13195 (pass_cse_reciprocals): Replace declaration with that of...
13196 (make_pass_cse_reciprocals): ...new function.
13197 (pass_cse_sincos): Replace declaration with that of...
13198 (make_pass_cse_sincos): ...new function.
13199 (pass_optimize_bswap): Replace declaration with that of...
13200 (make_pass_optimize_bswap): ...new function.
13201 (pass_optimize_widening_mul): Replace declaration with that of...
13202 (make_pass_optimize_widening_mul): ...new function.
13203 (pass_warn_function_return): Replace declaration with that of...
13204 (make_pass_warn_function_return): ...new function.
13205 (pass_warn_function_noreturn): Replace declaration with that of...
13206 (make_pass_warn_function_noreturn): ...new function.
13207 (pass_cselim): Replace declaration with that of...
13208 (make_pass_cselim): ...new function.
13209 (pass_phiopt): Replace declaration with that of...
13210 (make_pass_phiopt): ...new function.
13211 (pass_forwprop): Replace declaration with that of...
13212 (make_pass_forwprop): ...new function.
13213 (pass_phiprop): Replace declaration with that of...
13214 (make_pass_phiprop): ...new function.
13215 (pass_tree_ifcombine): Replace declaration with that of...
13216 (make_pass_tree_ifcombine): ...new function.
13217 (pass_dse): Replace declaration with that of...
13218 (make_pass_dse): ...new function.
13219 (pass_nrv): Replace declaration with that of...
13220 (make_pass_nrv): ...new function.
13221 (pass_rename_ssa_copies): Replace declaration with that of...
13222 (make_pass_rename_ssa_copies): ...new function.
13223 (pass_sink_code): Replace declaration with that of...
13224 (make_pass_sink_code): ...new function.
13225 (pass_fre): Replace declaration with that of...
13226 (make_pass_fre): ...new function.
13227 (pass_check_data_deps): Replace declaration with that of...
13228 (make_pass_check_data_deps): ...new function.
13229 (pass_copy_prop): Replace declaration with that of...
13230 (make_pass_copy_prop): ...new function.
13231 (pass_vrp): Replace declaration with that of...
13232 (make_pass_vrp): ...new function.
13233 (pass_uncprop): Replace declaration with that of...
13234 (make_pass_uncprop): ...new function.
13235 (pass_return_slot): Replace declaration with that of...
13236 (make_pass_return_slot): ...new function.
13237 (pass_reassoc): Replace declaration with that of...
13238 (make_pass_reassoc): ...new function.
13239 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13240 (make_pass_rebuild_cgraph_edges): ...new function.
13241 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13242 (make_pass_remove_cgraph_callee_edges): ...new function.
13243 (pass_build_cgraph_edges): Replace declaration with that of...
13244 (make_pass_build_cgraph_edges): ...new function.
13245 (pass_local_pure_const): Replace declaration with that of...
13246 (make_pass_local_pure_const): ...new function.
13247 (pass_tracer): Replace declaration with that of...
13248 (make_pass_tracer): ...new function.
13249 (pass_warn_unused_result): Replace declaration with that of...
13250 (make_pass_warn_unused_result): ...new function.
13251 (pass_diagnose_tm_blocks): Replace declaration with that of...
13252 (make_pass_diagnose_tm_blocks): ...new function.
13253 (pass_lower_tm): Replace declaration with that of...
13254 (make_pass_lower_tm): ...new function.
13255 (pass_tm_init): Replace declaration with that of...
13256 (make_pass_tm_init): ...new function.
13257 (pass_tm_mark): Replace declaration with that of...
13258 (make_pass_tm_mark): ...new function.
13259 (pass_tm_memopt): Replace declaration with that of...
13260 (make_pass_tm_memopt): ...new function.
13261 (pass_tm_edges): Replace declaration with that of...
13262 (make_pass_tm_edges): ...new function.
13263 (pass_split_functions): Replace declaration with that of...
13264 (make_pass_split_functions): ...new function.
13265 (pass_feedback_split_functions): Replace declaration with that of...
13266 (make_pass_feedback_split_functions): ...new function.
13267 (pass_strength_reduction): Replace declaration with that of...
13268 (make_pass_strength_reduction): ...new function.
13269 (pass_ipa_lower_emutls): Replace declaration with that of...
13270 (make_pass_ipa_lower_emutls): ...new function.
13271 (pass_ipa_function_and_variable_visibility): Replace declaration with
13273 (make_pass_ipa_function_and_variable_visibility): ...new function.
13274 (pass_ipa_tree_profile): Replace declaration with that of...
13275 (make_pass_ipa_tree_profile): ...new function.
13276 (pass_early_local_passes): Replace declaration with that of...
13277 (make_pass_early_local_passes): ...new function.
13278 (pass_ipa_whole_program_visibility): Replace declaration with that
13280 (make_pass_ipa_whole_program_visibility): ...new function.
13281 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13282 (make_pass_ipa_lto_gimple_out): ...new function.
13283 (pass_ipa_increase_alignment): Replace declaration with that of...
13284 (make_pass_ipa_increase_alignment): ...new function.
13285 (pass_ipa_inline): Replace declaration with that of...
13286 (make_pass_ipa_inline): ...new function.
13287 (pass_ipa_free_lang_data): Replace declaration with that of...
13288 (make_pass_ipa_free_lang_data): ...new function.
13289 (pass_ipa_free_inline_summary): Replace declaration with that of...
13290 (make_pass_ipa_free_inline_summary): ...new function.
13291 (pass_ipa_cp): Replace declaration with that of...
13292 (make_pass_ipa_cp): ...new function.
13293 (pass_ipa_reference): Replace declaration with that of...
13294 (make_pass_ipa_reference): ...new function.
13295 (pass_ipa_pure_const): Replace declaration with that of...
13296 (make_pass_ipa_pure_const): ...new function.
13297 (pass_ipa_pta): Replace declaration with that of...
13298 (make_pass_ipa_pta): ...new function.
13299 (pass_ipa_lto_finish_out): Replace declaration with that of...
13300 (make_pass_ipa_lto_finish_out): ...new function.
13301 (pass_ipa_tm): Replace declaration with that of...
13302 (make_pass_ipa_tm): ...new function.
13303 (pass_ipa_profile): Replace declaration with that of...
13304 (make_pass_ipa_profile): ...new function.
13305 (pass_ipa_cdtor_merge): Replace declaration with that of...
13306 (make_pass_ipa_cdtor_merge): ...new function.
13307 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13309 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13310 (pass_init_datastructures): Replace declaration with that of...
13311 (make_pass_init_datastructures): ...new function.
13312 (pass_fixup_cfg): Replace declaration with that of...
13313 (make_pass_fixup_cfg): ...new function.
13314 (pass_expand): Replace declaration with that of...
13315 (make_pass_expand): ...new function.
13316 (pass_instantiate_virtual_regs): Replace declaration with that of...
13317 (make_pass_instantiate_virtual_regs): ...new function.
13318 (pass_rtl_fwprop): Replace declaration with that of...
13319 (make_pass_rtl_fwprop): ...new function.
13320 (pass_rtl_fwprop_addr): Replace declaration with that of...
13321 (make_pass_rtl_fwprop_addr): ...new function.
13322 (pass_jump): Replace declaration with that of...
13323 (make_pass_jump): ...new function.
13324 (pass_jump2): Replace declaration with that of...
13325 (make_pass_jump2): ...new function.
13326 (pass_lower_subreg): Replace declaration with that of...
13327 (make_pass_lower_subreg): ...new function.
13328 (pass_cse): Replace declaration with that of...
13329 (make_pass_cse): ...new function.
13330 (pass_fast_rtl_dce): Replace declaration with that of...
13331 (make_pass_fast_rtl_dce): ...new function.
13332 (pass_ud_rtl_dce): Replace declaration with that of...
13333 (make_pass_ud_rtl_dce): ...new function.
13334 (pass_rtl_dce): Replace declaration with that of...
13335 (make_pass_rtl_dce): ...new function.
13336 (pass_rtl_dse1): Replace declaration with that of...
13337 (make_pass_rtl_dse1): ...new function.
13338 (pass_rtl_dse2): Replace declaration with that of...
13339 (make_pass_rtl_dse2): ...new function.
13340 (pass_rtl_dse3): Replace declaration with that of...
13341 (make_pass_rtl_dse3): ...new function.
13342 (pass_rtl_cprop): Replace declaration with that of...
13343 (make_pass_rtl_cprop): ...new function.
13344 (pass_rtl_pre): Replace declaration with that of...
13345 (make_pass_rtl_pre): ...new function.
13346 (pass_rtl_hoist): Replace declaration with that of...
13347 (make_pass_rtl_hoist): ...new function.
13348 (pass_rtl_store_motion): Replace declaration with that of...
13349 (make_pass_rtl_store_motion): ...new function.
13350 (pass_cse_after_global_opts): Replace declaration with that of...
13351 (make_pass_cse_after_global_opts): ...new function.
13352 (pass_rtl_ifcvt): Replace declaration with that of...
13353 (make_pass_rtl_ifcvt): ...new function.
13354 (pass_into_cfg_layout_mode): Replace declaration with that of...
13355 (make_pass_into_cfg_layout_mode): ...new function.
13356 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13357 (make_pass_outof_cfg_layout_mode): ...new function.
13358 (pass_loop2): Replace declaration with that of...
13359 (make_pass_loop2): ...new function.
13360 (pass_rtl_loop_init): Replace declaration with that of...
13361 (make_pass_rtl_loop_init): ...new function.
13362 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13363 (make_pass_rtl_move_loop_invariants): ...new function.
13364 (pass_rtl_unswitch): Replace declaration with that of...
13365 (make_pass_rtl_unswitch): ...new function.
13366 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13367 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13368 (pass_rtl_doloop): Replace declaration with that of...
13369 (make_pass_rtl_doloop): ...new function.
13370 (pass_rtl_loop_done): Replace declaration with that of...
13371 (make_pass_rtl_loop_done): ...new function.
13372 (pass_web): Replace declaration with that of...
13373 (make_pass_web): ...new function.
13374 (pass_cse2): Replace declaration with that of...
13375 (make_pass_cse2): ...new function.
13376 (pass_df_initialize_opt): Replace declaration with that of...
13377 (make_pass_df_initialize_opt): ...new function.
13378 (pass_df_initialize_no_opt): Replace declaration with that of...
13379 (make_pass_df_initialize_no_opt): ...new function.
13380 (pass_reginfo_init): Replace declaration with that of...
13381 (make_pass_reginfo_init): ...new function.
13382 (pass_inc_dec): Replace declaration with that of...
13383 (make_pass_inc_dec): ...new function.
13384 (pass_stack_ptr_mod): Replace declaration with that of...
13385 (make_pass_stack_ptr_mod): ...new function.
13386 (pass_initialize_regs): Replace declaration with that of...
13387 (make_pass_initialize_regs): ...new function.
13388 (pass_combine): Replace declaration with that of...
13389 (make_pass_combine): ...new function.
13390 (pass_if_after_combine): Replace declaration with that of...
13391 (make_pass_if_after_combine): ...new function.
13392 (pass_ree): Replace declaration with that of...
13393 (make_pass_ree): ...new function.
13394 (pass_partition_blocks): Replace declaration with that of...
13395 (make_pass_partition_blocks): ...new function.
13396 (pass_match_asm_constraints): Replace declaration with that of...
13397 (make_pass_match_asm_constraints): ...new function.
13398 (pass_regmove): Replace declaration with that of...
13399 (make_pass_regmove): ...new function.
13400 (pass_split_all_insns): Replace declaration with that of...
13401 (make_pass_split_all_insns): ...new function.
13402 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13403 (make_pass_fast_rtl_byte_dce): ...new function.
13404 (pass_lower_subreg2): Replace declaration with that of...
13405 (make_pass_lower_subreg2): ...new function.
13406 (pass_mode_switching): Replace declaration with that of...
13407 (make_pass_mode_switching): ...new function.
13408 (pass_sms): Replace declaration with that of...
13409 (make_pass_sms): ...new function.
13410 (pass_sched): Replace declaration with that of...
13411 (make_pass_sched): ...new function.
13412 (pass_ira): Replace declaration with that of...
13413 (make_pass_ira): ...new function.
13414 (pass_reload): Replace declaration with that of...
13415 (make_pass_reload): ...new function.
13416 (pass_clean_state): Replace declaration with that of...
13417 (make_pass_clean_state): ...new function.
13418 (pass_branch_prob): Replace declaration with that of...
13419 (make_pass_branch_prob): ...new function.
13420 (pass_value_profile_transformations): Replace declaration with that
13422 (make_pass_value_profile_transformations): ...new function.
13423 (pass_postreload_cse): Replace declaration with that of...
13424 (make_pass_postreload_cse): ...new function.
13425 (pass_gcse2): Replace declaration with that of...
13426 (make_pass_gcse2): ...new function.
13427 (pass_split_after_reload): Replace declaration with that of...
13428 (make_pass_split_after_reload): ...new function.
13429 (pass_branch_target_load_optimize1): Replace declaration with that
13431 (make_pass_branch_target_load_optimize1): ...new function.
13432 (pass_thread_prologue_and_epilogue): Replace declaration with that
13434 (make_pass_thread_prologue_and_epilogue): ...new function.
13435 (pass_stack_adjustments): Replace declaration with that of...
13436 (make_pass_stack_adjustments): ...new function.
13437 (pass_peephole2): Replace declaration with that of...
13438 (make_pass_peephole2): ...new function.
13439 (pass_if_after_reload): Replace declaration with that of...
13440 (make_pass_if_after_reload): ...new function.
13441 (pass_regrename): Replace declaration with that of...
13442 (make_pass_regrename): ...new function.
13443 (pass_cprop_hardreg): Replace declaration with that of...
13444 (make_pass_cprop_hardreg): ...new function.
13445 (pass_reorder_blocks): Replace declaration with that of...
13446 (make_pass_reorder_blocks): ...new function.
13447 (pass_branch_target_load_optimize2): Replace declaration with that
13449 (make_pass_branch_target_load_optimize2): ...new function.
13450 (pass_leaf_regs): Replace declaration with that of...
13451 (make_pass_leaf_regs): ...new function.
13452 (pass_split_before_sched2): Replace declaration with that of...
13453 (make_pass_split_before_sched2): ...new function.
13454 (pass_compare_elim_after_reload): Replace declaration with that of...
13455 (make_pass_compare_elim_after_reload): ...new function.
13456 (pass_sched2): Replace declaration with that of...
13457 (make_pass_sched2): ...new function.
13458 (pass_stack_regs): Replace declaration with that of...
13459 (make_pass_stack_regs): ...new function.
13460 (pass_stack_regs_run): Replace declaration with that of...
13461 (make_pass_stack_regs_run): ...new function.
13462 (pass_df_finish): Replace declaration with that of...
13463 (make_pass_df_finish): ...new function.
13464 (pass_compute_alignments): Replace declaration with that of...
13465 (make_pass_compute_alignments): ...new function.
13466 (pass_duplicate_computed_gotos): Replace declaration with that of...
13467 (make_pass_duplicate_computed_gotos): ...new function.
13468 (pass_variable_tracking): Replace declaration with that of...
13469 (make_pass_variable_tracking): ...new function.
13470 (pass_free_cfg): Replace declaration with that of...
13471 (make_pass_free_cfg): ...new function.
13472 (pass_machine_reorg): Replace declaration with that of...
13473 (make_pass_machine_reorg): ...new function.
13474 (pass_cleanup_barriers): Replace declaration with that of...
13475 (make_pass_cleanup_barriers): ...new function.
13476 (pass_delay_slots): Replace declaration with that of...
13477 (make_pass_delay_slots): ...new function.
13478 (pass_split_for_shorten_branches): Replace declaration with that of...
13479 (make_pass_split_for_shorten_branches): ...new function.
13480 (pass_split_before_regstack): Replace declaration with that of...
13481 (make_pass_split_before_regstack): ...new function.
13482 (pass_convert_to_eh_region_ranges): Replace declaration with that
13484 (make_pass_convert_to_eh_region_ranges): ...new function.
13485 (pass_shorten_branches): Replace declaration with that of...
13486 (make_pass_shorten_branches): ...new function.
13487 (pass_set_nothrow_function_flags): Replace declaration with that of...
13488 (make_pass_set_nothrow_function_flags): ...new function.
13489 (pass_dwarf2_frame): Replace declaration with that of...
13490 (make_pass_dwarf2_frame): ...new function.
13491 (pass_final): Replace declaration with that of...
13492 (make_pass_final): ...new function.
13493 (pass_rtl_seqabstr): Replace declaration with that of...
13494 (make_pass_rtl_seqabstr): ...new function.
13495 (pass_release_ssa_names): Replace declaration with that of...
13496 (make_pass_release_ssa_names): ...new function.
13497 (pass_early_inline): Replace declaration with that of...
13498 (make_pass_early_inline): ...new function.
13499 (pass_inline_parameters): Replace declaration with that of...
13500 (make_pass_inline_parameters): ...new function.
13501 (pass_update_address_taken): Replace declaration with that of...
13502 (make_pass_update_address_taken): ...new function.
13503 (pass_convert_switch): Replace declaration with that of...
13504 (make_pass_convert_switch): ...new function.
13505 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13506 to a subclass of simple_ipa_opt_pass along with...
13507 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13508 (make_pass_ipa_tree_profile): ...new function.
13509 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13510 subclass of gimple_opt_pass along with...
13511 (pass_data_sra_early): ...new pass_data instance and...
13512 (make_pass_sra_early): ...new function.
13513 (pass_sra): Convert from a global struct to a subclass of
13514 gimple_opt_pass along with...
13515 (pass_data_sra): ...new pass_data instance and...
13516 (make_pass_sra): ...new function.
13517 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13518 gimple_opt_pass along with...
13519 (pass_data_early_ipa_sra): ...new pass_data instance and...
13520 (make_pass_early_ipa_sra): ...new function.
13521 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13522 subclass of gimple_opt_pass along with...
13523 (pass_data_ccp): ...new pass_data instance and...
13524 (make_pass_ccp): ...new function.
13525 (pass_fold_builtins): Convert from a global struct to a subclass of
13526 gimple_opt_pass along with...
13527 (pass_data_fold_builtins): ...new pass_data instance and...
13528 (make_pass_fold_builtins): ...new function.
13529 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13530 subclass of gimple_opt_pass along with...
13531 (pass_data_copy_prop): ...new pass_data instance and...
13532 (make_pass_copy_prop): ...new function.
13533 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13534 global struct to a subclass of gimple_opt_pass along with...
13535 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13536 (make_pass_rename_ssa_copies): ...new function.
13537 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13538 subclass of gimple_opt_pass along with...
13539 (pass_data_dce): ...new pass_data instance and...
13540 (make_pass_dce): ...new function.
13541 (pass_dce_loop): Convert from a global struct to a subclass of
13542 gimple_opt_pass along with...
13543 (pass_data_dce_loop): ...new pass_data instance and...
13544 (make_pass_dce_loop): ...new function.
13545 (pass_cd_dce): Convert from a global struct to a subclass of
13546 gimple_opt_pass along with...
13547 (pass_data_cd_dce): ...new pass_data instance and...
13548 (make_pass_cd_dce): ...new function.
13549 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13550 subclass of gimple_opt_pass along with...
13551 (pass_data_dominator): ...new pass_data instance and...
13552 (make_pass_dominator): ...new function.
13553 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13554 gimple_opt_pass along with...
13555 (pass_data_phi_only_cprop): ...new pass_data instance and...
13556 (make_pass_phi_only_cprop): ...new function.
13557 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13558 subclass of gimple_opt_pass along with...
13559 (pass_data_dse): ...new pass_data instance and...
13560 (make_pass_dse): ...new function.
13561 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13562 a subclass of gimple_opt_pass along with...
13563 (pass_data_forwprop): ...new pass_data instance and...
13564 (make_pass_forwprop): ...new function.
13565 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13566 struct to a subclass of gimple_opt_pass along with...
13567 (pass_data_tree_ifcombine): ...new pass_data instance and...
13568 (make_pass_tree_ifcombine): ...new function.
13569 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13570 subclass of gimple_opt_pass along with...
13571 (pass_data_ch): ...new pass_data instance and...
13572 (make_pass_ch): ...new function.
13573 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13574 subclass of gimple_opt_pass along with...
13575 (pass_data_tree_loop): ...new pass_data instance and...
13576 (make_pass_tree_loop): ...new function.
13577 (pass_tree_loop_init): Convert from a global struct to a subclass of
13578 gimple_opt_pass along with...
13579 (pass_data_tree_loop_init): ...new pass_data instance and...
13580 (make_pass_tree_loop_init): ...new function.
13581 (pass_lim): Convert from a global struct to a subclass of
13582 gimple_opt_pass along with...
13583 (pass_data_lim): ...new pass_data instance and...
13584 (make_pass_lim): ...new function.
13585 (pass_tree_unswitch): Convert from a global struct to a subclass of
13586 gimple_opt_pass along with...
13587 (pass_data_tree_unswitch): ...new pass_data instance and...
13588 (make_pass_tree_unswitch): ...new function.
13589 (pass_predcom): Convert from a global struct to a subclass of
13590 gimple_opt_pass along with...
13591 (pass_data_predcom): ...new pass_data instance and...
13592 (make_pass_predcom): ...new function.
13593 (pass_vectorize): Convert from a global struct to a subclass of
13594 gimple_opt_pass along with...
13595 (pass_data_vectorize): ...new pass_data instance and...
13596 (make_pass_vectorize): ...new function.
13597 (pass_graphite): Convert from a global struct to a subclass of
13598 gimple_opt_pass along with...
13599 (pass_data_graphite): ...new pass_data instance and...
13600 (make_pass_graphite): ...new function.
13601 (pass_graphite_transforms): Convert from a global struct to a subclass
13602 of gimple_opt_pass along with...
13603 (pass_data_graphite_transforms): ...new pass_data instance and...
13604 (make_pass_graphite_transforms): ...new function.
13605 (pass_check_data_deps): Convert from a global struct to a subclass of
13606 gimple_opt_pass along with...
13607 (pass_data_check_data_deps): ...new pass_data instance and...
13608 (make_pass_check_data_deps): ...new function.
13609 (pass_iv_canon): Convert from a global struct to a subclass of
13610 gimple_opt_pass along with...
13611 (pass_data_iv_canon): ...new pass_data instance and...
13612 (make_pass_iv_canon): ...new function.
13613 (pass_scev_cprop): Convert from a global struct to a subclass of
13614 gimple_opt_pass along with...
13615 (pass_data_scev_cprop): ...new pass_data instance and...
13616 (make_pass_scev_cprop): ...new function.
13617 (pass_record_bounds): Convert from a global struct to a subclass of
13618 gimple_opt_pass along with...
13619 (pass_data_record_bounds): ...new pass_data instance and...
13620 (make_pass_record_bounds): ...new function.
13621 (pass_complete_unroll): Convert from a global struct to a subclass of
13622 gimple_opt_pass along with...
13623 (pass_data_complete_unroll): ...new pass_data instance and...
13624 (make_pass_complete_unroll): ...new function.
13625 (pass_complete_unrolli): Convert from a global struct to a subclass of
13626 gimple_opt_pass along with...
13627 (pass_data_complete_unrolli): ...new pass_data instance and...
13628 (make_pass_complete_unrolli): ...new function.
13629 (pass_parallelize_loops): Convert from a global struct to a subclass
13630 of gimple_opt_pass along with...
13631 (pass_data_parallelize_loops): ...new pass_data instance and...
13632 (make_pass_parallelize_loops): ...new function.
13633 (pass_loop_prefetch): Convert from a global struct to a subclass of
13634 gimple_opt_pass along with...
13635 (pass_data_loop_prefetch): ...new pass_data instance and...
13636 (make_pass_loop_prefetch): ...new function.
13637 (pass_iv_optimize): Convert from a global struct to a subclass of
13638 gimple_opt_pass along with...
13639 (pass_data_iv_optimize): ...new pass_data instance and...
13640 (make_pass_iv_optimize): ...new function.
13641 (pass_tree_loop_done): Convert from a global struct to a subclass of
13642 gimple_opt_pass along with...
13643 (pass_data_tree_loop_done): ...new pass_data instance and...
13644 (make_pass_tree_loop_done): ...new function.
13645 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
13646 struct to a subclass of gimple_opt_pass along with...
13647 (pass_data_cse_reciprocals): ...new pass_data instance and...
13648 (make_pass_cse_reciprocals): ...new function.
13649 (pass_cse_sincos): Convert from a global struct to a subclass of
13650 gimple_opt_pass along with...
13651 (pass_data_cse_sincos): ...new pass_data instance and...
13652 (make_pass_cse_sincos): ...new function.
13653 (pass_optimize_bswap): Convert from a global struct to a subclass of
13654 gimple_opt_pass along with...
13655 (pass_data_optimize_bswap): ...new pass_data instance and...
13656 (make_pass_optimize_bswap): ...new function.
13657 (pass_optimize_widening_mul): Convert from a global struct to a
13658 subclass of gimple_opt_pass along with...
13659 (pass_data_optimize_widening_mul): ...new pass_data instance and...
13660 (make_pass_optimize_widening_mul): ...new function.
13661 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
13662 subclass of gimple_opt_pass along with...
13663 (pass_data_phiopt): ...new pass_data instance and...
13664 (make_pass_phiopt): ...new function.
13665 (pass_cselim): Convert from a global struct to a subclass of
13666 gimple_opt_pass along with...
13667 (pass_data_cselim): ...new pass_data instance and...
13668 (make_pass_cselim): ...new function.
13669 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
13670 subclass of gimple_opt_pass along with...
13671 (pass_data_phiprop): ...new pass_data instance and...
13672 (make_pass_phiprop): ...new function.
13673 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
13674 subclass of gimple_opt_pass along with...
13675 (pass_data_pre): ...new pass_data instance and...
13676 (make_pass_pre): ...new function.
13677 (pass_fre): Convert from a global struct to a subclass of
13678 gimple_opt_pass along with...
13679 (pass_data_fre): ...new pass_data instance and...
13680 (make_pass_fre): ...new function.
13681 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
13682 subclass of gimple_opt_pass along with...
13683 (pass_data_reassoc): ...new pass_data instance and...
13684 (make_pass_reassoc): ...new function.
13685 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
13686 subclass of gimple_opt_pass along with...
13687 (pass_data_sink_code): ...new pass_data instance and...
13688 (make_pass_sink_code): ...new function.
13689 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
13690 subclass of gimple_opt_pass along with...
13691 (pass_data_strlen): ...new pass_data instance and...
13692 (make_pass_strlen): ...new function.
13693 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
13694 struct to a subclass of gimple_opt_pass along with...
13695 (pass_data_build_alias): ...new pass_data instance and...
13696 (make_pass_build_alias): ...new function.
13697 (pass_build_ealias): Convert from a global struct to a subclass of
13698 gimple_opt_pass along with...
13699 (pass_data_build_ealias): ...new pass_data instance and...
13700 (make_pass_build_ealias): ...new function.
13701 (pass_ipa_pta): Convert from a global struct to a subclass of
13702 simple_ipa_opt_pass along with...
13703 (pass_data_ipa_pta): ...new pass_data instance and...
13704 (make_pass_ipa_pta): ...new function.
13705 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
13706 subclass of gimple_opt_pass along with...
13707 (pass_data_uncprop): ...new pass_data instance and...
13708 (make_pass_uncprop): ...new function.
13709 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
13710 global struct to a subclass of gimple_opt_pass along with...
13711 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
13712 (make_pass_late_warn_uninitialized): ...new function.
13713 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
13714 to a subclass of gimple_opt_pass along with...
13715 (pass_data_init_datastructures): ...new pass_data instance and...
13716 (make_pass_init_datastructures): ...new function.
13717 (pass_early_warn_uninitialized): Convert from a global struct to a
13718 subclass of gimple_opt_pass along with...
13719 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
13720 (make_pass_early_warn_uninitialized): ...new function.
13721 (pass_update_address_taken): Convert from a global struct to a
13722 subclass of gimple_opt_pass along with...
13723 (pass_data_update_address_taken): ...new pass_data instance and...
13724 (make_pass_update_address_taken): ...new function.
13725 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
13726 struct to a subclass of gimple_opt_pass along with...
13727 (pass_data_release_ssa_names): ...new pass_data instance and...
13728 (make_pass_release_ssa_names): ...new function.
13729 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
13730 subclass of gimple_opt_pass along with...
13731 (pass_data_stdarg): ...new pass_data instance and...
13732 (make_pass_stdarg): ...new function.
13733 * tree-switch-conversion.c (pass_convert_switch): Convert from a
13734 global struct to a subclass of gimple_opt_pass along with...
13735 (pass_data_convert_switch): ...new pass_data instance and...
13736 (make_pass_convert_switch): ...new function.
13737 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
13738 to a subclass of gimple_opt_pass along with...
13739 (pass_data_tail_recursion): ...new pass_data instance and...
13740 (make_pass_tail_recursion): ...new function.
13741 (pass_tail_calls): Convert from a global struct to a subclass of
13742 gimple_opt_pass along with...
13743 (pass_data_tail_calls): ...new pass_data instance and...
13744 (make_pass_tail_calls): ...new function.
13745 * tree-vect-generic.c (pass_lower_vector): Convert from a global
13746 struct to a subclass of gimple_opt_pass along with...
13747 (pass_data_lower_vector): ...new pass_data instance and...
13748 (make_pass_lower_vector): ...new function.
13749 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
13750 gimple_opt_pass along with...
13751 (pass_data_lower_vector_ssa): ...new pass_data instance and...
13752 (make_pass_lower_vector_ssa): ...new function.
13753 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
13754 to a subclass of gimple_opt_pass along with...
13755 (pass_data_slp_vectorize): ...new pass_data instance and...
13756 (make_pass_slp_vectorize): ...new function.
13757 (pass_ipa_increase_alignment): Convert from a global struct to a
13758 subclass of simple_ipa_opt_pass along with...
13759 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
13760 (make_pass_ipa_increase_alignment): ...new function.
13761 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
13762 gimple_opt_pass along with...
13763 (pass_data_vrp): ...new pass_data instance and...
13764 (make_pass_vrp): ...new function.
13765 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
13766 subclass of simple_ipa_opt_pass along with...
13767 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
13768 (make_pass_ipa_free_lang_data): ...new function.
13769 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
13770 gimple_opt_pass along with...
13771 (pass_data_tsan): ...new pass_data instance and...
13772 (make_pass_tsan): ...new function.
13773 (pass_tsan_O0): Convert from a global struct to a subclass of
13774 gimple_opt_pass along with...
13775 (pass_data_tsan_O0): ...new pass_data instance and...
13776 (make_pass_tsan_O0): ...new function.
13777 * var-tracking.c (pass_variable_tracking): Convert from a global
13778 struct to a subclass of rtl_opt_pass along with...
13779 (pass_data_variable_tracking): ...new pass_data instance and...
13780 (make_pass_variable_tracking): ...new function.
13781 * web.c (pass_web): Convert from a global struct to a subclass of
13782 rtl_opt_pass along with...
13783 (pass_data_web): ...new pass_data instance and...
13784 (make_pass_web): ...new function.
13785 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
13786 declaration with that of...
13787 (make_pass_mode_switch_use): ...new function.
13788 (pass_resolve_sw_modes): Replace declaration with that of...
13789 (make_pass_resolve_sw_modes): ...new function.
13790 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
13791 from a global struct to a subclass of rtl_opt_pass along with...
13792 (pass_data_mode_switch_use): ...new pass_data instance and...
13793 (make_pass_mode_switch_use): ...new function.
13794 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
13795 from a global struct to a subclass of rtl_opt_pass along with...
13796 (pass_data_resolve_sw_modes): ...new pass_data instance and...
13797 (make_pass_resolve_sw_modes): ...new function.
13798 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
13799 struct to a subclass of rtl_opt_pass along with...
13800 (pass_data_insert_vzeroupper): ...new pass_data instance and...
13801 (make_pass_insert_vzeroupper): ...new function.
13802 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
13803 global struct to a subclass of rtl_opt_pass along with...
13804 (pass_data_work_around_errata): ...new pass_data instance and...
13805 (make_pass_work_around_errata): ...new function.
13806 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
13807 struct to a subclass of rtl_opt_pass along with...
13808 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
13809 (make_pass_mips_machine_reorg2): ...new function.
13811 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13813 * passes.c (pass_manager::operator new): New.
13815 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13817 Handwritten part of conversion of passes to C++ classes.
13819 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13820 (toplev.o): Add dep on PASS_MANAGER_H.
13821 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13822 of early local pases to reflect this moving from a global to a
13823 member of gcc::pass_manager.
13824 (cgraph_add_new_function): Likewise.
13825 * lto-cgraph.c (lto_output_node): Update for conversion of
13826 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13827 * passes.c (opt_pass::clone): New.
13828 (opt_pass::gate): New.
13829 (opt_pass::execute): New.
13830 (opt_pass::opt_pass): New.
13831 (pass_manager::execute_early_local_passes): New.
13832 (pass_manager::execute_pass_mode_switching): new.
13833 (finish_optimization_passes): Convert to...
13834 (pass_manager::finish_optimization_passes): ...this.
13835 (finish_optimization_passes): Update for conversion of passes to
13837 (register_dump_files_1): Use has_gate since we cannot portably
13838 check a vtable entry against NULL.
13839 (dump_one_pass): Likewise.
13840 (ipa_write_summaries_2): Likewise.
13841 (ipa_write_optimization_summaries_1): Likewise.
13842 (ipa_read_summaries_1): Likewise.
13843 (ipa_read_optimization_summaries_1): Likewise.
13844 (execute_ipa_stmt_fixups): Likewise.
13845 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13846 pass-creation functions rather than wiring up globals, and
13847 storing the results in fields of pass_manager generated using
13848 pass-instances.def.
13849 (pass_manager::dump_profile_report): Update for conversion of
13850 passes to C++ classes.
13851 (pass_manager::execute_ipa_summary_passes): Likewise.
13852 (execute_one_ipa_transform_pass): Likewise.
13853 (execute_one_pass): Use has_gate and has_execute since we cannot
13854 portably check a vtable entry against NULL.
13855 * pass_manager.h (pass_manager::finish_optimization_passes): New.
13856 (pass_manager): Use pass-instances.def to add fields for the
13857 various pass instances.
13858 * toplev.c (finalize): Update for move of
13859 finish_optimization_passes to a method of gcc::pass_manager.
13860 * toplev.h (finish_optimization_passes): Move to method of class
13862 * tree-pass.h (struct pass_data): New.
13863 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
13864 (opt_pass::gate): Convert to virtual function.
13865 (opt_pass::~opt_pass): New.
13866 (opt_pass::clone): New.
13867 (opt_pass::execute): Convert to virtual function.
13868 (opt_pass::opt_pass): New.
13869 (opt_pass::ctxt_): new.
13870 (gimple_opt_pass): Convert to subclass of opt_pass.
13871 (gimple_opt_pass::gimple_opt_pass): New.
13872 (rtl_opt_pass): Convert to subclass of opt_pass.
13873 (rtl_opt_pass::rtl_opt_pass): New.
13874 (ipa_opt_pass_d): Convert to subclass of opt_pass.
13875 (ipa_opt_pass_d::ipa_opt_pass_d): New.
13876 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
13877 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
13878 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
13879 invocation of pass_mode_switching to reflect this moving from a
13880 global to a member of gcc::pass_manager.
13881 (ix86_option_override): Rework how pass_insert_vzeroupper is
13882 added to the pass_manager to reflect autogenerated changes.
13883 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
13885 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
13887 PR rtl-optimization/57708
13888 * recog.c (peep2_find_free_register): Validate all regs in a
13891 2013-08-05 Jan Hubicka <jh@suse.cz>
13894 * cgraph.c (verify_cgraph_node): Accept local flags from other
13896 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
13897 (function_and_variable_visibility): Likewise.
13898 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
13900 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13902 * graph.c (init_graph_slim_pretty_print): Remove.
13903 (print_graph_cfg): Do not call it. Use local pretty printer.
13904 (start_graph_dump): Likewise.
13906 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13908 * gimple-pretty-print.c (buffer): Remove.
13909 (initialized): Likewise.
13910 (maybe_init_pretty_print): Likewise.
13911 (print_gimple_stmt): Do not call it. Use non-static local
13912 pretty_printer variable.
13913 (print_gimple_expr): Likewise.
13914 (print_gimple_seq): Likewise.
13915 (gimple_dump_bb): Likewise.
13917 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13919 * asan.c (asan_pp): Remove.
13920 (asan_pp_initialized): Likewise.
13921 (asan_pp_initialize): Likewise.
13922 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
13923 (asan_emit_stack_protection): Tidy. Use local pretty printer.
13924 (asan_add_global): Likewise.
13926 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
13928 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
13929 * diagnostic.h (diagnostic_flush_buffer): Adjust.
13930 * pretty-print.c (pp_formatted_text_data): Likewise.
13931 (pp_indent): Rename from pp_base_indent.
13932 (pp_format): Rename from pp_base_format.
13933 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
13934 (pp_format_verbatim): Rename from pp_base_format_verbatim.
13935 (pp_flush): Rename from pp_base_flush.
13936 (pp_set_line_maximum_length): Rename from
13937 pp_base_set_line_maximum_length.
13938 (pp_clear_output_area): Rename from pp_base_clear_output_area.
13939 (pp_set_prefix): Rename from pp_base_set_prefix.
13940 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
13941 (pp_emit_prefix): Rename from pp_base_emit_prefix.
13942 (pp_append_text): Rename from pp_base_append_text.
13943 (pp_formatted_text): Rename from pp_base_formatted_text.
13944 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
13945 (pp_remaining_character_count_for_line): Rename from
13946 pp_base_remaining_character_count_for_line.
13947 (pp_newline): Rename from pp_base_newline.
13948 (pp_character): Rename from pp_base_character.
13949 (pp_string): Rename from pp_base_string.
13950 (pp_maybe_space): Rename from pp_base_maybe_space.
13951 * asan.c (asan_pp_string): Adjust.
13952 (asan_emit_stack_protection): Likewise.
13953 (asan_add_global): Likewise.
13954 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
13955 * tree-mudflap.c (mf_varname_tree): Likewise.
13956 * tree-pretty-print.c (pp_tree_identifier): Rename from
13957 pp_base_tree_identifier.
13958 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
13959 Declare as function.
13961 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13963 * pretty-print.h (pp_bar_bar): New.
13964 (pp_ampersand_ampersand): Likewise.
13965 (pp_less_equal): Likewise.
13966 (pp_greater_equal): Likewise.
13967 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
13968 printer functions instead of pp_string or operators and punctuators.
13969 (dump_gimple_call): Likewise.
13970 (dump_gimple_omp_for): Likewise.
13971 (dump_gimple_transaction): Likewise.
13972 (dump_gimple_phi): Likewise.
13973 (pp_gimple_stmt_1): Likewise.
13974 * sched-vis.c (print_insn): Likewise.
13975 * tree-mudflap.c (mf_varname_tree): Likewise.
13976 * tree-pretty-print.c (dump_block_node): Likewise.
13977 (dump_generic_node): Likewise.
13979 2013-08-02 Jan Hubicka <jh@suse.cz>
13981 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
13983 * lto-streamer-out.c (tree_is_indexable): Results decls and
13984 parm decls are not indexable.
13985 (DFS_write_tree_body): Do not follow args and results.
13986 (hash_tree): Likewise.
13987 (output_functions): Rearrange so struct function is needed
13988 only when real body is output; be able to also ouptut abstract
13989 functions; output DECL_ARGUMENTS and DECL_RESULT.
13990 (lto_output): When not in WPA, ale store abstract functions.
13991 (write_symbol): Do not care about RESULT_DECL.
13992 (output_symbol_p): Handle correctly sbtract decls.
13993 * lto-streamer-in.c (input_function): Rearrange so struct
13994 function can be NULL at entry; allow streaming of
13995 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
13996 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
13997 sanity check during LTO.
13998 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
13999 RESULT_DECl and DECL_ARGUMENTS.
14000 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
14003 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14005 * pretty-print.h (pp_underscore): New.
14007 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
14008 printer functions instead of pp_character.
14009 (dump_binary_rhs): Likewise.
14010 (dump_ternary_rhs): Likewise.
14011 (dump_gimple_call_args): Likewise.
14012 (pp_points_to_solution): Likewise.
14013 (dump_gimple_call): Likewise.
14014 (dump_gimple_switch): Likewise.
14015 (dump_gimple_cond): Likewise.
14016 (dump_gimple_bind): Likewise.
14017 (dump_gimple_try): Likewise.
14018 (dump_gimple_omp_for): Likewise.
14019 (dump_gimple_omp_continue): Likewise.
14020 (dump_gimple_omp_single): Likewise.
14021 (dump_gimple_omp_sections): Likewise.
14022 (dump_gimple_omp_block): Likewise.
14023 (dump_gimple_omp_critical): Likewise.
14024 (dump_gimple_transaction): Likewise.
14025 (dump_gimple_asm): Likewise.
14026 (dump_gimple_phi): Likewise.
14027 (dump_gimple_omp_parallel): Likewise.
14028 (dump_gimple_omp_task): Likewise.
14029 (dump_gimple_omp_atomic_load): Likewise.
14030 (dump_gimple_omp_atomic_store): Likewise.
14031 (dump_gimple_mem_ops): Likewise.
14032 (pp_gimple_stmt_1): Likewise.
14033 (pp_cfg_jump): Likewise.
14034 (dump_implicit_edges): Likewise.
14035 (gimple_dump_bb_for_graph): Likewise.
14036 * graph.c (draw_cfg_node): Likewise.
14037 * langhooks.c (lhd_print_error_function): Likewise.
14038 * sched-vis.c (print_exp): Likewise.
14039 (print_value): Likewise.
14040 (print_pattern): Likewise.
14041 (print_insn): Likewise.
14042 (rtl_dump_bb_for_graph): Likewise.
14043 * tree-pretty-print.c (dump_function_declaration): Likewise.
14044 (dump_array_domain): Likewise.
14045 (dump_omp_clause): Likewise.
14046 (dump_location): Likewise.
14047 (dump_generic_node): Likewise.
14048 (print_struct_decl): Likewise.
14049 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14051 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14053 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14054 candidate table when replacing a candidate statement.
14055 (replace_rhs_if_not_dup): Likewise.
14056 (replace_one_candidate): Likewise.
14058 2013-08-02 Jan Hubicka <jh@suse.cz>
14059 Martin Liska <marxin.liska@gmail.com>
14061 * cgraphunit.c (add_new_function): Fix logic when adding from
14063 (assemble_thunk): Rename to ...
14064 (expand_thunk); .. this one; export; get it working with
14065 general functions; make produced gimple valid.
14066 * cgraph.h (expand_thunk): Declare.
14068 2013-08-02 Jan Hubicka <jh@suse.cz>
14070 * ipa-cp.c (gather_context_independent_values): Use
14071 ipa_get_param_move_cost.
14072 (get_replacement_map): Remove PARAM; move parameter folding
14074 (create_specialized_node): Update.
14075 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14076 assert that we have gimple body; update move_cost.
14077 (count_formal_params): Assert that we have gimple body.
14078 (ipa_dump_param): New function.
14079 (ipa_alloc_node_params): Break out from ...
14080 (ipa_initialize_node_params): ... here.
14081 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14082 (ipa_write_node_info): Stream move costs.
14083 (ipa_read_node_info): Read move costs.
14084 (ipa_update_after_lto_read): Do not recompute node params.
14085 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14086 (ipa_get_param): Check we are not in WPA.
14087 (ipa_get_param_move_cost): New.
14088 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14089 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14090 parm numbers to be present.
14092 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14094 PR rtl-optimization/58048
14095 * lra-constraints.c (process_alt_operands): Don't check asm
14096 operand on register.
14098 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14100 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14101 the implied StoreLoad barrier for atomic operations if before.
14103 2013-08-02 Jan Hubicka <jh@suse.cz>
14104 Martin Liska <marxin.liska@gmail.com>
14106 * cgraph.c (cgraph_function_body_availability): Do not check
14108 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14109 symtab_node_availability): Declare.
14110 * ipa.c (can_replace_by_local_alias): New.
14111 (function_and_variable_visibility): Use it.
14112 * symtab.c (symtab_for_node_and_aliases,
14113 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14115 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14117 PR rtl-optimization/57963
14118 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14119 (lra_constraints): Use them.
14121 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14123 * config/arm/types.md (define_attr "type"): Add "load_acq"
14125 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14127 (cortex_a53_store1): Likewise.
14129 2013-08-01 Jan Hubicka <jh@suse.cz>
14131 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14132 partitions are not needed.
14134 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14136 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14137 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14138 MAYBE_NON_Q_CLASS_P where appropriate.
14140 2013-08-01 Jan Hubicka <jh@suse.cz>
14142 * cgraph.h (release_function_body): Declare.
14143 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14144 of unused delcarations.
14146 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14148 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14149 RTL form when subtracting a constant.
14151 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14153 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14154 Generate canonical plus rtx with negated immediate instead of minus
14156 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14158 2013-08-01 Jan Hubicka <jh@suse.cz>
14160 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14161 (cgraph_release_function_body): Likewise.
14162 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
14163 * cgraph.h (cgrpah_node): Rename abstract_and_needed
14164 to used_as_abstract_origin.
14165 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
14166 symbols used as abstract origins.
14167 * cgraphunit.c (analyze_functions): Update.
14168 * ipa.c (symtab_remove_unreachable_nodes): Recompute
14169 used_as_abstract_origin.
14170 * tree-inline.c (tree_function_versioning): Update
14171 used_as_abstract_origin; be ready for DECL_RESULT and
14172 DECL_ARGUMENTS to be NULL.
14174 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
14175 for abstract functions.
14176 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
14179 2013-08-01 Jan Hubicka <jh@suse.cz>
14181 * profile.c (compute_value_histograms): Fix thinko.
14183 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14185 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
14186 aarch-common-protos.h to extra_headers.
14187 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14188 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
14189 * config/aarch64/t-aarch64 (aarch-common.o): Define.
14191 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14193 * config/aarch64/aarch64.md (define_attr "type"): Delete.
14194 Include "../arm/types.md". Define "type" attribute for all patterns.
14195 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
14198 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
14200 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
14201 to support power8 load fusion.
14202 (fusion_gpr_mem_load): Likewise.
14204 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
14206 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
14207 declarations for power8 load fusion.
14208 (emit_fusion_gpr_load): Likewise.
14210 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14211 tuning for power8, turn on fusion mode by default. Turn on sign
14212 extending fusion mode if normal fusion mode is on, and we are at
14214 (fusion_gpr_load_p): New function, return true if we can fuse an
14215 addis instruction with a dependent load to a GPR.
14216 (emit_fusion_gpr_load): Emit the instructions for power8 load
14219 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14220 (VSX load fusion peepholes): New peepholes to fuse together an
14221 addi instruction with a VSX load instruction.
14223 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14224 peepholes to fuse an addis instruction with a load to a GPR base
14225 register. If we are supporting sign extending fusions, convert
14226 sign extending loads to zero extending loads and add an explicit
14229 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14231 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14232 aarch-common-protos.h to extra_headers.
14233 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14234 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14235 (arm_early_store_addr_dep): Likewise.
14236 (arm_no_early_alu_shift_dep): Likewise.
14237 (arm_no_early_alu_shift_value_dep): Likewise.
14238 (arm_no_early_mul_dep): Likewise.
14239 (arm_no_early_store_addr_dep): Likewise.
14240 (arm_mac_accumulator_is_mul_result): Likewise.
14241 (arm_mac_accumulator_is_result): Likewise.
14242 * config/arm/aarch-common.c: ... here. New file.
14243 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14245 (arm_early_store_addr_dep): Likewise.
14246 (arm_no_early_alu_shift_dep): Likewise.
14247 (arm_no_early_alu_shift_value_dep): Likewise.
14248 (arm_no_early_mul_dep): Likewise.
14249 (arm_no_early_store_addr_dep): Likewise.
14250 (arm_mac_accumulator_is_mul_result): Likewise.
14251 (arm_mac_accumulator_is_result): Likewise.
14252 * config/arm/aarch-common-protos.h: ... here. New file.
14253 * config/arm/t-arm (aarch-common.o): Define.
14255 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14257 * config/arm/arm.md: Include new file "types.md".
14258 (define_attr "type"): Move from here to ...
14259 (define_attr "mul32"): Likewise.
14260 (define_attr "mul64"): Likewise.
14261 * config/arm/types.md: ... here. New file.
14263 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14265 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14266 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14268 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14270 * gen-pass-instances.awk: Fix offset of substr().
14272 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14274 * Makefile.in (pass-instances.def): New.
14275 (passes.o): Replace dependency on passes.def with one on
14278 * gen-pass-instances.awk: New.
14280 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14281 rather than passes.def, updating local definition of NEXT_PASS
14282 macro to add an extra NUM parameter (currently unused).
14284 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14286 * Makefile.in (PASS_MANAGER_H): New.
14287 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14288 (passes.o): Likewise.
14289 (statistics.o): Likewise.
14290 (cgraphunit.o): Likewise.
14291 (context.o): Depend on PASS_MANAGER_H.
14293 * pass_manager.h: New.
14295 * cgraphunit.c (cgraph_add_new_function): Update for moves
14296 of globals to fields of pass_manager.
14297 (analyze_function): Likewise.
14298 (expand_function): Likewise.
14299 (ipa_passes): Likewise.
14300 (compile): Likewise.
14302 * context.c (context::context): New.
14303 * context.h (context::context): New.
14304 (context::get_passes): New.
14305 (context::passes_): New.
14307 * lto-cgraph.c (input_node): Update for moves of globals to
14308 fields of pass_manager.
14310 * passes.c (all_passes): Remove, in favor of a field of the
14311 same name within the new class pass_manager.
14312 (all_small_ipa_passes): Likewise.
14313 (all_lowering_passes): Likewise.
14314 (all_regular_ipa_passes): Likewise.
14315 (all_late_ipa_passes): Likewise.
14316 (all_lto_gen_passes): Likewise.
14317 (passes_by_id): Likewise.
14318 (passes_by_id_size): Likewise.
14319 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14320 the new class pass_manager.
14321 (set_pass_for_id): Convert to...
14322 (pass_manager::set_pass_for_id): ...method.
14323 (get_pass_for_id): Convert to...
14324 (pass_manager::get_pass_for_id): ...method.
14325 (register_one_dump_file): Move body of implementation into...
14326 (pass_manager::register_one_dump_file): ...here.
14327 (register_dump_files_1): Convert to...
14328 (pass_manager::register_dump_files_1): ...method.
14329 (register_dump_files): Convert to...
14330 (pass_manager::register_dump_files): ...method.
14331 (create_pass_tab): Update for moves of globals to fields of
14333 (dump_passes): Move body of implementation into...
14334 (pass_manager::dump_passes): ...here.
14335 (register_pass): Move body of implementation into...
14336 (pass_manager::register_pass): ...here.
14337 (init_optimization_passes): Convert into...
14338 (pass_manager::pass_manager): ...constructor for new
14339 pass_manager class, and initialize the pass_lists array.
14340 (check_profile_consistency): Update for moves of globals to
14341 fields of pass_manager.
14342 (dump_profile_report): Move body of implementation into...
14343 (pass_manager::dump_profile_report): ...here.
14344 (ipa_write_summaries_1): Update for moves of pass lists from
14345 being globals to fields of pass_manager.
14346 (ipa_write_optimization_summaries): Likewise.
14347 (ipa_read_summaries): Likewise.
14348 (ipa_read_optimization_summaries): Likewise.
14349 (execute_all_ipa_stmt_fixups): Likewise.
14351 * statistics.c (statistics_fini): Update for moves of globals to
14352 fields of pass_manager.
14354 * toplev.c (general_init): Replace call to
14355 init_optimization_passes with construction of the pass_manager
14358 * tree-pass.h (all_passes): Remove, in favor of a field of the
14359 same name within the new class pass_manager.
14360 (all_small_ipa_passes): Likewise.
14361 (all_lowering_passes): Likewise.
14362 (all_regular_ipa_passes): Likewise.
14363 (all_lto_gen_passes): Likewise.
14364 (all_late_ipa_passes): Likewise.
14365 (passes_by_id): Likewise.
14366 (passes_by_id_size): Likewise.
14367 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14368 the new class pass_manager.
14369 (get_pass_for_id): Remove.
14371 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14373 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14376 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14378 * arm.md (mulhi3): New expand pattern.
14380 2013-07-30 Jan Hubicka <jh@suse.cz>
14381 Martin Liska <marxin.liska@gmail.com>
14383 * profile.c (compute_value_histograms): Do not ICE when
14384 there is mismatch only on some counters.
14386 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14388 PR rtl-optimization/57637
14389 * function.c (move_insn_for_shrink_wrap): Also check the
14390 GEN set of the LIVE problem for the liveness analysis
14391 if it exists, otherwise give up.
14393 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14395 PR tree-optimization/57993
14396 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14397 replaced statement in the candidate table.
14398 (phi_add_costs): Return infinite cost when the hidden basis does
14399 not dominate all phis on which the candidate is dependent.
14400 (replace_one_candidate): Record replaced statement in the
14403 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14405 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14406 (ashlv2si3): New expander.
14407 (*ashlv2si3_i): New define_insn_and_split.
14408 * predicates.md (float_operation): Allow patterns with three
14409 basic sub-patterns.
14411 PR rtl-optimization/58021
14412 * mode-switching.c (create_pre_exit): Always split off preceding
14413 insns if we are not at the basic block head.
14415 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14417 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14418 (UCLIBC_DYNAMIC_LINKER): New macro.
14419 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14421 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14422 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14424 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14425 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14426 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14427 for SF and DF modes. Use ieee_quad_format for TF mode.
14428 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14429 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14430 (mips_option_override): Handle `-mnan=legacy'.
14431 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14432 `-mabs=2008' and `-mnan=2008'.
14433 (OPTION_DEFAULT_SPECS): Add "nan" default.
14434 (ASM_SPEC): Handle `-mnan='.
14435 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14436 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14437 comment accordingly.
14438 (neg<mode>2): Likewise.
14439 * config/mips/mips.opt (mabs, mnan): New options.
14440 * doc/install.texi (Configuration): Document `--with-nan=' option.
14441 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14443 (MIPS Options): Document them.
14444 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14445 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14446 * configure: Regenerate.
14447 * config.in: Regenerate.
14449 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14451 * config/i386/i386.md (float post-reload splitters): Do not check
14452 for subregs of SSE registers.
14454 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14455 H.J. Lu <hongjiu.lu@intel.com>
14459 * config/i386/i386.md (post-reload splitter
14460 to avoid partial SSE reg dependency stalls): New pattern.
14462 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14464 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14466 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14468 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14469 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14471 2013-07-26 Cary Coutant <ccoutant@google.com>
14473 * dwarf2out.c (die_checksum_ordered): Don't include template
14474 instantiations in signature.
14475 (is_template_parameter): New function.
14476 (is_template_instantiation): New function.
14477 (generate_skeleton_bottom_up): Don't include template instantiations
14479 (generate_skeleton): Likewise.
14480 (break_out_comdat_types): Move recursive call to break out nested
14482 (prune_unused_types_mark_generic_parms_dies): Call
14483 is_template_parameter.
14485 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14487 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14488 uses vector registers.
14489 * config/aarch64/iterators.md: Add attributes rtn and vas.
14491 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14492 Richard Earnshaw <richard.earnshaw@arm.com>
14494 * combine.c (simplify_comparison): Re-canonicalize operands
14496 * config/arm/arm.md (movcond_addsi): New splitter.
14498 2013-07-25 Sterling Augustine <saugustine@google.com>
14500 * dwarf2out.c (size_of_pubnames): Move code to...
14501 (include_pubname_in_output): ...here. New.
14502 (want_pubnames): Rearrange.
14503 (output_pubnames): Call include_pubname_in_output. Move assertion.
14505 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14507 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14509 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14512 * doc/extend.texi: Remove obsolete builtins. Fix
14513 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14515 2013-07-25 Jan Hubicka <jh@suse.cz>
14517 * cgraph.c (release_function_body): Break out from ...
14518 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14519 and DECL_ARGUMENTS.
14520 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14521 old_tree in the map.
14522 (create_specialized_node): Update.
14523 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14525 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14526 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14527 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14528 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14531 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14533 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14534 addsi3_carryin_alt2_<optab>): Correct output template.
14536 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14538 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14539 Adjust for arm_restrict_it.
14540 Remove trailing whitespace.
14542 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14544 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14545 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14547 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14549 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14551 PR rtl-optimization/57960
14552 * lra-constraints.c (process_alt_operands): Use the right mode
14553 when checking strict_low.
14555 2013-07-25 Jan Hubicka <jh@suse.cz>
14557 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14558 * cgraph.c (cgraph_remove_node): Do not release function body
14559 when in cgraph streaming.
14560 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14561 in other partitions are not considered reachable; fix handling of
14564 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14566 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14568 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14573 * config/arm/arm.md ("*sibcall_insn): Replace use of
14574 Ss with US. Adjust output for v5 and v4t.
14575 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14577 * config/arm/constraints.md ("Ss"): Rename to US.
14579 2013-07-25 Terry Guo <terry.guo@arm.com>
14581 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14582 shift_add/shift_sub0/shift_sub1 RTXs.
14584 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14585 Anton Blanchard <anton@au1.ibm.com>
14587 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14588 (altivec_vpks<VI_char>ss): Likewise.
14589 (altivec_vpks<VI_char>us): Likewise.
14590 (altivec_vpku<VI_char>us): Likewise.
14591 (altivec_vpku<VI_char>um): Likewise.
14593 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14595 Introduce context class.
14597 * Makefile.in (CONTEXT_H): New.
14598 (OBJS): Add context.o.
14599 (toplev.o): Add CONTEXT_H to dependencies.
14602 * toplev.c (general_init): Create the singleton gcc::context instance.
14608 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14610 PR rtl-optimization/57968
14611 * mode-switching.c (create_pre_exit): Allow instructions that
14612 don't set a return register to need a non-exit mode.
14614 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14615 Anton Blanchard <anton@au1.ibm.com>
14617 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14618 operands to vperm for little endian.
14619 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14620 of lvsl to create the control mask for a vperm for little endian.
14622 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14623 Anton Blanchard <anton@au1.ibm.com>
14625 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14626 two operands for little-endian.
14628 2013-07-23 Steve Ellcey <sellcey@mips.com>
14630 * config/mips/mips.c (mips_case_values_threshold): New.
14631 (TARGET_CASE_VALUES_THRESHOLD): Define.
14633 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14634 Anton Blanchard <anton@au1.ibm.com>
14636 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
14637 selection of field for vector splat in little endian mode.
14639 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
14641 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
14642 expanders to rs6000.md.
14643 (ior<mode>3): Likewise.
14644 (and<mode>3): Likewise.
14645 (one_cmpl<mode>2): Likewise.
14646 (nor<mode>3): Likewise.
14647 (andc<mode>3): Likewise.
14648 (eqv<mode>3): Likewise.
14649 (nand<mode>3): Likewise.
14650 (orc<mode>3): Likewise.
14652 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
14655 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
14656 to split multi-word logical operations.
14657 (rs6000_split_logical_di): Likewise.
14658 (rs6000_split_logical): Likewise.
14660 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
14661 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
14662 and allow TImode operations in 32-bit.
14663 (vsx_and<mode>3_64bit): Likewise.
14664 (vsx_ior<mode>3_32bit): Likewise.
14665 (vsx_ior<mode>3_64bit): Likewise.
14666 (vsx_xor<mode>3_32bit): Likewise.
14667 (vsx_xor<mode>3_64bit): Likewise.
14668 (vsx_one_cmpl<mode>2_32bit): Likewise.
14669 (vsx_one_cmpl<mode>2_64bit): Likewise.
14670 (vsx_nor<mode>3_32bit): Likewise.
14671 (vsx_nor<mode>3_64bit): Likewise.
14672 (vsx_andc<mode>3_32bit): Likewise.
14673 (vsx_andc<mode>3_64bit): Likewise.
14674 (vsx_eqv<mode>3_32bit): Likewise.
14675 (vsx_eqv<mode>3_64bit): Likewise.
14676 (vsx_nand<mode>3_32bit): Likewise.
14677 (vsx_nand<mode>3_64bit): Likewise.
14678 (vsx_orc<mode>3_32bit): Likewise.
14679 (vsx_orc<mode>3_64bit): Likewise.
14681 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
14682 logical types in GPRs.
14684 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
14685 logical insns to rs6000.md, and allow TImode operations in
14687 (altivec_ior<mode>3): Likewise.
14688 (altivec_xor<mode>3): Likewise.
14689 (altivec_one_cmpl<mode>2): Likewise.
14690 (altivec_nor<mode>3): Likewise.
14691 (altivec_andc<mode>3): Likewise.
14693 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
14694 attributes for moving the 128-bit logical operations into
14696 (BOOL_REGS_OUTPUT): Likewise.
14697 (BOOL_REGS_OP1): Likewise.
14698 (BOOL_REGS_OP2): Likewise.
14699 (BOOL_REGS_UNARY): Likewise.
14700 (BOOL_REGS_AND_CR0): Likewise.
14701 (one_cmpl<mode>2): Add support for DI logical operations on
14702 32-bit, splitting the operations to 32-bit.
14703 (anddi3): Likewise.
14704 (iordi3): Likewise.
14705 (xordi3): Likewise.
14706 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
14707 changes to combine the 32/64-bit code, allow logical operations on
14708 TI mode in 32-bit, and to use similar match_operator patterns like
14709 scalar mode uses. Combine the Altivec and VSX code for logical
14710 operations, and move it here.
14711 (ior<mode>3, 128-bit types): Likewise.
14712 (xor<mode>3, 128-bit types): Likewise.
14713 (one_cmpl<mode>3, 128-bit types): Likewise.
14714 (nor<mode>3, 128-bit types): Likewise.
14715 (andc<mode>3, 128-bit types): Likewise.
14716 (eqv<mode>3, 128-bit types): Likewise.
14717 (nand<mode>3, 128-bit types): Likewise.
14718 (orc<mode>3, 128-bit types): Likewise.
14719 (and<mode>3_internal): Likewise.
14720 (bool<mode>3_internal): Likewise.
14721 (boolc<mode>3_internal1): Likewise.
14722 (boolc<mode>3_internal2): Likewise.
14723 (boolcc<mode>3_internal1): Likewise.
14724 (boolcc<mode>3_internal2): Likewise.
14725 (eqv<mode>3_internal1): Likewise.
14726 (eqv<mode>3_internal2): Likewise.
14727 (one_cmpl1<mode>3_internal): Likewise.
14729 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14731 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14732 Rename flag_stack_usage to flag_stack_usage_info.
14734 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14736 * config/microblaze/sync.md: New file.
14737 * config/microblaze/microblaze.md: Include sync.md
14738 * config/microblaze/microblaze.c: Add print_operand 'y'.
14739 * config/microblaze/constraints.md: Add memory_contraint
14740 'Q' which is a single register.
14742 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
14744 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
14746 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14748 * reload.c (find_reloads): Exit loop once we find this operand
14749 cannot be reloaded somehow for this alternative.
14751 * reload.c (find_reloads): Exit loop once we find a hard register.
14753 * rtlanal.c (computed_jump_p): Exit loop once we find label
14756 * i386.c (ix86_pad_returns): Exit loop after setting replace.
14758 * cfgloopmanip.c (remove_path): Exit loop after setting
14761 * gensupport.c (subst_dup): Avoid loop if code is not
14762 MATCH_DUP nor MATCH_OP_DUP.
14764 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
14766 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
14768 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14770 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
14771 true for SP_REGNUM if mode == ptr_mode.
14772 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
14773 with value R0_REGNUM + 31.
14775 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14777 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
14778 pad pointer-typed argument downward.
14780 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14782 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
14783 and __ILP32__ when the ILP32 model is in use.
14785 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14787 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
14788 (aarch64_load_symref_appropriately): In the case of
14789 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
14790 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
14791 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
14792 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
14793 if 'mode' doesn't equal to 'ptr_mode'.
14794 (aarch64_output_mi_thunk): Add an assertion on the alignment of
14795 'vcall_offset'; change to call aarch64_emit_move differently depending
14796 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
14797 to calculate the upper bound of 'vcall_offset'.
14798 (aarch64_cannot_force_const_mem): Change to also return true if
14800 (aarch64_legitimize_reload_address): In the case of large
14801 displacements, add new local variable 'xmode' and an assertion
14802 based on it; change to use 'xmode' to generate the new rtx and
14804 (aarch64_asm_trampoline_template): Change to generate the template
14805 differently depending on TARGET_ILP32 or not; change to use
14806 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
14807 (aarch64_trampoline_size): Removed.
14808 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
14809 and replace immediate literals with it. Change to use 'ptr_mode'
14810 instead of 'DImode' and call convert_memory_address if the mode
14811 of 'fnaddr' doesn't equal to 'ptr_mode'.
14812 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
14814 (aarch64_elf_asm_destructor): Likewise.
14815 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
14816 on TARGET_ILP32 instead of aarch64_trampoline_size.
14817 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14818 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14819 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14820 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14821 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14822 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14823 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14824 depending on the value of 'mode'.
14825 (add_losym_<mode>): New.
14826 (ldr_got_small_<mode>): New, based on ldr_got_small.
14827 (ldr_got_small): Remove.
14828 (ldr_got_small_sidi): New.
14829 * config/aarch64/iterators.md (P): New.
14830 (PTR): Change to 'ptr_mode' in the condition.
14832 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14834 * config.gcc (aarch64*-*-*): Support --with-abi.
14835 (aarch64*-*-elf): Support --with-multilib-list.
14836 (aarch64*-*-linux*): Likewise.
14837 (supported_defaults): Add abi to aarch64*-*-*.
14838 * configure.ac: Mention AArch64 for --with-multilib-list.
14839 * configure: Re-generated.
14840 * config/aarch64/biarchilp32.h: New file.
14841 * config/aarch64/biarchlp64.h: New file.
14842 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14844 (MULTILIB_DEFAULTS): Ditto.
14845 (DRIVER_SELF_SPECS): Ditto.
14846 (ASM_SPEC): Update to also substitute -mabi.
14847 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14848 file whose name depends on -mabi= and -mbig-endian.
14849 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14851 (POINTER_SIZE): New define.
14852 (POINTERS_EXTEND_UNSIGNED): Ditto.
14853 (enum aarch64_abi_type): New enumeration tag.
14854 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
14855 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
14856 (TARGET_ILP32): New define.
14857 * config/aarch64/aarch64.opt (mabi): New.
14858 (aarch64_abi): New.
14859 (ilp32, lp64): New values for -mabi.
14860 * config/aarch64/t-aarch64 (comma): New define.
14861 (MULTILIB_OPTIONS): Ditto.
14862 (MULTILIB_DIRNAMES): Ditto.
14863 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
14864 * doc/invoke.texi: Document -mabi for AArch64.
14866 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
14868 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
14870 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14871 Anton Blanchard <anton@au1.ibm.com>
14873 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
14874 endianness when selecting field to splat.
14876 2013-07-22 Eric Christopher <echristo@gmail.com>
14878 * dwarf2out.c (die_odr_checksum): New function to use
14879 CHECKSUM_ macros and ULEB128 for DIE tag.
14880 (generate_type_signature): Use.
14882 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
14884 * config.gcc (sparc*-*-*): Accept leon3 processor.
14885 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
14886 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
14887 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
14888 * config/sparc/sparc.opt (enum processor_type): Add leon3.
14889 (mfix-ut699): Adjust comment.
14890 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
14891 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
14892 (CPP_CPU_SPEC): Likewise.
14893 (ASM_CPU_SPEC): Likewise.
14894 * config/sparc/sparc.c (leon3_cost): New constant.
14895 (sparc_option_override): Add leon3 support.
14896 (mem_ref): New function.
14897 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
14898 (sparc_do_work_around_errata): Look into the instruction in the delay
14899 slot and adjust accordingly. Add fix for the data cache nullify issues
14900 of the UT699. Change insertion position for the NOP.
14901 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
14902 (leon3_load): New reservation.
14903 (leon_store): Bump latency to 2.
14904 (grfpu): New automaton.
14905 (grfpu_alu): New unit.
14906 (grfpu_ds): Likewise.
14907 (leon_fp_alu): Adjust.
14908 (leon_fp_mult): Delete.
14909 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
14910 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
14911 * config/sparc/sparc.md (cpu): Add leon3.
14912 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
14913 (swapsi): Likewise.
14914 (atomic_test_and_set): Likewise.
14915 (ldstub): Likewise.
14917 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
14919 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
14920 default for R5900 targets.
14921 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
14922 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
14923 * config/mips/mips.c (mips_option_override): Report an error for
14924 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
14925 for -march=r5900 -mhard-float.
14927 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14929 * df-problems.c (can_move_insns_across): Exit loop once we
14930 find a non-fixed, non-global register.
14932 * ipa-pure-const.c (propagate_nothrow): Exit loop after
14935 * omega.c (omega_eliminate_red): Break after setting red_found.
14936 (omega_problem_has_red_equations): Similarly after setting found.
14937 (omega_query_variable): Similarly after setting coupled.
14939 2013-07-22 Marek Polacek <polacek@redhat.com>
14941 * gimplify.c: Don't include gimple.h twice.
14943 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14945 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
14946 instead of TARGET_THUMB1.
14947 (Pz): New constraint.
14948 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
14950 (compare_negsi_si): Likewise.
14951 (compare_addsi2_op0): Likewise.
14952 (compare_addsi2_op1): Likewise.
14953 (addsi3_carryin_<optab>): Likewise.
14954 (addsi3_carryin_alt2_<optab>): Likewise.
14955 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
14956 for arm_restrict_it.
14957 (subsi3_carryin): Likewise.
14958 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
14959 (minmax_arithsi): Disable for arm_restrict_it.
14960 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
14961 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
14962 (satsi_<SAT:code>_shift): Likewise.
14963 (arm_shiftsi3): Add alternative for 16-bit encoding.
14964 (arm32_movhf): Disable for arm_restrict_it.
14965 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
14966 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
14968 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
14970 * config/arm/arm.md (attribute "insn"): Delete.
14971 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
14972 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
14973 (not_shiftsi): Update for attribute change.
14974 (not_shiftsi_compare0): Likewise.
14975 (not_shiftsi_compare0_scratch): Likewise.
14976 (arm_one_cmplsi2): Likewise.
14977 (thumb1_one_cmplsi2): Likewise.
14978 (notsi_compare0): Likewise.
14979 (notsi_compare0_scratch): Likewise.
14980 (thumb1_movdi_insn): Likewise.
14981 (arm_movsi_insn): Likewise.
14982 (movhi_insn_arch4): Likewise.
14983 (movhi_bytes): Likewise.
14984 (arm_movqi_insn): Likewise.
14985 (thumb1_movqi_insn): Likewise.
14986 (arm32_movhf): Likewise.
14987 (thumb1_movhf): Likewise.
14988 (arm_movsf_soft_insn): Likewise.
14989 (thumb1_movsf_insn): Likewise.
14990 (thumb_movdf_insn): Likewise.
14991 (movsicc_insn): Likewise.
14992 (movsfcc_soft_insn): Likewise.
14993 (and_scc): Likewise.
14994 (cond_move): Likewise.
14995 (if_move_not): Likewise.
14996 (if_not_move): Likewise.
14997 (if_shift_move): Likewise.
14998 (if_move_shift): Likewise.
14999 (if_shift_shift): Likewise.
15000 (if_not_arith): Likewise.
15001 (if_arith_not): Likewise.
15002 (cond_move_not): Likewise.
15003 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
15004 (neon_mov<mode>): Likewise.
15005 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
15006 (thumb2_movsi_vfp): Likewise.
15007 (movsf_vfp): Likewise.
15008 (thumb2_movsf_vfp): Likewise.
15009 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15011 (cortexa7_older_only): Likewise.
15012 (cortexa7_younger): Likewise.
15013 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
15014 (1020alu_shift_op): Likewise.
15015 (1020alu_shift_reg_op): Likewise.
15016 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15017 (alu_shift_op): Likewise.
15018 (alu_shift_reg_op): Likewise.
15019 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15020 (11_alu_shift_op): Likewise.
15021 (11_alu_shift_reg_op): Likewise.
15022 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15023 (9_alu_shift_reg_op): Likewise.
15024 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15026 (cortex_a15_alu_shift): Likewise.
15027 (cortex_a15_alu_shift_reg): Likewise.
15028 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15029 (cortex_a5_alu_shift): Likewise.
15030 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15032 (cortex_a53_alu_shift): Likewise.
15033 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15035 (cortex_a7_alu_reg): Likewise.
15036 (cortex_a7_alu_shift): Likewise.
15037 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15038 (cortex_a8_alu_shift): Likewise.
15039 (cortex_a8_alu_shift_reg): Likewise.
15040 (cortex_a8_mov): Likewise.
15041 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15042 (cortex_a9_dp_shift): Likewise.
15043 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15044 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15045 (cortex_r4_mov): Likewise.
15046 (cortex_r4_alu_shift): Likewise.
15047 (cortex_r4_alu_shift_reg): Likewise.
15048 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15049 (526_alu_shift_op): Likewise.
15050 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15051 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15052 (626te_alu_shift_op): Likewise.
15053 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15054 (726te_alu_op): Likewise.
15055 (726te_alu_shift_op): Likewise.
15056 (726te_alu_shift_reg_op): Likewise.
15057 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15058 (mp626_alu_shift_op): Likewise.
15059 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15060 (pj4_alu_e1_conds): Likewise.
15061 (pj4_alu): Likewise.
15062 (pj4_alu_conds): Likewise.
15063 (pj4_shift): Likewise.
15064 (pj4_shift_conds): Likewise.
15065 (pj4_alu_shift): Likewise.
15066 (pj4_alu_shift_conds): Likewise.
15068 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15070 * config/arm/predicates.md (shiftable_operator_strict_it):
15072 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15073 Disable cond_exec version for arm_restrict_it.
15074 (thumb2_smaxsi3): Convert to generate cond_exec.
15075 (thumb2_sminsi3): Likewise.
15076 (thumb32_umaxsi3): Likewise.
15077 (thumb2_uminsi3): Likewise.
15078 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15079 (thumb2_neg_abssi2): Likewise.
15080 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15081 (thumb2_movsicc_insn): Adjust alternatives.
15082 (thumb2_mov_negscc): Disable for arm_restrict_it.
15083 (thumb2_mov_negscc_strict_it): New pattern.
15084 (thumb2_mov_notscc_strict_it): New pattern.
15085 (thumb2_mov_notscc): Disable for arm_restrict_it.
15086 (thumb2_ior_scc): Likewise.
15087 (thumb2_ior_scc_strict_it): New pattern.
15088 (thumb2_cond_move): Adjust for arm_restrict_it.
15089 (thumb2_cond_arith): Disable for arm_restrict_it.
15090 (thumb2_cond_arith_strict_it): New pattern.
15091 (thumb2_cond_sub): Adjust for arm_restrict_it.
15092 (thumb2_movcond): Likewise.
15093 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15094 (thumb2_zero_extendhisi2_v6): Likewise.
15095 (thumb2_zero_extendqisi2_v6): Likewise.
15096 (orsi_notsi_si): Likewise.
15097 (orsi_not_shiftsi_si): Likewise.
15099 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15101 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15102 instruction sequence is 1 byte shorter.
15104 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15106 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15107 it is not needed after split.
15109 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15112 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15113 second label for nonlocal goto receivers. Don't output pic base labels
15114 unless we're producing PIC; mark that action unreachable().
15115 (ix86_save_reg): If the function contains a nonlocal label, save the
15117 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15118 * config/darwin.c (emitted_pic_label_num): New GTY.
15119 (update_pic_label_number_if_needed): New.
15120 (machopic_output_function_base_name): Adjust for nonlocal receiver
15122 (machopic_should_output_picbase_label): New.
15123 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15124 (nonlocal_goto_receiver): New insn and split.
15126 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15128 * config/aarch64/aarch64-builtins.c
15129 (aarch64_fold_builtin): Fold abs in all modes.
15130 * config/aarch64/aarch64-simd-builtins.def
15131 (abs): Enable for all modes.
15132 * config/aarch64/arm_neon.h
15133 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15134 (vabs_f64): Add missing intrinsic.
15136 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15138 * config/aarch64/arm_neon.h (vabs_s64): New function
15140 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15143 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15144 * config/avr/avr.md (adjust_len): Add `round'.
15145 * config/avr/avr-protos.h (avr_out_round): New prototype.
15146 (avr_out_plus): Add `out_label' argument.
15147 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15148 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15149 Handle the case where `insn' is just a pattern.
15150 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15151 (avr_out_round): New function.
15152 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15154 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15156 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15157 Add check for flag_stack_usage to handle -fstack-usage support
15159 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15161 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
15162 interaction for new Power8 flags and VSX.
15164 2013-07-18 Sriraman Tallam <tmsriram@google.com>
15166 PR middle-end/57698
15167 * tree-inline.c (expand_call_inline): Emit errors during
15168 early_inlining only if optimization is not turned on.
15170 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15174 * passes.c (init_optimization_passes): Move the construction of
15175 the pass hierarchy into a new passes.def file.
15177 * Makefile.in (passes.o): Add dependency on passes.def.
15179 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15181 * passes.c (init_optimization_passes): Introduce macros for
15182 constructing the tree of passes (INSERT_PASSES_AFTER,
15183 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
15184 TERMINATE_PASS_LIST).
15186 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
15187 Wei Mi <wmi@google.com>
15189 PR rtl-optimization/57878
15190 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
15192 (reload_pseudo_compare_func): Check nregs first for reload
15195 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15197 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
15199 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
15201 * read-rtl.c (validate_const_int): Once an invalid character is
15202 seen, quit the loop.
15204 * gengtype.c (write_roots): Similarly once we find the "deletable"
15205 or "if_marked" option.
15207 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15209 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15210 "xtab" and "sat". Move value "clz" from here to ...
15211 (attriubte "type"): ... here.
15212 (satsi_<SAT:code>): Delete "insn" attribute.
15213 (satsi_<SAT:code>_shift): Likewise.
15214 (arm_zero_extendqisi2addsi): Likewise.
15215 (arm_extendqisi2addsi): Likewise.
15216 (clzsi2): Update for attribute changes.
15217 (rbitsi2): Likewise.
15218 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15220 (arm_usatsihi): Likewise.
15221 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15223 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15225 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15226 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15227 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15228 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15229 in alphabetical order.
15230 (attribute "core_cycles"): Update for attribute changes.
15231 (arm_addsi3): Likewise.
15232 (addsi3_compare0): Likewise.
15233 (addsi3_compare0_scratch): Likewise.
15234 (addsi3_compare_op1): Likewise.
15235 (addsi3_compare_op2): Likewise.
15236 (compare_addsi2_op0): Likewise.
15237 (compare_addsi2_op1): Likewise.
15238 (addsi3_carryin_shift_<optab>): Likewise.
15239 (subsi3_carryin_shift): Likewise.
15240 (rsbsi3_carryin_shift): Likewise.
15241 (arm_subsi3_insn): Likewise.
15242 (subsi3_compare0): Likewise.
15243 (subsi3_compare): Likewise.
15244 (arm_andsi3_insn): Likewise.
15245 (thumb1_andsi3_insn): Likewise.
15246 (andsi3_compare0): Likewise.
15247 (andsi3_compare0_scratch): Likewise.
15248 (zeroextractsi_compare0_scratch
15249 (andsi_not_shiftsi_si): Likewise.
15250 (iorsi3_insn): Likewise.
15251 (iorsi3_compare0): Likewise.
15252 (iorsi3_compare0_scratch): Likewise.
15253 (arm_xorsi3): Likewise.
15254 (thumb1_xorsi3_insn): Likewise.
15255 (xorsi3_compare0): Likewise.
15256 (xorsi3_compare0_scratch): Likewise.
15257 (satsi_<SAT:code>_shift): Likewise.
15259 (arm_shiftsi3): Likewise.
15260 (shiftsi3_compare0): Likewise.
15261 (not_shiftsi): Likewise.
15262 (not_shiftsi_compare0): Likewise.
15263 (not_shiftsi_compare0_scratch): Likewise.
15264 (arm_one_cmplsi2): Likewise.
15265 (thumb_one_complsi2): Likewise.
15266 (notsi_compare0): Likewise.
15267 (notsi_compare0_scratch): Likewise.
15268 (thumb1_zero_extendhisi2): Likewise.
15269 (arm_zero_extendhisi2): Likewise.
15270 (arm_zero_extendhisi2_v6): Likewise.
15271 (arm_zero_extendhisi2addsi): Likewise.
15272 (thumb1_zero_extendqisi2): Likewise.
15273 (thumb1_zero_extendqisi2_v6): Likewise.
15274 (arm_zero_extendqisi2): Likewise.
15275 (arm_zero_extendqisi2_v6): Likewise.
15276 (arm_zero_extendqisi2addsi): Likewise.
15277 (thumb1_extendhisi2): Likewise.
15278 (arm_extendhisi2): Likewise.
15279 (arm_extendhisi2_v6): Likewise.
15280 (arm_extendqisi): Likewise.
15281 (arm_extendqisi_v6): Likewise.
15282 (arm_extendqisi2addsi): Likewise.
15283 (thumb1_extendqisi2): Likewise.
15284 (thumb1_movdi_insn): Likewise.
15285 (arm_movsi_insn): Likewise.
15286 (movsi_compare0): Likewise.
15287 (movhi_insn_arch4): Likewise.
15288 (movhi_bytes): Likewise.
15289 (arm_movqi_insn): Likewise.
15290 (thumb1_movqi_insn): Likewise.
15291 (arm32_movhf): Likewise.
15292 (thumb1_movhf): Likewise.
15293 (arm_movsf_soft_insn): Likewise.
15294 (thumb1_movsf_insn): Likewise.
15295 (movdf_soft_insn): Likewise.
15296 (thumb_movdf_insn): Likewise.
15297 (arm_cmpsi_insn): Likewise.
15298 (cmpsi_shiftsi): Likewise.
15299 (cmpsi_shiftsi_swp): Likewise.
15300 (arm_cmpsi_negshiftsi_si): Likewise.
15301 (movsicc_insn): Likewise.
15302 (movsfcc_soft_insn): Likewise.
15303 (arith_shiftsi): Likewise.
15304 (arith_shiftsi_compare0
15305 (arith_shiftsi_compare0_scratch
15306 (sub_shiftsi): Likewise.
15307 (sub_shiftsi_compare0
15308 (sub_shiftsi_compare0_scratch
15309 (and_scc): Likewise.
15310 (cond_move): Likewise.
15311 (if_plus_move): Likewise.
15312 (if_move_plus): Likewise.
15313 (if_move_not): Likewise.
15314 (if_not_move): Likewise.
15315 (if_shift_move): Likewise.
15316 (if_move_shift): Likewise.
15317 (if_shift_shift): Likewise.
15318 (if_not_arith): Likewise.
15319 (if_arith_not): Likewise.
15320 (cond_move_not): Likewise.
15321 (thumb1_ashlsi3): Set type attribute.
15322 (thumb1_ashrsi3): Likewise.
15323 (thumb1_lshrsi3): Likewise.
15324 (thumb1_rotrsi3): Likewise.
15325 (shiftsi3_compare0_scratch): Likewise.
15326 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15327 (neon_mov<mode>): Likewise.
15328 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15330 (thumb2_movsi_insn): Likewise.
15331 (thumb2_cmpsi_neg_shiftsi): Likewise.
15332 (thumb2_extendqisi_v6): Likewise.
15333 (thumb2_zero_extendhisi2_v6): Likewise.
15334 (thumb2_zero_extendqisi2_v6): Likewise.
15335 (thumb2_shiftsi3_short): Likewise.
15336 (thumb2_addsi3_compare0_scratch): Likewise.
15337 (orsi_not_shiftsi_si): Likewise.
15338 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15339 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15341 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15342 (1020alu_shift_op): Likewise.
15343 (1020alu_shift_reg_op): Likewise.
15344 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15345 (alu_shift_op): Likewise.
15346 (alu_shift_reg_op): Likewise.
15347 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15348 (11_alu_shift_op): Likewise.
15349 (11_alu_shift_reg_op): Likewise.
15350 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15351 (9_alu_shift_reg_op): Likewise.
15352 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15354 (cortex_a15_alu_shift): Likewise.
15355 (cortex_a15_alu_shift_reg): Likewise.
15356 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15358 (cortex_a5_alu_shift): Likewise.
15359 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15361 (cortex_a53_alu_shift): Likewise.
15362 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15364 (cortex_a7_alu_reg): Likewise.
15365 (cortex_a7_alu_shift): Likewise.
15366 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15368 (cortex_a8_alu_shift): Likewise.
15369 (cortex_a8_alu_shift_reg): Likewise.
15370 (cortex_a8_mov): Likewise.
15371 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15372 (cortex_a9_dp_shift): Likewise.
15373 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15375 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15377 (cortex_r4_mov): Likewise.
15378 (cortex_r4_alu_shift): Likewise.
15379 (cortex_r4_alu_shift_reg): Likewise.
15380 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15381 (526_alu_shift_op): Likewise.
15382 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15383 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15384 (626te_alu_shift_op): Likewise.
15385 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15386 (726te_alu_op): Likewise.
15387 (726te_alu_shift_op): Likewise.
15388 (726te_alu_shift_reg_op): Likewise.
15389 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15390 (mp626_alu_shift_op): Likewise.
15391 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15392 (pj4_alu_e1_conds): Likewise.
15393 (pj4_alu): Likewise.
15394 (pj4_alu_conds): Likewise.
15395 (pj4_shift): Likewise.
15396 (pj4_shift_conds): Likewise.
15397 (pj4_alu_shift): Likewise.
15398 (pj4_alu_shift_conds): Likewise.
15399 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15401 (cortexa7_older_only): Likewise.
15402 (cortexa7_younger): Likewise.
15404 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15406 * ipa-pure-const.c (generate_summary): Rename to...
15407 (pure_const_generate_summary): ... this.
15409 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15411 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15413 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15416 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15419 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15421 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15422 enabled without -march=zEC12.
15423 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15426 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15428 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15429 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15430 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15431 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15432 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15433 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15434 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15436 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15437 and ISA_HAS_NMADD3_NMSUB3.
15438 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15439 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15440 (nmsub4<mode>, nmsub3<mode>): Likewise.
15441 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15443 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15445 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15446 TARGET_MIPS5400 checking.
15448 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15449 Peter Bergner <bergner@vnet.ibm.com>
15451 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15452 registers in the comment.
15453 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15454 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15455 rather than FIRST_PSEUDO_REGISTERS.
15457 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15459 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15460 enable extra ISA flags with TARGET_HTM.
15462 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15464 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15467 2013-07-15 Cong Hou <congh@google.com>
15469 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15470 in compare function for sorting.
15472 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15474 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15475 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15476 * config/rs6000/rs6000.opt: Add -mhtm option.
15477 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15478 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15479 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15480 __HTM__ if the HTM instructions are available.
15481 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15482 htm_spr_reg_operand): New define_predicates.
15483 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15484 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15486 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15487 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15488 HTM builtin functions.
15489 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15490 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15491 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15492 (rs6000_builtin_mask_calculate): Likewise.
15493 (rs6000_option_override_internal): Likewise.
15494 (bdesc_htm): Add new HTM builtin support.
15495 (htm_spr_num): New function.
15496 (htm_spr_regno): Likewise.
15497 (rs6000_htm_spr_icode): Likewise.
15498 (htm_expand_builtin): Likewise.
15499 (htm_init_builtins): Likewise.
15500 (rs6000_expand_builtin): Add support for HTM builtin functions.
15501 (rs6000_init_builtins): Likewise.
15502 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15504 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15505 (TARGET_HTM, MASK_HTM): Define macros.
15506 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15507 (FIXED_REGISTERS): Likewise.
15508 (CALL_USED_REGISTERS): Likewise.
15509 (CALL_REALLY_USED_REGISTERS): Likewise.
15510 (REG_ALLOC_ORDER): Likewise.
15511 (enum reg_class): Likewise.
15512 (REG_CLASS_NAMES): Likewise.
15513 (REG_CLASS_CONTENTS): Likewise.
15514 (REGISTER_NAMES): Likewise.
15515 (ADDITIONAL_REGISTER_NAMES): Likewise.
15516 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15517 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15518 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15519 * config/rs6000/htm.md: New file.
15520 * config/rs6000/htmintrin.h: New file.
15521 * config/rs6000/htmxlintrin.h: New file.
15523 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15525 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15526 Define SYMBOL_TINY_GOT, update comment.
15527 * config/aarch64/aarch64.c
15528 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15529 (aarch64_expand_mov_immediate): Likewise.
15530 (aarch64_print_operand): Likewise.
15531 (aarch64_classify_symbol): Likewise.
15532 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15533 (ldr_got_tiny): Define.
15535 2013-07-13 Tobias Grosser <tobias@grosser.es>
15537 PR tree-optimization/54094
15538 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15539 scheduling dimension for the parallelism check from the polyhedral
15540 information in the AST.
15541 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15544 2013-07-13 Jason Merrill <jason@redhat.com>
15546 * print-tree.c (debug_vec_tree): Use debug_raw.
15547 (debug_raw (vec<tree, va_gc> &)): New.
15548 (debug_raw (vec<tree, va_gc> *)): New.
15549 * tree.h: Declare them.
15551 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15553 * ifcvt.c (ifcvt_after_combine): New static variable.
15554 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15556 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15557 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15558 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15560 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15562 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15564 2013-07-12 Michael Matz <matz@suse.de>
15566 PR middle-end/55771
15567 * convert.c (convert_to_real): Reject non-float inner types.
15569 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15571 * config/aarch64/aarch64-protos.h
15572 (aarch64_simd_immediate_valid_for_move): Remove.
15573 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15574 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15576 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15578 2013-07-11 Steve Ellcey <sellcey@mips.com>
15580 * config/mips/mips.c (mips_conditional_register_usage): Do not
15581 use t[0-7] registers in MIPS16 mode when optimizing for size.
15583 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15585 * config/i386/i386.c (dispatch_function_versions): Fix array
15586 indexing of function_version_info to match actual_versions.
15588 2013-07-11 Teresa Johnson <tejohnson@google.com>
15590 * vec.h (struct va_gc): Move release out-of-line.
15591 (va_gc::release): Call ggc_free on released vec.
15593 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15595 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15596 Require GOT register as additional operand in UNSPEC.
15597 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15598 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15599 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15600 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15601 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15602 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15603 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15605 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15608 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15609 name seen by assembler/linker rather if available.
15611 2013-07-11 Andreas Schwab <schwab@suse.de>
15613 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15615 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15617 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15619 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15621 * doc/tm.texi.in: Move hook documentation to ....
15622 * target.def: ... here.
15624 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15625 text on @hook line.
15626 * doc/tm.texi: Regenerate.
15628 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
15631 * doc/invoke.texi: Document -Wconditionally-supported.
15633 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15636 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
15639 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15642 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
15643 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
15644 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
15645 Remove duplicate devices.
15646 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
15647 * config/avr/t-multilib: Regenerate.
15648 * config/avr/avr-tables.opt: Regenerate.
15649 * doc/avr-mmcu.texi: Regenerate.
15651 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15654 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
15656 2013-07-10 Graham Stott <graham.stott@btinternet.com>
15658 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
15659 the cost of MULT when optimizing for size.
15661 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15663 * config/cr16/cr16-protos.h: Don't include target.h.
15665 2013-07-09 Joseph Myers <joseph@codesourcery.com>
15667 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
15668 adjust register size for TDmode and TFmode for VSX registers.
15670 2013-07-08 Kai Tietz <ktietz@redhat.com>
15673 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
15674 hook_bool_const_tree_true.
15676 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15678 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
15679 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
15680 * config/s390/s390.md: Define FPR*_REGNUM constants.
15681 Fix FPR2_REGNUM constant (18 -> 17).
15682 ("*trunc<BFP:mode><DFP_ALL:mode>2")
15683 ("*trunc<DFP_ALL:mode><BFP:mode>2")
15684 ("trunc<BFP:mode><DFP_ALL:mode>2")
15685 ("trunc<DFP_ALL:mode><BFP:mode>2")
15686 ("*extend<BFP:mode><DFP_ALL:mode>2")
15687 ("*extend<DFP_ALL:mode><BFP:mode>2")
15688 ("extend<BFP:mode><DFP_ALL:mode>2")
15689 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
15692 2013-07-08 Graham Stott <graham.stott@btinternet.com>
15694 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
15696 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15698 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
15699 and cfun_fpr_bit_p to cfun_fpr_save_p.
15700 (s390_frame_area, s390_register_info, s390_frame_info)
15701 (s390_emit_prologue, s390_emit_epilogue)
15702 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
15704 * config/s390/s390.h: Define *_REGNUM macros for floating point
15707 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
15709 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
15711 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
15713 PR rtl-optimization/57786
15714 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
15715 and break out of the loop when it is set to false.
15717 2013-07-08 Jakub Jelinek <jakub@redhat.com>
15720 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
15721 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
15722 (const_int 63)) 0)).
15723 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
15724 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
15725 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
15727 PR rtl-optimization/57829
15728 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
15729 mask bits outside of mode are just sign-extension from mode to HWI.
15731 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
15733 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
15734 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
15735 adjust_address instead of change_address to keep info about alignment.
15736 (emit_strmov): Remove.
15737 (emit_memmov): New function.
15738 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
15739 (expand_movmem_epilogue): Likewise and return updated rtx for
15741 (expand_constant_movmem_prologue): Likewise and return updated rtx for
15742 destination and source.
15743 (decide_alignment): Refactor, handle vector_loop.
15744 (ix86_expand_movmem): Likewise.
15745 (ix86_expand_setmem): Likewise.
15746 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
15748 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
15750 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
15751 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
15753 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15755 * config/i386/sse.md (sse_movlhps): Change alternative 3
15756 of operand 2 to "m".
15758 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15761 * config/i386/sse.md (iptr): New mode attribute.
15762 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
15763 (<sse>_vm<plusminus_insn><mode>3): Ditto.
15764 (<sse>_vmmul<mode>3): Ditto.
15765 (<sse>_vmdiv<mode>3): Ditto.
15766 (sse_vmrcpv4sf2): Ditto.
15767 (<sse>_vmsqrt<mode>2): Ditto.
15768 (sse_vmrsqrtv4sf2): Ditto.
15769 (<sse>_vm<code><mode>3): Ditto.
15770 (avx_vmcmp<mode>3): Ditto.
15771 (<sse>_vmmaskcmp<mode>3): Ditto.
15772 (<sse>_comi): Ditto.
15773 (<sse>_ucomi): Ditto.
15774 (*xop_vmfrcz_<mode>): Ditto.
15775 (*fmai_fmadd_<mode>): Ditto.
15776 (*fmai_fmsub_<mode>): Ditto.
15777 (*fmai_fnmadd_<mode>): Ditto.
15778 (*fmai_fnmsub_<mode>): Ditto.
15779 (*fma4i_vmfmadd_<mode>): Ditto.
15780 (*fma4i_vmfmsub_<mode>): Ditto.
15781 (*fma4i_vmfnmadd_<mode>): Ditto.
15782 (*fma4i_vmfnmsub_<mode>): Ditto.
15783 (*xop_vmfrcz_<mode>): Ditto.
15784 (sse_cvtps2pi): Ditto.
15785 (sse_cvttps2pi): Ditto.
15786 (sse_cvtss2si): Ditto.
15787 (sse_cvtss2si_2): Ditto.
15788 (sse_cvtss2siq_2): Ditto.
15789 (sse_cvttss2si): Ditto.
15790 (sse_cvttss2siq): Ditto.
15791 (sse_cvtsd2si): Ditto.
15792 (sse_cvtsd2si_2): Ditto.
15793 (sse_cvtsd2siq_2): Ditto.
15794 (sse_cvttsd2si): Ditto.
15795 (sse_cvttsd2siq): Ditto.
15796 (sse_cvtsd2ss): Ditto.
15797 (sse_cvtss2sd): Ditto.
15798 (avx2_pbroadcast<mode>): Ditto.
15799 (avx2_pbroadcast<mode>_1): Ditto.
15800 (*avx_vperm_broadcast_v4sf): Ditto.
15802 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
15803 (sse_movlhps): Ditto.
15804 (sse_storehps): Ditto.
15805 (sse_loadhps): Ditto.
15806 (sse_storelps): Ditto.
15807 (sse_loadlps): Ditto.
15808 (*vec_concatv4sf): Ditto.
15809 (*vec_interleave_highv2df): Ditto.
15810 (*vec_interleave_lowv2df): Ditto.
15811 (*vec_extractv2df_1_sse): Ditto.
15812 (*vec_extractv2df_0_sse): Ditto.
15813 (sse2_storelpd): Ditto.
15814 (sse2_loadlpd): Ditto.
15815 (sse2_movsd): Ditto.
15816 (*vec_concatv4si): Ditto.
15817 (vec_concatv2di): Ditto.
15819 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15820 for Intel asm dialect.
15821 (mmx_punpcklwd): Ditto.
15822 (mmx_punpckldq): Ditto.
15824 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15825 for intel assembler dialect.
15827 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15830 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15831 for BUILT_IN_C{LZ,LRSB}*.
15832 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15833 * tree-vrp.c (extract_range_basic): Handle
15834 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15835 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15836 fall thru to code calling set_value*.
15837 * builtins.c (expand_builtin): Remove *IMAX cases.
15838 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15839 if width is bigger than 2*HWI.
15841 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15843 PR rtl-optimization/55342
15844 * lra-int.h (lra_subreg_reload_pseudos): New.
15845 * lra.c: Add undoing optional reloads to the block diagram.
15846 (lra_subreg_reload_pseudos): New.
15847 (lra_optional_reload_pseudos): Change comments.
15848 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15849 lra_optional_reload_pseudos after undo transformations.
15850 * lra-assigns.c (pseudo_prefix_title): New.
15851 (lra_setup_reg_renumber): Use it.
15852 (spill_for): Ditto. Check subreg reload pseudos too.
15853 (assign_by_spills): Consider subreg reload pseudos too.
15854 * lra-constraints.c (simplify_operand_subreg): Use
15855 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
15856 (curr_insn_transform): Recognize and do optional reloads.
15857 (undo_optional_reloads): New.
15858 (lra_undo_inheritance): Call undo_optional_reloads.
15860 2013-07-05 Thomas Quinot <quinot@adacore.com>
15862 * tree-complex.c (expand_complex_operations_1): Fix typo.
15864 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
15866 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
15867 (tune_params): New member 'const vec_costs'.
15868 * config/aarch64/aarch64.c (generic_vector_cost): New.
15869 (generic_tunings): New member 'generic_vector_cost'.
15870 (aarch64_builtin_vectorization_cost): New.
15871 (aarch64_add_stmt_cost): New.
15872 (TARGET_VECTORIZE_ADD_STMT_COST): New.
15873 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
15875 2013-07-03 Jakub Jelinek <jakub@redhat.com>
15878 * config/i386/predicates.md (vsib_address_operand): Disallow
15879 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
15881 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
15883 PR middle-end/55030
15884 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
15885 expand_builtin_setjmp_receiver.
15886 (expand_label): Adjust, call expand_builtin_setjmp_receiver
15887 with NULL for the label parameter.
15888 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
15889 the frame-pointer. Adjust comments.
15890 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
15891 only if LABEL is non-NULL.
15893 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
15895 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
15896 (ARM_ABI_AAPCS64): Ditto.
15898 (ARM_DEFAULT_ABI): Ditto.
15900 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
15902 * config/aarch64/aarch64-builtins.c
15903 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
15904 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
15906 * config/aarch64/aarch64-simd.md
15907 (aarch64_ld1<VALL:mode>): New.
15908 (aarch64_st1<VALL:mode>): Likewise.
15909 * config/aarch64/arm_neon.h
15910 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
15912 2013-07-02 Sriraman Tallam <tmsriram@google.com>
15914 * config/i386/i386.c (gate_insert_vzeroupper): Check if
15916 (ix86_option_override_internal):Turn on all -mavx target flags by
15917 default as they are dependent on AVX anyway.
15919 2013-07-02 Cary Coutant <ccoutant@google.com>
15921 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
15922 deterministic hash.
15923 (loc_checksum_ordered): Likewise.
15924 (hash_loc_operands): Remove inline keyword.
15926 2013-07-02 Jakub Jelinek <jakub@redhat.com>
15928 PR tree-optimization/57741
15929 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
15930 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
15931 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
15932 Allow REAL_CST step_exprs if flag_associative_math.
15933 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
15935 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15937 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
15939 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15941 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
15943 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15945 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
15947 (iorsi3_insn): Likewise.
15948 (arm_xorsi3): Likewise.
15950 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
15952 * arm.md (attribute "wtype"): Delete. Move attribute values from here
15954 (attribute "type"): ... here, and prefix with "wmmx_".
15955 (attribute "core_cycles"): Update for attribute changes.
15956 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
15957 (tbcstv4hi): Likewise.
15958 (tbcstv2si): Likewise.
15959 (iwmmxt_iordi3): Likewise.
15960 (iwmmxt_xordi3): Likewise.
15961 (iwmmxt_anddi3): Likewise.
15962 (iwmmxt_nanddi3): Likewise.
15963 (iwmmxt_arm_movdi): Likewise.
15964 (iwmmxt_movsi_insn): Likewise.
15965 (mov<mode>_internal): Likewise.
15966 (and<mode>3_iwmmxt): Likewise.
15967 (ior<mode>3_iwmmxt): Likewise.
15968 (xor<mode>3_iwmmxt): Likewise.
15969 (add<mode>3_iwmmxt): Likewise.
15970 (ssaddv8qi3): Likewise.
15971 (ssaddv4hi3): Likewise.
15972 (ssaddv2si3): Likewise.
15973 (usaddv8qi3): Likewise.
15974 (usaddv4hi3): Likewise.
15975 (usaddv2si3): Likewise.
15976 (sub<mode>3_iwmmxt): Likewise.
15977 (sssubv8qi3): Likewise.
15978 (sssubv4hi3): Likewise.
15979 (sssubv2si3): Likewise.
15980 (ussubv8qi3): Likewise.
15981 (ussubv4hi3): Likewise.
15982 (ussubv2si3): Likewise.
15983 (mulv4hi3_iwmmxt): Likewise.
15984 (smulv4hi3_highpart): Likewise.
15985 (umulv4hi3_highpart): Likewise.
15986 (iwmmxt_wmacs): Likewise.
15987 (iwmmxt_wmacsz): Likewise.
15988 (iwmmxt_wmacu): Likewise.
15989 (iwmmxt_wmacuz): Likewise.
15990 (iwmmxt_clrdi): Likewise.
15991 (iwmmxt_clrv8qi): Likewise.
15992 (iwmmxt_clr4hi): Likewise.
15993 (iwmmxt_clr2si): Likewise.
15994 (iwmmxt_uavgrndv8qi3): Likewise.
15995 (iwmmxt_uavgrndv4hi3): Likewise.
15996 (iwmmxt_uavgv8qi3): Likewise.
15997 (iwmmxt_uavgv4hi3): Likewise.
15998 (iwmmxt_tinsrb): Likewise.
15999 (iwmmxt_tinsrh): Likewise.
16000 (iwmmxt_tinsrw): Likewise.
16001 (iwmmxt_textrmub): Likewise.
16002 (iwmmxt_textrmsb): Likewise.
16003 (iwmmxt_textrmuh): Likewise.
16004 (iwmmxt_textrmsh): Likewise.
16005 (iwmmxt_textrmw): Likewise.
16006 (iwmxxt_wshufh): Likewise.
16007 (eqv8qi3): Likewise.
16008 (eqv4hi3): Likewise.
16009 (eqv2si3): Likewise.
16010 (gtuv8qi3): Likewise.
16011 (gtuv4hi3): Likewise.
16012 (gtuv2si3): Likewise.
16013 (gtv8qi3): Likewise.
16014 (gtv4hi3): Likewise.
16015 (gtv2si3): Likewise.
16016 (smax<mode>3_iwmmxt): Likewise.
16017 (umax<mode>3_iwmmxt): Likewise.
16018 (smin<mode>3_iwmmxt): Likewise.
16019 (umin<mode>3_iwmmxt): Likewise.
16020 (iwmmxt_wpackhss): Likewise.
16021 (iwmmxt_wpackwss): Likewise.
16022 (iwmmxt_wpackdss): Likewise.
16023 (iwmmxt_wpackhus): Likewise.
16024 (iwmmxt_wpackwus): Likewise.
16025 (iwmmxt_wpackdus): Likewise.
16026 (iwmmxt_wunpckihb): Likewise.
16027 (iwmmxt_wunpckihh): Likewise.
16028 (iwmmxt_wunpckihw): Likewise.
16029 (iwmmxt_wunpckilb): Likewise.
16030 (iwmmxt_wunpckilh): Likewise.
16031 (iwmmxt_wunpckilw): Likewise.
16032 (iwmmxt_wunpckehub): Likewise.
16033 (iwmmxt_wunpckehuh): Likewise.
16034 (iwmmxt_wunpckehuw): Likewise.
16035 (iwmmxt_wunpckehsb): Likewise.
16036 (iwmmxt_wunpckehsh): Likewise.
16037 (iwmmxt_wunpckehsw): Likewise.
16038 (iwmmxt_wunpckelub): Likewise.
16039 (iwmmxt_wunpckeluh): Likewise.
16040 (iwmmxt_wunpckeluw): Likewise.
16041 (iwmmxt_wunpckelsb): Likewise.
16042 (iwmmxt_wunpckelsh): Likewise.
16043 (iwmmxt_wunpckelsw): Likewise.
16044 (ror<mode>3): Likewise.
16045 (ashr<mode>3_iwmmxt): Likewise.
16046 (lshr<mode>3_iwmmxt): Likewise.
16047 (ashl<mode>3_iwmmxt): Likewise.
16048 (ror<mode>3_di): Likewise.
16049 (ashr<mode>3_di): Likewise.
16050 (lshr<mode>3_di): Likewise.
16051 (ashl<mode>3_di): Likewise.
16052 (iwmmxt_wmadds): Likewise.
16053 (iwmmxt_wmaddu): Likewise.
16054 (iwmmxt_tmia): Likewise.
16055 (iwmmxt_tmiaph): Likewise.
16056 (iwmmxt_tmiabb): Likewise.
16057 (iwmmxt_tmiatb): Likewise.
16058 (iwmmxt_tmiabt): Likewise.
16059 (iwmmxt_tmiatt): Likewise.
16060 (iwmmxt_tmovmskb): Likewise.
16061 (iwmmxt_tmovmskh): Likewise.
16062 (iwmmxt_tmovmskw): Likewise.
16063 (iwmmxt_waccb): Likewise.
16064 (iwmmxt_wacch): Likewise.
16065 (iwmmxt_waccw): Likewise.
16066 (iwmmxt_waligni): Likewise.
16067 (iwmmxt_walignr): Likewise.
16068 (iwmmxt_walignr0): Likewise.
16069 (iwmmxt_walignr1): Likewise.
16070 (iwmmxt_walignr2): Likewise.
16071 (iwmmxt_walignr3): Likewise.
16072 (iwmmxt_wsadb): Likewise.
16073 (iwmmxt_wsadh): Likewise.
16074 (iwmmxt_wsadbz): Likewise.
16075 (iwmmxt_wsadhz): Likewise.
16076 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16077 (iwmmxt_wabsdiffb): Likewise.
16078 (iwmmxt_wabsdiffh): Likewise.
16079 (iwmmxt_wabsdiffw): Likewise.
16080 (iwmmxt_waddsubhx): Likewise
16081 (iwmmxt_wsubaddhx): Likewise.
16082 (addc<mode>3): Likewise.
16083 (iwmmxt_avg4): Likewise.
16084 (iwmmxt_avg4r): Likewise.
16085 (iwmmxt_wmaddsx): Likewise.
16086 (iwmmxt_wmaddux): Likewise.
16087 (iwmmxt_wmaddsn): Likewise.
16088 (iwmmxt_wmaddun): Likewise.
16089 (iwmmxt_wmulwsm): Likewise.
16090 (iwmmxt_wmulwum): Likewise.
16091 (iwmmxt_wmulsmr): Likewise.
16092 (iwmmxt_wmulumr): Likewise.
16093 (iwmmxt_wmulwsmr): Likewise.
16094 (iwmmxt_wmulwumr): Likewise.
16095 (iwmmxt_wmulwl): Likewise.
16096 (iwmmxt_wqmulm): Likewise.
16097 (iwmmxt_wqmulwm): Likewise.
16098 (iwmmxt_wqmulmr): Likewise.
16099 (iwmmxt_wqmulwmr): Likewise.
16100 (iwmmxt_waddbhusm): Likewise.
16101 (iwmmxt_waddbhusl): Likewise.
16102 (iwmmxt_wqmiabb): Likewise.
16103 (iwmmxt_wqmiabt): Likewise.
16104 (iwmmxt_wqmiatb): Likewise.
16105 (iwmmxt_wqmiatt): Likewise.
16106 (iwmmxt_wqmiabbn): Likewise.
16107 (iwmmxt_wqmiabtn): Likewise.
16108 (iwmmxt_wqmiatbn): Likewise.
16109 (iwmmxt_wqmiattn): Likewise.
16110 (iwmmxt_wmiabb): Likewise.
16111 (iwmmxt_wmiabt): Likewise.
16112 (iwmmxt_wmiatb): Likewise.
16113 (iwmmxt_wmiatt): Likewise.
16114 (iwmmxt_wmiabbn): Likewise.
16115 (iwmmxt_wmiabtn): Likewise.
16116 (iwmmxt_wmiatbn): Likewise.
16117 (iwmmxt_wmiattn): Likewise.
16118 (iwmmxt_wmiawbb): Likewise.
16119 (iwmmxt_wmiawbt): Likewise.
16120 (iwmmxt_wmiawtb): Likewise.
16121 (iwmmxt_wmiawtt): Likewise.
16122 (iwmmxt_wmiawbbn): Likewise.
16123 (iwmmxt_wmiawbtn): Likewise.
16124 (iwmmxt_wmiawtbn): Likewise.
16125 (iwmmxt_wmiawttn): Likewise.
16126 (iwmmxt_wmerge): Likewise.
16127 (iwmmxt_tandc<mode>3): Likewise.
16128 (iwmmxt_torc<mode>3): Likewise.
16129 (iwmmxt_torvsc<mode>3): Likewise.
16130 (iwmmxt_textrc<mode>3): Likewise.
16131 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16132 (wmmxt_pack): Likewise.
16133 (wmmxt_mult_c1): Likewise.
16134 (wmmxt_mult_c2): Likewise.
16135 (wmmxt_alu_c1): Likewise.
16136 (wmmxt_alu_c2): Likewise.
16137 (wmmxt_alu_c3): Likewise.
16138 (wmmxt_transfer_c1): Likewise.
16139 (wmmxt_transfer_c2): Likewise.
16140 (wmmxt_transfer_c3): Likewise.
16141 (marvell_f_iwmmxt_wstr): Likewise.
16142 (marvell_f_iwmmxt_wldr): Likewise.
16144 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16146 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16148 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16151 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16152 * lra-constraints.c (need_for_split_p): Check call used hard regs
16153 living through calls.
16155 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16156 call used regs for call insn.
16158 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16161 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
16162 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
16164 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16166 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
16168 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16170 * lra-constraints.c (need_for_split_p): Check call used hard regs
16171 living through calls.
16173 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
16176 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
16177 to tie with any other modes. Eliminate Altivec vector mode tests,
16178 since these are a subset of ALTIVEC or VSX vector modes. Simplify
16179 code, to return 0 if testing MODE2 for a condition, if we've
16180 already tested MODE1 for the same condition.
16182 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16184 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
16187 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16189 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16190 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
16192 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16194 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
16196 (aarch64_symbolic_constant_p): Remove.
16197 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
16198 static. Fix line length and white space.
16199 (aarch64_symbolic_constant_p): Remove.
16200 * config/aarch64/predicates.md (aarch64_valid_symref):
16201 Use aarch64_classify_symbol_expression.
16203 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16205 * config/arm/constraints.md (Ts): New constraint.
16206 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
16208 (compare_scc): Use "Ts" constraint for operand 0.
16209 (ior_scc_scc): Likewise.
16210 (and_scc_scc): Likewise.
16211 (and_scc_scc_nodom): Likewise.
16212 (ior_scc_scc_cmp): Likewise for operand 7.
16213 (and_scc_scc_cmp): Likewise.
16214 * config/arm/thumb2.md (thumb2_movsi_insn):
16215 Add alternatives for 16-bit encodings.
16216 (thumb2_movhi_insn): Likewise.
16217 (thumb2_movsicc_insn): Likewise.
16218 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16219 (thumb2_negscc): Use "Ts" constraint.
16220 Move mvn instruction outside cond_exec block.
16221 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16222 for 16-bit encodings.
16224 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16226 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16228 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16229 (mulsi3subsi): Likewise.
16230 (mulsidi3adddi): Likewise.
16231 (mulsidi3_v6): Likewise.
16232 (umulsidi3_v6): Likewise.
16233 (umulsidi3adddi_v6): Likewise.
16234 (smulsi3_highpart_v6): Likewise.
16235 (umulsi3_highpart_v6): Likewise.
16236 (mulhisi3tb): Likewise.
16237 (mulhisi3bt): Likewise.
16238 (mulhisi3tt): Likewise.
16239 (maddhisi4): Likewise.
16240 (maddhisi4tb): Likewise.
16241 (maddhisi4tt): Likewise.
16242 (maddhidi4): Likewise.
16243 (maddhidi4tb): Likewise.
16244 (maddhidi4tt): Likewise.
16245 (zeroextractsi_compare0_scratch): Likewise.
16246 (insv_zero): Likewise.
16247 (insv_t2): Likewise.
16248 (anddi_notzesidi_di): Likewise.
16249 (anddi_notsesidi_di): Likewise.
16250 (andsi_notsi_si): Likewise.
16251 (iordi_zesidi_di): Likewise.
16252 (xordi_zesidi_di): Likewise.
16253 (andsi_iorsi3_notsi): Likewise.
16254 (smax_0): Likewise.
16255 (smax_m1): Likewise.
16256 (smin_0): Likewise.
16257 (not_shiftsi): Likewise.
16258 (unaligned_loadsi): Likewise.
16259 (unaligned_loadhis): Likewise.
16260 (unaligned_loadhiu): Likewise.
16261 (unaligned_storesi): Likewise.
16262 (unaligned_storehi): Likewise.
16263 (extv_reg): Likewise.
16264 (extzv_t2): Likewise.
16265 (divsi3): Likewise.
16266 (udivsi3): Likewise.
16267 (arm_zero_extendhisi2addsi): Likewise.
16268 (arm_zero_extendqisi2addsi): Likewise.
16269 (compareqi_eq0): Likewise.
16270 (arm_extendhisi2_v6): Likewise.
16271 (arm_extendqisi2addsi): Likewise.
16272 (arm_movt): Likewise.
16273 (thumb2_ldrd): Likewise.
16274 (thumb2_ldrd_base): Likewise.
16275 (thumb2_ldrd_base_neg): Likewise.
16276 (thumb2_strd): Likewise.
16277 (thumb2_strd_base): Likewise.
16278 (thumb2_strd_base_neg): Likewise.
16279 (arm_negsi2): Add alternative for 16-bit encoding.
16280 (arm_one_cmplsi2): Likewise.
16282 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16284 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16285 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16286 (movdfcc): Likewise.
16287 * config/arm/vfp.md (*thumb2_movsf_vfp):
16288 Disable predication for arm_restrict_it.
16289 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16290 (*thumb2_movdfcc_vfp): Likewise.
16291 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16292 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16293 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16294 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16295 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16296 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16297 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16298 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16299 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16300 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16301 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16302 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16303 Disable predication for arm_restrict_it.
16305 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16307 * config/i386/bmiintrin.h (_bextr_u32): New.
16308 (_bextr_u64): Ditto.
16310 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16312 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16313 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16314 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16315 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16316 * config/mips/n32-elf.h: ...this new file.
16318 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16321 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16322 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16324 2013-06-27 Catherine Moore <clm@codesourcery.com>
16326 * config/mips/mips-tables.opt: Regenerate.
16327 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16328 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16329 * doc/invoke.texi: Add -m14kc.
16331 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16334 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16335 constraints of operand 1 and 2.
16338 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16339 to match RTL canonicalization. Swap predicates and
16340 constraints of operand 1 and 2.
16342 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16344 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16345 Process OP_INOUT regs for splitting too.
16347 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16349 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16350 decl before the loop, initialize to NULL.
16351 (vectorizable_load): Initialize ptr_incr to NULL.
16353 2013-06-27 Martin Jambor <mjambor@suse.cz>
16356 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16357 * ipa-ref.c (ipa_maybe_record_reference): New function.
16358 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16359 * ipa-cp.c (create_specialized_node): Record potential references from
16361 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16363 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16365 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16366 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16367 'mode' to force_reg.
16368 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16369 (aarch64_expand_mov_immediate): Likewise.
16371 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16373 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16374 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16376 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16378 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16379 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16380 (struct machine_function): Add tbegin_p.
16381 (s390_canonicalize_comparison): Fold CC mode compares to
16382 conditional jump if possible.
16383 (s390_emit_jump): Return the emitted jump.
16384 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16385 Handle CCRAWmode compares.
16386 (s390_option_override): Default to -mhtm if available.
16387 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16388 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16389 FPRs instead of df_regs_ever_live_p.
16390 (s390_optimize_nonescaping_tx): New function.
16391 (s390_init_frame_layout): Extend clobbered_regs array to cover
16393 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16394 (s390_expand_tbegin): New function.
16395 (enum s390_builtin): New enum definition.
16396 (code_for_builtin): New array definition.
16397 (s390_init_builtins): New function.
16398 (s390_expand_builtin): New function.
16399 (TARGET_INIT_BUILTINS): Define.
16400 (TARGET_EXPAND_BUILTIN): Define.
16401 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16402 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16403 (s390_alc_comparison): Likewise.
16404 * config/s390/s390-modes.def: Add CCRAWmode.
16405 * config/s390/s390.h (processor_flags): Add PF_TX.
16406 (TARGET_CPU_HTM): Define macro.
16407 (TARGET_HTM): Define macro.
16408 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16409 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16410 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16411 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16412 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16413 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16414 ("*cc_to_int"): Move up.
16415 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16416 constants other than 0.
16417 ("*ccraw_to_int"): New insn and splitter definition.
16418 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16419 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16420 ("tx_assist"): New expander.
16421 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16422 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16423 * config/s390/s390.opt: Add -mhtm option.
16424 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16425 * config/s390/htmxlintrin.h: New file.
16426 * config/s390/htmintrin.h: New file.
16427 * config/s390/s390intrin.h: New file.
16428 * doc/extend.texi: Document htm builtins.
16429 * config.gcc: Add the new header files to extra_headers.
16431 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16433 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16434 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16436 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16437 Pat Haugen <pthaugen@us.ibm.com>
16438 Peter Bergner <bergner@vnet.ibm.com>
16440 * config/rs6000/power8.md: New.
16441 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16442 setting for power8 entry.
16443 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16444 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16445 test for Power4/Power5 only.
16446 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16448 (force_new_group): Adjust comment.
16449 * config/rs6000/rs6000.md: Include power8.md.
16451 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16453 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16454 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16456 (tune_params): Update comment.
16457 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16458 (arm_max_conditional_execute): New function.
16459 (thumb2_final_prescan_insn): Use max_insn_skipped and
16460 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16462 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16464 PR tree-optimization/57705
16465 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16466 SSA_NAME step, provided that it is not defined inside the loop.
16467 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16468 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16470 2013-06-25 Martin Jambor <mjambor@suse.cz>
16472 PR middle-end/57670
16473 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16474 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16476 (ipa_note_param_call): Initialize member_ptr flag.
16477 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16478 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16479 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16480 (ipa_read_indirect_edge_info): Likewise.
16482 2013-06-25 Richard Biener <rguenther@suse.de>
16484 PR middle-end/56977
16485 * passes.c (init_optimization_passes): Move pass_fold_builtins
16486 and pass_dce earlier with -Og.
16488 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16490 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16491 <BIT_FIELD_REF>: Remove trailing TAB.
16492 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16495 2013-06-24 Martin Jambor <mjambor@suse.cz>
16497 PR tree-optimization/57358
16498 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16499 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16500 (ipa_analyze_params_uses): Generate pessimistic info when true.
16502 2013-06-24 Martin Jambor <mjambor@suse.cz>
16504 PR tree-optimization/57539
16505 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16506 global.inlined_to of the new node to it. All callers changed.
16507 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16508 inlining_into, pass it to cgraph_clone_node.
16509 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16510 ipa_free_edge_args_substructures.
16511 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16512 rdesc linked list. Do not assert rdesc edges have inlined caller.
16513 Assert we have found an rdesc in the rdesc list.
16515 2013-06-24 Richard Biener <rguenther@suse.de>
16517 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16518 (pointer_set_lookup): Declare.
16519 (class pointer_map): New template class implementing a
16520 generic pointer to T map.
16521 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16522 pointer_map<T>::contains, pointer_map<T>::insert,
16523 pointer_map<T>::traverse): New functions.
16524 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16525 (pointer_set_lookup): New function.
16526 (pointer_set_contains): Use pointer_set_lookup.
16527 (pointer_set_insert): Likewise.
16528 (insert_aux): Remove.
16529 (struct pointer_map_t): Embed a pointer_set_t.
16530 (pointer_map_create): Adjust.
16531 (pointer_map_destroy): Likewise.
16532 (pointer_map_contains): Likewise.
16533 (pointer_map_insert): Likewise.
16534 (pointer_map_traverse): Likewise.
16535 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16536 pointer_map<unsigned> instead of a pointer_map_t.
16537 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16538 (streamer_tree_cache_lookup): Likewise.
16539 (streamer_tree_cache_create): Likewise.
16540 (streamer_tree_cache_delete): Likewise.
16541 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16542 pointer_map<unsigned> instead of a pointer_map_t.
16543 (lto_init_tree_ref_encoder): Adjust.
16544 (lto_destroy_tree_ref_encoder): Likewise.
16545 * lto-section-out.c (lto_output_decl_index): Likewise.
16546 (lto_record_function_out_decl_state): Likewise.
16547 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16549 2013-06-24 Richard Biener <rguenther@suse.de>
16551 PR tree-optimization/57488
16552 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16554 2013-06-24 Alan Modra <amodra@gmail.com>
16556 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16557 (gen_easy_altivec_constant): Likewise.
16558 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16559 easy_vector_constant_msb): Likewise.
16561 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16564 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16565 add missing return true.
16567 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16570 * config/sh/predicates.md (general_extend_operand): Invoke
16571 general_movsrc_operand for memory operands.
16572 (general_movsrc_operand): Allow reg+reg addressing, do not use
16573 general_operand for memory operands.
16575 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16577 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16578 when current target options does not apply.
16579 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16580 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16581 * config/i386/bmiintrin.h: Pass appropriate target
16582 attributes to header.
16583 * config/i386/mmintrin.h: Ditto.
16584 * config/i386/nmmintrin.h: Ditto.
16585 * config/i386/avx2intrin.h: Ditto.
16586 * config/i386/fxsrintrin.h: Ditto.
16587 * config/i386/tbmintrin.h: Ditto.
16588 * config/i386/xsaveintrin.h: Ditto.
16589 * config/i386/f16cintrin.h: Ditto.
16590 * config/i386/xtestintrin.h: Ditto.
16591 * config/i386/xsaveoptintrin.h: Ditto.
16592 * config/i386/bmi2intrin.h: Ditto.
16593 * config/i386/lzcntintrin.h: Ditto.
16594 * config/i386/smmintrin.h: Ditto.
16595 * config/i386/wmmintrin.h: Ditto.
16596 * config/i386/x86intrin.h: Remove all header include guards.
16597 * config/i386/prfchwintrin.h: Ditto.
16598 * config/i386/pmmintrin.h: Ditto.
16599 * config/i386/tmmintrin.h: Ditto.
16600 * config/i386/xmmintrin.h: Ditto.
16601 * config/i386/popcntintrin.h: Ditto.
16602 * config/i386/rdseedintrin.h: Ditto.
16603 * config/i386/ammintrin.h: Ditto.
16604 * config/i386/emmintrin.h: Ditto.
16605 * config/i386/immintrin.h: Remove all header include guards.
16606 * config/i386/fma4intrin.h: Ditto.
16607 * config/i386/lwpintrin.h: Ditto.
16608 * config/i386/xopintrin.h: Ditto.
16609 * config/i386/ia32intrin.h: Ditto.
16610 * config/i386/avxintrin.h: Ditto.
16611 * config/i386/rtmintrin.h: Ditto.
16612 * config/i386/fmaintrin.h: Ditto.
16613 * config/i386/mm3dnow.h: Ditto.
16615 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16617 * common/config/i386/i386-common.c: Handle LZCNT.
16619 2013-06-22 Andi Kleen <ak@linux.intel.com>
16621 * doc/extend.texi: Use __atomic_store_n instead of
16622 __atomic_store in HLE example.
16624 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16626 * config/sh/sh.c: Remove <cstdlib> workaround.
16628 2013-06-21 Andi Kleen <ak@linux.intel.com>
16630 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
16632 2013-06-21 Andi Kleen <ak@linux.intel.com>
16634 * doc/extend.texi: Document that __atomic_clear and
16635 __atomic_test_and_set should only be used with bool.
16637 2013-06-20 Jan Hubicka <jh@suse.cz>
16639 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
16640 types_same_for_odr.
16641 * tree.c (decls_same_for_odr): New function.
16642 (same_for_edr): New function.
16643 (types_same_for_odr): New function.
16644 (get_binfo_at_offset): Use it.
16645 * tree.h (types_same_for_odr): Declare.
16647 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
16648 Jason Merrill <jason@redhat.com>
16650 * system.h: Include <cstdlib> as well as <stdlib.h>.
16652 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
16655 * config/i386/i386.c (construct_container): Report error if
16656 long double is used with disabled x87 float returns.
16658 2013-06-20 Jan Hubicka <jh@suse.cz>
16660 * lto-cgraph.c (input_symtab): Do not set cgraph state.
16662 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
16664 PR rtl-optimization/57425
16665 PR rtl-optimization/57569
16666 * alias.c (write_dependence_p): Remove parameters mem_mode and
16667 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
16668 Changed all callers.
16669 (canon_anti_dependence): Get comments and semantics in sync.
16670 Add parameter mem_canonicalized. Changed all callers.
16671 * rtl.h (canon_anti_dependence): Update prototype.
16673 2013-06-20 Richard Biener <rguenther@suse.de>
16675 * data-streamer-in.c (streamer_read_uhwi): Optimize single
16676 byte case, inline streamer_read_uchar and defer section
16679 2013-06-20 Richard Biener <rguenther@suse.de>
16681 PR tree-optimization/57584
16682 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
16683 SSA names into the expanded expression that take part in
16684 abnormal coalescing.
16686 2013-06-19 Sharad Singhai <singhai@google.com>
16688 * gcov.c (print_usage): Handle new option.
16689 (process_args): Ditto.
16690 (get_gcov_intermediate_filename): New function.
16691 (output_intermediate_file): New function.
16692 (output_gcov_file): New function
16693 (generate_results): Handle new option.
16694 (release_function): Relase demangled name.
16695 (read_graph_file): Handle demangled name.
16696 (output_lines): Ditto.
16697 * doc/gcov.texi: Document gcov intermediate format.
16699 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
16702 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
16703 (lra_emit_add): Use the functions. Add comment about Y as an
16706 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
16709 * collect2.c (collect_atexit): New.
16710 (collect_exit): Delete.
16711 (main): Register collect_atexit with atexit.
16712 (collect_wait): Change collect_exit to exit.
16714 * collect2.h (collect_exit): Delete.
16715 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
16717 2013-06-19 Wei Mi <wmi@google.com>
16719 PR rtl-optimization/57518
16720 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
16721 if regno is used in paradoxical subreg.
16722 (update_equiv_regs): Check pdx_subregs[regno] before
16723 set a reg to be equivalent with a mem.
16725 2013-06-19 Matthias Klose <doko@ubuntu.com>
16728 * file-find.h (find_a_file): Add a mode parameter.
16729 * file-find.c (find_a_file): Likewise.
16730 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
16731 with X_OK for the executables.
16732 * collect2.c (main): Call find_a_file with X_OK.
16734 2013-06-19 Steve Ellcey <sellcey@mips.com>
16737 * config/mips/mips.md (casesi_internal_mips16_<mode>):
16738 Use NEXT_INSN instead of next_real_insn.
16740 2013-06-19 Jan Hubicka <jh@suse.cz>
16742 * cgraph.h (const_value_known_p): Replace by ...
16743 (ctor_for_folding): .. this one.
16744 * cgraphunit.c (process_function_and_variable_attributes): Use it.
16745 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
16746 * expr.c (expand_expr_real_1): Likewise.
16747 (string_constant): Likewise.
16748 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
16749 * ipa.c (process_references): Likewise.
16750 (symtab_remove_unreachable_nodes): Likewise.
16751 * ipa-inline-analysis.c (param_change_prob): Likewise.
16752 * gimple-fold.c (canonicalize_constructor_val): Likewise.
16753 (get_base_constructor): Likwise.
16754 * varpool.c (varpool_remove_node): Likewise.
16755 (varpool_remove_initializer): LIkewise.
16756 (dump_varpool_node): LIkwise.
16757 (const_value_known_p): Rewrite to ...
16758 (ctor_for_folding): ... this one.
16760 2013-06-19 Jakub Jelinek <jakub@redhat.com>
16763 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
16764 PERSONALITY in $PATH derived prefixes.
16766 2013-06-19 Jeff Law <law@redhat.com>
16768 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
16771 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
16772 (simplify_bitwise_binary): Use it to simpify certain binary ops on
16775 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
16777 * config/arm/vfp.md: Move VFP instruction classification documentation
16779 * config/arm/arm.md: ... here. Update instruction classification
16782 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
16784 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
16785 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
16786 pattern. Use more efficient sequences on ARMv5 and Thumb2.
16788 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
16791 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16792 with NEXT_INSN. Use tablejump_p to check for jump table data
16795 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
16798 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
16799 that now in C++ the value is correct per the C++ standards.
16801 2013-06-19 Richard Biener <rguenther@suse.de>
16803 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
16804 for global context.
16806 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16809 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16812 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16813 with next_active_insn.
16815 2013-06-18 Sriraman Tallam <tmsriram@google.com>
16817 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16818 functions are inlined during failures to flag an error.
16819 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16820 in early inline pass.
16822 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16824 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16827 2013-06-18 Julian Brown <julian@codesourcery.com>
16829 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16830 Permit virtual register pre-reload if !strict.
16831 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16833 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16835 * config/arm/neon.md (movmisalign<mode>): Use
16836 neon_perm_struct_or_reg_operand instead of
16837 neon_struct_or_register_operand.
16838 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16839 neon_permissive_struct_operand instead of neon_struct_operand.
16840 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16841 neon_vector_mem_operand.
16842 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16843 neon_vector_mem_operand.
16844 (neon_permissive_struct_operand): New.
16845 (neon_struct_or_register_operand): Rename to...
16846 (neon_perm_struct_or_reg_operand): This. Adjust call to
16847 neon_vector_mem_operand.
16849 2013-06-18 Richard Biener <rguenther@suse.de>
16851 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
16852 * lto-streamer.h: Include pointer-set.h.
16853 (struct lto_decl_slot): Remove.
16854 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
16855 Remove next_index entry.
16856 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16857 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16858 (lto_init_tree_ref_encoder): Adjust.
16859 (lto_destroy_tree_ref_encoder): Likewise.
16860 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16861 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16862 (lto_output_decl_index): Adjust.
16863 (lto_new_out_decl_state): Likewise.
16864 (lto_record_function_out_decl_state): Likewise.
16865 * lto-streamer-out.c (copy_function): Likewise.
16867 2013-06-18 Richard Biener <rguenther@suse.de>
16869 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
16870 * cgraphunit.c: Include cfgloop.h.
16871 (init_lowered_empty_function): Initialize the loop tree.
16872 (assemble_thunk): Insert new BBs into loops.
16874 2013-06-18 Richard Biener <rguenther@suse.de>
16876 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
16877 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
16878 the map from cache entry to cache index optional.
16879 (streamer_tree_cache_replace_tree): Adjust accordingly.
16880 (streamer_tree_cache_append): Likewise.
16881 (streamer_tree_cache_delete): Likewise.
16882 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
16883 streamer cache map from cache entry to cache index.
16884 * lto-streamer-out.c (create_output_block): Adjust.
16886 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
16888 * config/arm/arm.md (attribute "insn"): Move multiplication and
16889 division attributes to...
16890 (attribute "type"): ... here. Remove mult.
16891 (attribute "mul32"): New attribute.
16892 (attribute "mul64"): Add umaal.
16893 (*arm_mulsi3): Update attributes.
16894 (*arm_mulsi3_v6): Likewise.
16895 (*thumb_mulsi3): Likewise.
16896 (*thumb_mulsi3_v6): Likewise.
16897 (*mulsi3_compare0): Likewise.
16898 (*mulsi3_compare0_v6): Likewise.
16899 (*mulsi_compare0_scratch): Likewise.
16900 (*mulsi_compare0_scratch_v6): Likewise.
16901 (*mulsi3addsi): Likewise.
16902 (*mulsi3addsi_v6): Likewise.
16903 (*mulsi3addsi_compare0): Likewise.
16904 (*mulsi3addsi_compare0_v6): Likewise.
16905 (*mulsi3addsi_compare0_scratch): Likewise.
16906 (*mulsi3addsi_compare0_scratch_v6): Likewise.
16907 (*mulsi3subsi): Likewise.
16908 (*mulsidi3adddi): Likewise.
16909 (*mulsi3addsi_v6): Likewise.
16910 (*mulsidi3adddi_v6): Likewise.
16911 (*mulsidi3_nov6): Likewise.
16912 (*mulsidi3_v6): Likewise.
16913 (*umulsidi3_nov6): Likewise.
16914 (*umulsidi3_v6): Likewise.
16915 (*umulsidi3adddi): Likewise.
16916 (*umulsidi3adddi_v6): Likewise.
16917 (*smulsi3_highpart_nov6): Likewise.
16918 (*smulsi3_highpart_v6): Likewise.
16919 (*umulsi3_highpart_nov6): Likewise.
16920 (*umulsi3_highpart_v6): Likewise.
16921 (mulhisi3): Likewise.
16922 (*mulhisi3tb): Likewise.
16923 (*mulhisi3bt): Likewise.
16924 (*mulhisi3tt): Likewise.
16925 (maddhisi4): Likewise.
16926 (*maddhisi4tb): Likewise.
16927 (*maddhisi4tt): Likewise.
16928 (maddhidi4): Likewise.
16929 (*maddhidi4tb): Likewise.
16930 (*maddhidi4tt): Likewise.
16931 (divsi3): Likewise.
16932 (udivsi3): Likewise.
16933 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
16934 (thumb2_mulsi_short_compare0): Likewise.
16935 (thumb2_mulsi_short_compare0_scratch): Likewise.
16936 * config/arm/arm1020e.md (1020mult1): Update attribute change.
16937 (1020mult2): Likewise.
16938 (1020mult3): Likewise.
16939 (1020mult4): Likewise.
16940 (1020mult5): Likewise.
16941 (1020mult6): Likewise.
16942 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
16944 (cortex_a15_mult64): Likewise.
16945 (cortex_a15_sdiv): Likewise.
16946 (cortex_a15_udiv): Likewise.
16947 * config/arm/arm1026ejs.md (mult1): Update attribute change.
16953 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
16954 (pj4_ir_div): Likewise.
16955 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
16956 (11_mult2): Likewise.
16957 (11_mult3): Likewise.
16958 (11_mult4): Likewise.
16959 (11_mult5): Likewise.
16960 (11_mult6): Likewise.
16961 (11_mult7): Likewise.
16962 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
16963 (cortex_a8_mla): Likewise.
16964 (cortex_a8_mull): Likewise.
16965 (cortex_a8_smulwy): Likewise.
16966 (cortex_a8_smlald): Likewise.
16967 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
16968 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
16969 (cortex_r4_mul_3): Likewise.
16970 (cortex_r4_mla_4): Likewise.
16971 (cortex_r4_mla_3): Likewise.
16972 (cortex_r4_smlald): Likewise.
16973 (cortex_r4_mull): Likewise.
16974 (cortex_r4_sdiv): Likewise.
16975 (cortex_r4_udiv): Likewise.
16976 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
16977 (cortex_a7_idiv): Likewise.
16978 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
16979 (9_mult2): Likewise.
16980 (9_mult3): Likewise.
16981 (9_mult4): Likewise.
16982 (9_mult5): Likewise.
16983 (9_mult6): Likewise.
16984 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
16985 (cortex_a53_sdiv): Likewise.
16986 (cortex_a53_udiv): Likewise.
16987 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
16988 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
16989 (mp626_mult2): Likewise.
16990 (mp626_mult3): Likewise.
16991 (mp626_mult4): Likewise.
16992 * config/arm/fa526.md (526_mult1): Update attribute change.
16993 (526_mult2): Likewise.
16994 * config/arm/arm-generic.md (mult): Update attribute change.
16995 (mult_ldsched_strongarm): Likewise.
16996 (mult_ldsched): Likewise.
16997 (multi_cycle): Likewise.
16998 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
16999 * config/arm/fa606te.md (606te_mult1): Update attribute change.
17000 (606te_mult2): Likewise.
17001 (606te_mult3): Likewise.
17002 (606te_mult4): Likewise.
17003 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
17004 (cortex_a9_mac16): Likewise.
17005 (cortex_a9_multiply): Likewise.
17006 (cortex_a9_mac): Likewise.
17007 (cortex_a9_multiply_long): Likewise.
17008 * config/arm/fa626te.md (626te_mult1): Update attribute change.
17009 (626te_mult2): Likewise.
17010 (626te_mult3): Likewise.
17011 (626te_mult4): Likewise.
17013 2013-06-18 Richard Biener <rguenther@suse.de>
17016 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17018 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17021 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17022 with next_active_insn.
17024 2013-06-18 Alan Modra <amodra@gmail.com>
17026 * config/rs6000/rs6000.h (enum data_align): New.
17027 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17028 (DATA_ABI_ALIGNMENT): Define.
17029 (CONSTANT_ALIGNMENT): Correct comment.
17030 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17031 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17033 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17035 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17036 ATTRIBUTE_UNUSED marking.
17038 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17040 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17041 alternative and update.
17042 (aarch64_dup_lanedi): Delete.
17043 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17044 * config/aarch64/aarch64-simd-builtins.def: Update.
17046 2013-06-17 Richard Biener <rguenther@suse.de>
17048 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17049 (lto_input_scc): Declare.
17050 (lto_input_tree_1): Likewise.
17051 (struct lto_stats_d): Add num_tree_bodies_output and
17052 num_pickle_refs_output.
17053 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17054 (lto_read_tree_1): Split out from ...
17055 (lto_read_tree): ... this.
17056 (lto_input_scc): New function.
17057 (lto_input_tree_1): Split out from ...
17058 (lto_input_tree): ... this. Handle LTO_tree_scc.
17059 (lto_data_in_create): Create the streamer cache without hashes.
17060 * lto-streamer-out.c (create_output_block): Create the streamer
17061 cache with hashes when not doing WPA.
17062 (lto_write_tree_1): Split out from ...
17063 (lto_write_tree): ... this.
17064 (get_symbol_initial_value): New function.
17065 (lto_output_tree_1): Split out from ...
17066 (lto_output_tree): ... this. Write trees as series of SCCs
17067 using a DFS walk via DFS_write_tree.
17068 (struct sccs, struct scc_entry): New types.
17069 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17070 (DFS_write_tree_body): New function.
17071 (DFS_write_tree): Likewise.
17072 (hash_tree): Likewise.
17073 (scc_entry_compare): Likewise.
17074 (hash_scc): Likewise.
17075 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17076 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17077 TREE_CHAIN as regular reference.
17078 (streamer_read_integer_cst): Remove.
17079 (streamer_get_pickled_tree): Adjust.
17080 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17081 of DECL_EXTERNALs in BLOCK_VARS for now.
17082 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17084 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17085 Add hash value argument and record that if hashes are recorded
17087 (streamer_tree_cache_insert_1): Adjust.
17088 (streamer_tree_cache_insert): Likewise.
17089 (streamer_tree_cache_insert_at): Rename to ...
17090 (streamer_tree_cache_replace_tree): ... this and adjust.
17091 (streamer_tree_cache_append): Adjust.
17092 (record_common_node): Likewise.
17093 (streamer_tree_cache_create): Add argument whether to
17094 record hash values together with trees.
17095 (streamer_tree_cache_delete): Adjust.
17096 * tree-streamer.h (struct streamer_tree_cache_d): Add
17098 (streamer_read_integer_cst): Remove.
17099 (streamer_tree_cache_insert): Adjust.
17100 (streamer_tree_cache_append): Likewise.
17101 (streamer_tree_cache_insert_at): Rename to ...
17102 (streamer_tree_cache_replace_tree): ... this and adjust.
17103 (streamer_tree_cache_create): Add argument whether to record hashes.
17104 (streamer_tree_cache_get): Rename to ...
17105 (streamer_tree_cache_get_tree): ... this.
17106 (streamer_tree_cache_get_hash): New function.
17107 * tree.c (cache_integer_cst): New function.
17108 * tree.h (cache_integer_cst): Declare.
17109 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17110 * lto-symtab.c (lto_varpool_replace_node): Only release
17111 DECL_INITIAL of non-prevailing decls.
17112 * varpool.c (varpool_remove_initializer): Do not release
17113 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17115 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17117 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17118 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17119 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17120 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17121 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17122 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17123 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17124 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17125 instead of TARGET_64BIT.
17126 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17127 Require ISA_HAS_<D>DIV.
17129 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17131 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17132 (mips*-*-linux*): Move default with_llsc setting to where other
17134 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17136 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17137 Likewise. Remove default with_tune setting. Move default float
17138 setting to its own block. Handle with_llsc in the same block as above.
17140 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17142 PR rtl-optimization/57425
17143 PR rtl-optimization/57569
17144 * alias.c (write_dependence_p): Add new parameters mem_mode,
17145 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17146 Changed all callers.
17147 (canon_anti_dependence): New function.
17148 * cse.c (check_dependence): Use canon_anti_dependence.
17149 * cselib.c (cselib_invalidate_mem): Likewise.
17150 * rtl.h (canon_anti_dependence): Declare.
17152 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17154 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17155 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17156 ".set mips3" for 64-bit targets.
17158 2013-06-15 Dehao Chen <dehao@google.com>
17160 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17161 * gimple-low.c (gimple_check_call_matching_types): Likewise.
17162 (gimple_check_call_args): Likewise.
17163 * value-prof.c (check_ic_target): Likewise.
17164 * ipa-inline.c (early_inliner): Likewise.
17165 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
17166 * cgraph.c (cgraph_create_edge_1): Likewise.
17167 (cgraph_make_edge_direct): Likewise.
17169 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
17172 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
17173 rs6000_output_move_128bit to handle emitting quad memory
17174 operations. Set attribute length to 8 bytes.
17176 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
17178 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
17180 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
17181 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
17182 (aarch64_<su>mlsl<mode>): Likewise.
17184 2013-06-14 Mike Stump <mikestump@comcast.net>
17186 * Makefile.in (TARGET_H): Add insn-codes.h.
17188 2013-06-14 Alan Modra <amodra@gmail.com>
17190 PR middle-end/57134
17191 PR middle-end/57586
17192 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
17193 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
17194 bitfield expansion when EXPAND_MEMORY.
17195 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
17197 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
17199 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
17200 test for clearing quad memory on 32-bit later.
17202 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17204 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
17205 (fold_negate_expr): Likewise.
17206 (fold_real_zero_addition_p): Handle vectors.
17207 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
17209 2013-06-14 Alan Modra <amodra@gmail.com>
17211 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17213 2013-06-13 Jan Hubicka <jh@suse.cz>
17215 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17216 Local comdats are not externally visible.
17217 * symtab.c (dump_symtab_base): Dump externally visible.
17218 (verify_symtab_base): Verify back links in the symtab hash.
17220 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17222 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17223 CONVERT_EXPR as equal nodes.
17225 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17227 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17229 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17231 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17232 Generalize to complex and vector.
17233 * tree.c (build_all_ones_cst): New function.
17234 * tree.h (build_all_ones_cst): Declare it.
17236 2013-06-13 Alan Modra <amodra@gmail.com>
17238 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17239 * config/rs6000/rs6000.md (signbittf2): New insn.
17240 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17241 (abstf2_internal, cmptf_internal2): Likewise.
17242 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17244 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17245 Pat Haugen <pthaugen@us.ibm.com>
17246 Peter Bergner <bergner@vnet.ibm.com>
17248 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17249 power8 byte, half-word, and quad-word atomic instructions.
17250 (emit_store_conditional): Likewise.
17251 (rs6000_expand_atomic_compare_and_swap): Likewise.
17252 (rs6000_expand_atomic_op): Likewise.
17254 * config/rs6000/sync.md (larx): Add new modes for power8.
17256 (AINT): New mode iterator to include TImode as well as normal
17257 integer modes on power8.
17258 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17259 that VSX registers are not considered. Use AINT mode iterator
17260 instead of INT1 to allow inclusion of quad word atomic operations
17262 (load_locked<mode>): Likewise.
17263 (store_conditional<mode>): Likewise.
17264 (atomic_compare_and_swap<mode>): Likewise.
17265 (atomic_exchange<mode>): Likewise.
17266 (atomic_nand<mode>): Likewise.
17267 (atomic_fetch_<fetchop_name><mode>): Likewise.
17268 (atomic_nand_fetch<mode>): Likewise.
17269 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17271 (ATOMIC): On power8, add QImode, HImode modes.
17272 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17273 modes that promote to SImode.
17274 (load_lockedti): Convert TImode arguments to PTImode, so that we
17275 get a guaranteed even/odd register pair.
17276 (load_lockedpti): Likewise.
17277 (store_conditionalti): Likewise.
17278 (store_conditionalpti): Likewise.
17280 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17281 atomic load/store instructions.
17284 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17286 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17288 (insn_count): New attribute, with most cases extracted from...
17289 (length): ...here. Redefine most cases in terms of insn_count.
17290 (single_insn): Delete.
17291 (can_delay): Use insn_count to check for single instructions.
17292 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17293 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17294 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17295 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17296 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17297 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17298 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17299 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17300 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17301 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17302 rather than "length".
17303 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17304 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17305 Use "insn_count" rather than "length".
17306 * config/mips/mips-dsp.md
17307 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17308 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17311 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17313 PR tree-optimization/57361
17314 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17316 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17318 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17320 (aarch64_simd_combine<mode>): New instruction expansion.
17321 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17322 function prototype.
17323 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17324 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17326 2013-06-12 Jan Hubicka <jh@suse.cz>
17328 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17329 decl has when in streaming stage.
17330 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17331 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17333 2013-06-12 Roland Stigge <stigge@antcom.de>
17336 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17338 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17340 PR tree-optimization/57537
17341 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17342 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17344 2013-06-12 Richard Biener <rguenther@suse.de>
17346 * data-streamer.h (streamer_write_char_stream): CSE
17347 obs->current_pointer.
17348 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17349 streamer_write_char_stream manually and optimize the resulting loop.
17350 (streamer_write_hwi_stream): Likewise.
17352 2013-06-12 Jan Hubicka <jh@suse.cz>
17354 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17355 * cgraph.h (varpool_create_empty_node): Declare.
17356 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17358 * symtab.c (symtab_unregister_node): Be lax about missin entries
17360 (symtab_get_node): Update comment.
17361 * varpool.c (varpool_create_empty_node): Break out from ...
17362 (varpool_node_for_decl): ... here.
17363 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17365 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17367 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17368 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17369 part. Use straight-line flow at the end.
17370 <COMPONENT_REF>: Remove superfluous else.
17371 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17373 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17376 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17377 target hook even for !TREE_PUBLIC decls. If no resolution info
17378 is available, return false for common and external decls.
17380 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17382 * config/rl78/constraints.md (U): New constraint.
17383 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17386 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17389 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17390 to allow returning address to AT_PLATFORM name.
17392 2013-06-11 Jan Hubicka <jh@suse.cz>
17394 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17395 * cgraph.h (symtab_node_base): Add weakref flag.
17396 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17397 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17398 (output_weakrefs): Use weakref flag.
17399 * fold-const.c (simple_operand_p): Handle WEAK.
17400 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17401 * ipa.c (varpool_externally_visible_p): Drop weakref.
17402 (function_and_variable_visibility): Update comment; fix weakref
17403 sanity checks; do not clear DECL_WEAK on them.
17404 * lto-cgraph.c (lto_output_node): update.
17405 (lto_output_varpool_node): Update.
17406 (input_overwrite_node): Update.
17407 (input_node): Update.
17408 (input_varpool_node): Update.
17409 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17410 (lto_symtab_merge_symbols): Add sanity check.
17411 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17412 * passes.c (rest_of_decl_compilation): Set static flag, too.
17413 * symtab.c (dump_symtab_base): Dump weakref.
17414 (verify_symtab_base): Sanity check weakrefs.
17415 (symtab_make_decl_local): Remove duplicated code.
17416 (symtab_alias_ultimate_target): Simplify.
17417 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17419 2013-06-11 Tom de Vries <tom@codesourcery.com>
17421 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17422 sequence_vect == NULL.
17424 2013-06-11 DJ Delorie <dj@redhat.com>
17426 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17427 (rl78_unwind_word_mode): New.
17429 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17431 * final.c (debug_prefix_maps): Make static.
17433 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17435 * function.c (initial_trampoline): Remove stray copy.
17437 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17439 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17441 2013-06-11 Martin Jambor <mjambor@suse.cz>
17443 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17444 within bounds at the beginning of the function.
17446 2013-06-11 Alan Modra <amodra@gmail.com>
17448 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17450 (get_named_section): Don't NULL !DECL_P decl.
17452 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17454 * doc/invoke.texi (core-avx2): Document.
17456 (atom): Updated with MOVBE.
17458 2013-06-11 Richard Biener <rguenther@suse.de>
17460 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17462 2013-06-11 Anton Blanchard <anton@samba.org>
17464 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17465 correct shift value in little-endian mode.
17467 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17470 * varasm.c (get_variable_align): Move #endif to the right place.
17472 2013-06-10 Cary Coutant <ccoutant@google.com>
17474 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17475 for hash so that hash table traversal order is deterministic.
17477 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17478 Pat Haugen <pthaugen@us.ibm.com>
17479 Peter Bergner <bergner@vnet.ibm.com>
17481 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17482 of vectors in GPRS if they are direct moves or quad word load or
17485 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17487 (direct_move_p): Likewise.
17488 (quad_load_store_p): Likewise.
17490 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17491 classes into bins based on the physical register type.
17492 (reg_class_to_reg_type): Likewise.
17493 (IS_STD_REG_TYPE): Likewise.
17494 (IS_FP_VECT_REG_TYPE): Likewise.
17495 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17496 use direct move instructions.
17497 (reload_gpr_vsx): Likewise.
17498 (reload_vsx_gpr): Likewise.
17499 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17500 information that is a simplification of register classes. Also
17501 precalculate direct move reload helpers.
17502 (direct_move_p): New function to return true if the operation can
17503 be done as a direct move instruciton.
17504 (quad_load_store_p): New function to return true if the operation
17505 is a quad memory operation.
17506 (rs6000_legitimize_address): If quad memory, only allow register
17507 indirect for TImode addresses.
17508 (rs6000_legitimate_address_p): Likewise.
17509 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17510 (rs6000_reload_register_type): Likewise.
17511 (register_to_reg_type): Return register type.
17512 (rs6000_secondary_reload_simple_move): New helper function for
17513 secondary reload and secondary memory needed to identify anything
17514 that is a simple move, and does not need reloading.
17515 (rs6000_secondary_reload_direct_move): New helper function for
17516 secondary reload to identify cases that can be done with several
17517 instructions via the direct move instructions.
17518 (rs6000_secondary_reload_move): New helper function for secondary
17519 reload to identify moves between register types that can be done.
17520 (rs6000_secondary_reload): Add support for quad memory operations
17521 and for direct move.
17522 (rs6000_secondary_memory_needed): Likewise.
17523 (rs6000_debug_secondary_memory_needed): Change argument names.
17524 (rs6000_output_move_128bit): New function to return the move to
17525 use for 128-bit moves, including knowing about the various
17526 limitations of quad memory operations.
17528 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17529 memory operations. call rs6000_output_move_128bit for the actual
17530 instruciton(s) to generate.
17531 (vsx_movti_64bit): Likewise.
17533 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17534 (UNSPEC_P8V_MTVSRWZ): Likewise.
17535 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17536 (UNSPEC_P8V_MTVSRD): Likewise.
17537 (UNSPEC_P8V_XXPERMDI): Likewise.
17538 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17539 (UNSPEC_FUSION_GPR): Likewise.
17540 (FMOVE128_GPR): New iterator for direct move.
17541 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17542 (f32_sv): Likewise.
17543 (f32_dm): Likewise.
17544 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17545 loads and direct move instructions.
17546 (zero_extendsidi2_lfiwzx): Likewise.
17547 (extendsidi2_lfiwax): Likewise.
17548 (extendsidi2_nocell): Likewise.
17549 (floatsi<mode>2_lfiwax): Likewise.
17550 (lfiwax): Likewise.
17551 (floatunssi<mode>2_lfiwzx): Likewise.
17552 (lfiwzx): Likewise.
17553 (fix_trunc<mode>_stfiwx): Likewise.
17554 (fixuns_trunc<mode>_stfiwx): Likewise.
17555 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17556 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17557 (parity<mode>2_cmpb): Set length/type attr.
17558 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17559 for 'mr.' to fast_compare.
17560 (bpermd_<mode>): Change type attr to popcnt.
17561 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17562 (p8_mtvsrwz_1): Likewise.
17563 (p8_mtvsrwz_2): Likewise.
17564 (reload_fpr_from_gpr<mode>): Likewise.
17565 (p8_mtvsrd_1): Likewise.
17566 (p8_mtvsrd_2): Likewise.
17567 (p8_xxpermdi_<mode>): Likewise.
17568 (reload_vsx_from_gpr<mode>): Likewise.
17569 (reload_vsx_from_gprsf): Likewise.
17570 (p8_mfvsrd_3_<mode>): LIkewise.
17571 (reload_gpr_from_vsx<mode>): Likewise.
17572 (reload_gpr_from_vsxsf): Likewise.
17573 (p8_mfvsrd_4_disf): Likewise.
17574 (multi-word GPR splits): Do not split direct moves or quad memory
17577 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17579 * tree-into-ssa.c (interesting_blocks): Make static.
17581 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17584 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17585 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17586 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17587 (get_variable_align): New function.
17588 (get_variable_section, emit_bss, emit_common,
17589 assemble_variable_contents, place_block_symbol): Use
17590 get_variable_align instead of DECL_ALIGN.
17591 (assemble_noswitch_variable): Add align argument, use it
17592 instead of DECL_ALIGN.
17593 (assemble_variable): Adjust caller. Use get_variable_align
17594 instead of DECL_ALIGN.
17595 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17597 (DATA_ABI_ALIGNMENT): Define.
17598 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17599 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17600 opt is false, only return the psABI mandated alignment increase.
17601 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17602 (DATA_ABI_ALIGNMENT): ... this.
17603 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17604 (DATA_ABI_ALIGNMENT): ... this.
17605 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17606 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17607 (DATA_ABI_ALIGNMENT): ... this.
17608 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17609 * doc/tm.texi: Regenerated.
17611 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17613 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17614 cmp_code to construct REG_EQUAL note.
17616 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17619 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17620 that operands[2] doesn't overlap with operands[0].
17622 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17623 Jan Hubicka <jh@suse.cz>
17625 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17626 hack to mark symbols as used.
17628 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
17630 PR rtl-optimization/57559
17631 * lra-constraints.c (process_alt_operands): Don't discourage
17632 memory with known offset for offsetable memory constraint.
17633 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
17635 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
17637 * varasm.c (struct oc_local_state): Reorder fields.
17638 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
17639 and adjust accordingly.
17640 (output_constructor): Reorder initialization code and adjust call to
17641 output_constructor_bitfield.
17643 2013-06-07 Jan Hubicka <jh@suse.cz>
17645 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
17647 2013-06-07 David Malcolm <dmalcolm@redhat.com>
17649 * tree-object-size.c (unknown): Make const.
17651 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17653 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
17654 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
17655 for last alternative in the cpu_facility attribute.
17657 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17660 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
17661 (xordi3): Change operand 2 constraint to arm_xordi_operand.
17662 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
17663 * config/arm/constraints.md (Dg): New constraint.
17664 * config/arm/neon.md (xordi3_neon): Remove.
17665 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
17666 * config/arm/predicates.md (arm_xordi_operand): New predicate.
17668 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17670 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
17671 Clean up alternatives.
17673 2013-06-07 Alan Modra <amodra@gmail.com>
17675 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
17678 2013-06-07 Alan Modra <amodra@gmail.com>
17680 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
17682 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17684 * config/arm/constraints.md (Df): New constraint.
17685 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
17686 Correct length attribute for last two alternatives.
17688 2013-06-07 Alan Modra <amodra@gmail.com>
17690 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
17691 override user -mfp-in-toc.
17692 (offsettable_ok_by_alignment): Consider just the current access
17693 rather than the whole object, unless BLKmode. Handle
17694 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
17695 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
17696 for -mcmodel=medium.
17697 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
17698 override user -mfp-in-toc or -msum-in-toc. Default to
17699 -mno-fp-in-toc for -mcmodel=medium.
17701 2013-06-06 DJ Delorie <dj@redhat.com>
17703 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
17704 TARGET_VALID_POINTER_MODE.
17706 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
17707 Pat Haugen <pthaugen@us.ibm.com>
17708 Peter Bergner <bergner@vnet.ibm.com>
17710 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
17711 Document new power8 builtins.
17713 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
17714 condition code register, to allow 128-bit logical operations to be
17715 done in the VSX or GPR registers.
17716 (nor<mode>3): Use the canonical form for nor.
17717 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
17718 vclz*, and vpopcnt* vector instructions.
17719 (nand<mode>3): Likewise.
17720 (orc<mode>3): Likewise.
17721 (clz<mode>2): LIkewise.
17722 (popcount<mode>2): Likewise.
17724 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
17725 that only the GPRs are recognized.
17727 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17728 support for new power8 builtins.
17730 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
17732 (xscvdpspn): Likewise.
17738 (vpopcnt): Likewise.
17739 (vpopcntb): Likewise.
17740 (vpopcnth): Likewise.
17741 (vpopcntw): Likewise.
17742 (vpopcntd): Likewise.
17744 (vmrgew): Likewise.
17745 (vmrgow): Likewise.
17747 (eqv_v16qi3): Likewise.
17748 (eqv_v8hi3): Likewise.
17749 (eqv_v4si3): Likewise.
17750 (eqv_v2di3): Likewise.
17751 (eqv_v4sf3): Likewise.
17752 (eqv_v2df3): Likewise.
17754 (nand_v16qi3): Likewise.
17755 (nand_v8hi3): Likewise.
17756 (nand_v4si3): Likewise.
17757 (nand_v2di3): Likewise.
17758 (nand_v4sf3): Likewise.
17759 (nand_v2df3): Likewise.
17761 (orc_v16qi3): Likewise.
17762 (orc_v8hi3): Likewise.
17763 (orc_v4si3): Likewise.
17764 (orc_v2di3): Likewise.
17765 (orc_v4sf3): Likewise.
17766 (orc_v2df3): Likewise.
17768 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
17769 allow power8 quad mode in 64-bit.
17770 (rs6000_builtin_vectorized_function): Add support to vectorize
17771 ISA 2.07 count leading zeros, population count builtins.
17772 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
17773 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
17774 (builtin_function_type): Add vgbbd builtin function which takes an
17776 (altivec_expand_vec_perm_const): Add support for new power8 merge
17779 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
17780 that does not include TImdoe for use with 32-bit.
17781 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
17783 (UNSPEC_VSX_CVDPSPN): Likewise.
17784 (vsx_xscvdpspn): Likewise.
17785 (vsx_xscvspdpn): Likewise.
17786 (vsx_xscvdpspn_scalar): Likewise.
17787 (vsx_xscvspdpn_directmove): Likewise.
17788 (vsx_and<mode>3): Split logical operations into 32-bit and
17789 64-bit. Add support to do logical operations on TImode as well as
17790 VSX vector types. Allow logical operations to be done in either
17791 VSX registers or in general purpose registers in 64-bit mode. Add
17792 splitters if GPRs were used. For AND, add clobber of CCmode to
17793 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
17795 (vsx_and<mode>3_32bit): Likewise.
17796 (vsx_and<mode>3_64bit): Likewise.
17797 (vsx_ior<mode>3): Likewise.
17798 (vsx_ior<mode>3_32bit): Likewise.
17799 (vsx_ior<mode>3_64bit): Likewise.
17800 (vsx_xor<mode>3): Likewise.
17801 (vsx_xor<mode>3_32bit): Likewise.
17802 (vsx_xor<mode>3_64bit): Likewise.
17803 (vsx_one_cmpl<mode>2): Likewise.
17804 (vsx_one_cmpl<mode>2_32bit): Likewise.
17805 (vsx_one_cmpl<mode>2_64bit): Likewise.
17806 (vsx_nor<mode>3): Likewise.
17807 (vsx_nor<mode>3_32bit): Likewise.
17808 (vsx_nor<mode>3_64bit): Likewise.
17809 (vsx_andc<mode>3): Likewise.
17810 (vsx_andc<mode>3_32bit): Likewise.
17811 (vsx_andc<mode>3_64bit): Likewise.
17812 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
17813 and xxlorc instructions.
17814 (vsx_eqv<mode>3_64bit): Likewise.
17815 (vsx_nand<mode>3_32bit): Likewise.
17816 (vsx_nand<mode>3_64bit): Likewise.
17817 (vsx_orc<mode>3_32bit): Likewise.
17818 (vsx_orc<mode>3_64bit): Likewise.
17820 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17822 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17824 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17825 (p8_vmrgow): Likewise.
17826 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17827 GPRs to be split under VSX.
17828 (p8v_clz<mode>2): Add power8 count leading zero support.
17829 (p8v_popcount<mode>2): Add power8 population count support.
17830 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17833 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17836 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17838 (vec_nand): Likewise.
17839 (vec_vclz): Likewise.
17840 (vec_vclzb): Likewise.
17841 (vec_vclzd): Likewise.
17842 (vec_vclzh): Likewise.
17843 (vec_vclzw): Likewise.
17844 (vec_vgbbd): Likewise.
17845 (vec_vmrgew): Likewise.
17846 (vec_vmrgow): Likewise.
17847 (vec_vpopcnt): Likewise.
17848 (vec_vpopcntb): Likewise.
17849 (vec_vpopcntd): Likewise.
17850 (vec_vpopcnth): Likewise.
17851 (vec_vpopcntw): Likewise.
17853 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17855 PR rtl-optimization/57468
17856 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
17859 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17861 PR rtl-optimization/57459
17862 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
17863 type when setting live regs.
17865 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17867 * config/s390/s390.opt (mlra): New option.
17868 * config/s390/s390.c (s390_decompose_address): Check displacement
17869 for all registers for LRA.
17870 (s390_secondary_reload): Don't used secondary reloads for LRA.
17871 (s390_lra_p): New function.
17872 (TARGET_LRA_P): Define.
17873 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
17874 of attribute cpu_facility to zarch for the last alternative.
17875 (*cmpmem_short): Ditto.
17877 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
17879 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
17880 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
17881 (arm_expand_prologue): Likewise.
17883 2013-06-06 Teresa Johnson <tejohnson@google.com>
17886 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
17887 as this is now done by redirect_edge_and_branch_force.
17888 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
17889 barriers, and fix interaction with splitting.
17890 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
17891 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
17892 reflect changes made in the routine.
17893 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
17894 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
17895 since this is called in cfglayout mode, and replace partition fixup
17896 with assert as that is now done by force_nonfallthru_and_redirect.
17897 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
17898 already be marked with region crossing note.
17899 (insert_section_boundary_note): Make non-static, gate on flag
17900 has_bb_partition, rewrite to also check for multiple partitions.
17901 (rest_of_handle_reorder_blocks): Remove call to
17902 insert_section_boundary_note, now done later during free_cfg.
17903 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
17904 * bb-reorder.h (insert_section_boundary_note): Declare.
17905 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
17906 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
17907 invoke insert_section_boundary_note.
17908 (try_redirect_by_replacing_jump): Remove unnecessary
17909 check for region crossing note.
17910 (fixup_partition_crossing): New function.
17911 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
17912 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
17913 in non-cfglayout mode.
17914 (force_nonfallthru_and_redirect): Fixup partition boundaries,
17915 remove old code that tried to do this. Emit barrier correctly
17916 when we are in cfglayout mode.
17917 (last_bb_in_partition): New function.
17918 (rtl_split_edge): Correctly fixup partition boundaries.
17919 (commit_one_edge_insertion): Remove old code that tried to
17920 fixup region crossing edge since this is now handled in
17921 split_block, and set up insertion point correctly since
17922 block may now end in a jump.
17923 (verify_hot_cold_block_grouping): Guard against checking when not in
17924 linearized RTL mode.
17925 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
17927 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
17928 rtl_verify_flow_info, so not called in cfglayout mode.
17929 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
17930 (fixup_reorder_chain): Remove old code that attempted to fixup region
17931 crossing note as this is now handled in force_nonfallthru_and_redirect.
17932 (duplicate_insn_chain): Don't duplicate switch section notes.
17933 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
17935 * basic-block.h (emit_barrier_after_bb): Declare.
17937 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17939 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
17940 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
17941 arm_usatsihi): Adjust alternatives for arm_restrict_it.
17943 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17945 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
17947 * config/arm/ldmstm.md: Regenerate.
17949 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17951 * config/arm/sync.md (atomic_loaddi_1):
17952 Disable predication for arm_restrict_it.
17953 (arm_load_exclusive<mode>): Likewise.
17954 (arm_load_exclusivesi): Likewise.
17955 (arm_load_exclusivedi): Likewise.
17956 (arm_load_acquire_exclusive<mode>): Likewise.
17957 (arm_load_acquire_exclusivesi): Likewise.
17958 (arm_load_acquire_exclusivedi): Likewise.
17959 (arm_store_exclusive<mode>): Likewise.
17960 (arm_store_exclusive<mode>): Likewise.
17961 (arm_store_release_exclusivedi): Likewise.
17962 (arm_store_release_exclusive<mode>): Likewise.
17964 2013-06-06 Richard Biener <rguenther@suse.de>
17966 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
17968 (lto_tag_is_tree_code_p): Adjust.
17969 (lto_tag_is_gimple_code_p): Likewise.
17970 (lto_gimple_code_to_tag): Likewise.
17971 (lto_tag_to_gimple_code): Likewise.
17972 (lto_tree_code_to_tag): Likewise.
17973 (lto_tag_to_tree_code): Likewise.
17974 * data-streamer.h (streamer_write_hwi_in_range): Use
17975 uhwi streaming to stream the normalized range.
17976 (streamer_read_hwi_in_range): Likewise.
17978 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17980 * config/arm/arm.md (enabled_for_depr_it): New attribute.
17981 (predicable_short_it): Likewise.
17982 (predicated): Likewise.
17983 (enabled): Handle above.
17984 (define_cond_exec): Set predicated attribute to yes.
17986 2013-06-05 Mike Stump <mikestump@comcast.net>
17988 * gdbinit.in (__FUNCTION__): Add.
17990 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
17992 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
17993 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
17995 2013-06-05 Jan Hubicka <jh@suse.cz>
17997 * varasm.c (mark_decl_referenced): Revert the removal until targets
18000 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
18002 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
18003 instead of mark_decl_referenced.
18005 2013-06-05 Jan Hubicka <jh@suse.cz>
18007 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
18008 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
18009 and symtab_used_from_object_file_p.
18010 (cgraph_make_node_local_1): Clear forced_by_abi.
18011 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
18012 * cgraph.h (symtab_node_base): Add forced_by_abi.
18013 (decide_is_variable_needed): Remove.
18014 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
18015 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
18016 (decide_is_symbol_needed): ... this one; handle symbols in general;
18017 always analyze virtuals; honnor forced_by_abi.
18018 (cgraph_finalize_function): Update.
18019 (varpool_finalize_decl): Update.
18020 (symbol_defined_and_needed): Remove.
18021 (analyze_functions): Update.
18022 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18023 output_refs, input_overwrite_node): Handle forced_by_abi.
18024 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18025 (address_taken_from_non_vtable_p): ... this one.
18026 (comdat_can_be_unshared_p_1): New function.
18027 (cgraph_comdat_can_be_unshared_p): Rename to ...
18028 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18029 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18030 (function_and_variable_visibility): Clear forced_by_abi as needed.
18031 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18032 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18033 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18034 * varpool.c (decide_is_variable_needed): Remove.
18036 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18038 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18039 (arm_option_override): Override arm_restrict_it where appropriate.
18040 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18041 * config/arm/arm.opt (mrestrict-it): New command-line option.
18042 * doc/invoke.texi: Document -mrestrict-it.
18044 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18046 * tsan.c (tsan_atomic_table): Make const.
18048 2013-06-05 Richard Biener <rguenther@suse.de>
18050 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18051 index associated with the tree we are supposed to replace.
18052 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18053 TREE_ASM_WRITTEN as zero for everything but SSA names.
18055 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18057 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18059 2013-06-04 Jan Hubicka <jh@suse.cz>
18061 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18062 (input_node, input_varpool_node): Handle correctly external same
18064 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18065 nodes at ltrans stage.
18067 2013-06-04 Jan Hubicka <jh@suse.cz>
18069 * ipa-inline.c (update_caller_keys): Fix availability test.
18070 (update_callee_keys): Likewise.
18071 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18072 to follow ELF standard.
18074 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18076 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18077 (mips64r5900el-*-elf*): New configurations.
18078 * config/mips/mips-cpus.def (r5900): New processor.
18079 * config/mips/mips-tables.opt: Regenerate.
18080 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18081 (mips_issue_rate): Handle PROCESSOR_R5900.
18082 (mips_reorg_process_insns): Force reorder mode for the R5900.
18083 * config/mips/mips.h (TARGET_MIPS5900): Define.
18084 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18086 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18088 * config/mips/mips.md (processor): Add r5900.
18089 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18091 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18093 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18094 into function to generate MOVI instruction.
18095 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18096 (aarch64_preferred_simd_mode): Turn into wrapper.
18097 (aarch64_output_scalar_simd_mov_immediate): New function.
18098 * config/aarch64/aarch64-protos.h: Add prototype for above.
18100 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18102 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18103 element_char member.
18104 (sizetochar): Return signed char.
18105 (aarch64_simd_valid_immediate): Remove elchar and other
18106 unnecessary variables.
18107 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18108 Calculate element_char as required.
18109 * config/aarch64/aarch64-protos.h: Update and move prototype
18110 for aarch64_output_simd_mov_immediate.
18111 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18114 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18116 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18117 information completed by aarch64_simd_valid_immediate.
18118 (aarch64_legitimate_constant_p): Update arguments.
18119 (aarch64_simd_valid_immediate): Work with struct rather than many
18121 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18122 (aarch64_simd_make_constant): Update arguments.
18123 (aarch64_output_simd_mov_immediate): Work with struct rather than
18124 many pointers. Output immediate directly rather than as operand.
18125 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18127 * config/aarch64/constraints.md (Dn): Update arguments.
18129 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18131 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18133 (aarch64_simd_immediate_valid_for_move): Remove.
18134 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18135 (aarch64_simd_make_constant): Update call.
18136 (aarch64_output_simd_mov_immediate): Update call.
18137 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18139 * config/aarch64/constraints.md (Dn): Update call.
18141 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18143 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18144 return type to bool for prototype.
18145 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18146 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18147 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18149 2013-06-04 Catherine Moore <clm@codesourcery.com>
18151 * config/mips/mips.opt (meva): New.
18152 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18153 (ASM_SPEC): Handle -meva.
18154 * doc/invoke.texi (meva): Document.
18156 2013-06-04 Alan Modra <amodra@gmail.com>
18158 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18161 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18163 * rtl.def: Add extra fourth optional field to define_cond_exec.
18164 * gensupport.c (process_one_cond_exec): Process attributes from
18166 * doc/md.texi: Document fourth field in define_cond_exec.
18168 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
18170 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
18171 out the processing order as in store_bit_field_1.
18173 2013-06-04 Jan Hubicka <jh@suse.cz>
18175 PR middle-end/57500
18176 * cgraphunit.c (cgraph_process_same_body_aliases): Create
18177 non-VAR_DECL node if it does not exist yet.
18179 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
18181 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
18182 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
18183 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
18184 target_cpu_default setting.
18186 2013-06-03 Teresa Johnson <tejohnson@google.com>
18188 * dumpfile.c (opt_info_switch_p): Change -fopt-info
18189 default to -fopt-info=optimized instead of all.
18190 * doc/invoke.texi: Ditto.
18191 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
18192 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
18193 (execute_vect_slp): Emit BB vectorization success under
18194 MSG_OPTIMIZED_LOCATIONS.
18195 * tree-vect-slp.c (vect_slp_transform_bb): Change
18196 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
18197 * tree-vect-loop.c (vect_transform_loop): Ditto.
18199 2013-06-03 Jason Merrill <jason@redhat.com>
18202 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
18203 Use TARGET_EXPR for C++.
18205 2013-06-03 Jakub Jelinek <jakub@redhat.com>
18207 PR rtl-optimization/57268
18208 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
18209 if DEBUG_INSN_P (insn).
18212 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18214 PR rtl-optimization/57268
18215 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18216 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18218 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18220 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18221 (ix86_avoid_lea_for_addr): Likewise.
18222 (exact_dependency_1): Likewise.
18223 (ix86_adjust_cost): Likewise.
18224 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18226 (do_reorder_for_imul): Fix typo, formatting and
18227 !reload_completed check removed.
18228 (ix86_sched_reorder): Fix typo and formatting.
18229 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18232 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18234 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18236 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18238 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18239 <STRING_CST>: Likewise.
18240 <VECTOR_CST>: Likewise.
18242 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18243 Mikael Morin <mikael@gcc.gnu.org>
18245 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18246 * config.in: Regenerated.
18247 * configure: Regenerated.
18249 2013-06-01 Jan Hubicka <jh@suse.cz>
18251 PR middle-end/57366
18252 * cgraphunit.c (compile): When weakref is not supported,
18253 set up transparent aliases before final output pass.
18254 * varasm.c (assemble_alias): Do not try to do it here.
18256 2013-06-01 Jan Hubicka <jh@suse.cz>
18258 PR middle-end/57467
18259 * passes.c (for_per_function): Skip unanalyzed functions.
18261 2013-06-01 Jan Hubicka <jh@suse.cz>
18263 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18264 (lto_symtab_merge_symbols_1): ... this one.
18265 (lto_symtab_merge_cgraph_nodes): Rename to ...
18266 (lto_symtab_merge_symbols): ... this one; simplify.
18267 * cgraph.c (same_body_aliases_done): Rename to ...
18268 (cpp_implicit_aliases_done): ... this one.
18269 (cgraph_create_function_alias): Update.
18270 (cgraph_same_body_alias): Update.
18271 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18272 (verify_edge_corresponds_to_fndecl): Simplify.
18273 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18274 (cgraph_node): Remove same_body_alias.
18275 (varpool_node): Remove alias_of and extra_name_alias.
18276 (same_body_aliases_done): Rename to ..
18277 (cpp_implicit_aliases_done): ... this one.
18278 (symtab_alias_ultimate_target): Add default parameter.
18279 (symtab_resolve_alias): New function.
18280 (fixup_same_cpp_alias_visibility): Declare.
18281 (cgraph_function_node): Add default parameter.
18282 (cgraph_node_asm_name): Likewise.
18283 (cgraph_function_or_thunk_node): Add default parameter; do
18284 not ICE when it is NULL.
18285 (varpool_variable_node): Likewise.
18286 * tree-emutls.c (create_emultls_var): Update.
18287 (ipa_lower_emutls): Update.
18288 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18289 (cgraph_reset_node): Reset alias info.
18290 (cgraph_finalize_function): Update.
18291 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18292 (analyze_function): Simplify.
18293 (cgraph_process_same_body_aliases): Simplify.
18294 (analyze_functions): Fixup same body aliases.
18295 (handle_alias_pairs): Simplify.
18296 (assemble_thunk): Update.
18297 (assemble_thunks_and_aliases): Update.
18298 (output_weakrefs): Rewrite.
18299 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18300 (lto_output_varpool_node): Likewise.
18301 (compute_ltrans_boundary): Remve assert.
18302 (get_alias_symbol): New functoin.
18303 (input_node): Rewrite alias handling.
18304 (input_varpool_node): Likewise.
18305 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18306 * ipa.c (process_references): Handle weakrefs correctly.
18307 (symtab_remove_unreachable_nodes): Likewise.
18308 * trans-mem.c (get_cg_data): Update.
18309 (ipa_tm_create_version_alias): Update.
18310 (ipa_tm_execute): Update.
18311 * symtab.c (dump_symtab_base): Dump aliases.
18312 (verify_symtab_base): Verify aliases.
18313 (symtab_node_availability): New function.
18314 (symtab_alias_ultimate_target): Simplify.
18315 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18316 handle all the fixup cases.
18317 (symtab_resolve_alias): New function.
18318 * passes.c (ipa_write_summaries): Handle weakrefs.
18319 * varpool.c (varpool_analyze_node): Simplify.
18320 (assemble_aliases): Update.
18321 (varpool_create_variable_alias): Simplify.
18322 (varpool_extra_name_alias): Simplify.
18323 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18324 (lto_symtab_merge_symbols): ... this one.
18326 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18329 PR rtl-optimization/57268
18330 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18331 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18333 2013-06-01 Tobias Burnus <burnus@net-b.de>
18335 Partially reverted:
18336 2013-05-31 Tobias Burnus <burnus@net-b.de>
18338 PR middle-end/57073
18339 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18342 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18344 PR rtl-optimization/57268
18345 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18346 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18348 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18350 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18351 unordered comparison operators when -fno-trapping-math is in effect
18353 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18354 and implement unordered comparison operators properly on the e500.
18356 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18358 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18359 for constant scalar integers.
18360 (simplify_relational_operation_1): Likewise.
18362 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18364 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18365 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18368 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18369 Igor Zamyatin <igor.zamyatin@intel.com>
18371 Silvermont (SLM) architecture performance tuning.
18372 * config/i386/i386.h (enum ix86_tune_indices): Add
18373 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18374 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18376 * config/i386/i386.c (initial_ix86_tune_features)
18377 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18378 (ix86_lea_outperforms): Handle Silvermont tuning.
18379 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18381 (ix86_use_lea_for_mov): Likewise.
18382 (ix86_avoid_lea_for_addr): Likewise.
18383 (ix86_lea_for_add_ok): Likewise.
18384 (exact_dependency_1): New function.
18385 (exact_store_load_dependency): Likewise.
18386 (ix86_adjust_cost): Handle Silvermont tuning.
18387 (do_reoder_for_imul): Likewise.
18388 (swap_top_of_ready_list): New function.
18389 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18391 * config/i386/i386.md (peepholes that split memory operand in fp
18394 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18396 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18397 Remove un-necessary braces.
18399 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18401 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18402 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18404 2013-05-31 Tobias Burnus <burnus@net-b.de>
18406 PR middle-end/57073
18407 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18409 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18412 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18413 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18414 * config/arm/neon.md (iordi3_neon): Remove.
18415 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18416 * config/arm/predicates.md (imm_for_neon_logic_operand):
18417 Move to earlier in the file.
18418 (neon_logic_op2): Likewise.
18419 (arm_iordi_operand_neon): New predicate.
18421 2013-05-31 Richard Biener <rguenther@suse.de>
18423 PR tree-optimization/57478
18424 PR tree-optimization/57453
18425 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18428 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18430 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18431 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18433 2013-05-30 Tobias Burnus <burnus@net-b.de>
18434 Thomas Koenig <tkoenig@gcc.gnu.org>
18436 PR middle-end/57073
18437 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18438 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18440 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18442 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18444 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18446 * target.def (register_usage_leveling_p): New hook.
18447 * targhooks.c (default_register_usage_leveling_p): New.
18448 * targhooks.h (default_register_usage_leveling_p): New prototype.
18449 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18450 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18451 * doc/tm.texi: Update.
18452 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18454 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18456 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18457 (*insv_reg<mode>): New define_insn.
18459 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18461 PR rtl-optimization/57439
18462 * postreload.c (move2add_valid_value_p): Check that we have
18463 a zero subreg_regno_offset when accessing the register in
18464 the requested mode.
18466 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18467 Igor Zamyatin <igor.zamyatin@intel.com>
18469 Silvermont (SLM) architecture pipeline model, tuning and
18471 * config.gcc: Add slm config options and target.
18473 * config/i386/slm.md: New.
18475 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18477 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18479 (ix86_target_macros_internal): Likewise.
18481 * config/i386/i386.c (slm_cost): New cost.
18482 (m_SLM): New macro flag.
18483 (initial_ix86_tune_features): Set m_SLM.
18484 (x86_accumulate_outgoing_args): Likewise.
18485 (x86_arch_always_fancy_math_387): Likewise.
18486 (processor_target_table): Add slm cost.
18487 (cpu_names): Add slm cpu name.
18488 (x86_option_override_internal): Set SLM ISA.
18489 (ix86_issue_rate): New case PROCESSOR_SLM.
18490 (ia32_multipass_dfa_lookahead): Likewise.
18491 (fold_builtin_cpu): Add slm.
18493 * config/i386/i386.h (TARGET_SLM): New target macro.
18494 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18495 (processor_type): Add PROCESSOR_SLM.
18497 * config/i386/i386.md (cpu): Add new value "slm".
18498 (slm.md): Include slm.md.
18500 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18501 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18503 * config/arm/arm-protos.h: Add and update function protos.
18504 * config/arm/arm.c (use_simple_return_p): New added.
18505 (thumb2_expand_return): Check simple_return flag.
18506 * config/arm/arm.md: Add simple_return and conditional simple_return.
18507 * config/arm/iterators.md: Add iterator for return and simple_return.
18509 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18511 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18512 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18513 (arm_emit_vfp_multi_reg_pop): Likewise.
18514 (thumb2_emit_ldrd_pop): Likewise.
18515 (arm_expand_epilogue): Add misc REG_CFA notes.
18516 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18518 2013-05-29 Lawrence Crowl <crowl@google.com>
18520 * config/arm/t-arm: Update for below.
18522 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18523 Change type to hash_table. Update dependent calls and types.
18525 * config/i386/t-cygming: Update for below.
18527 * config/i386/t-interix: Update for below.
18529 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18530 Change type to hash_table. Update dependent calls and types.
18531 (i386_find_on_wrapper_list::wrappers): Likewise.
18533 * config/ia64/t-ia64: Update for below.
18535 * config/ia64/ia64.c (bundle_state_table):
18536 Change type to hash_table. Update dependent calls and types.
18538 * config/mips/mips.c (mips_reorg_process_insns::htab):
18539 Change type to hash_table. Update dependent calls and types.
18541 * config/sol2.c (solaris_comdat_htab):
18542 Change type to hash_table. Update dependent calls and types.
18544 * config/t-sol2: Update for above.
18546 2013-05-29 Teresa Johnson <tejohnson@google.com>
18548 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18549 functions are not yet marked as defined.
18551 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18552 Pat Haugen <pthaugen@us.ibm.com>
18553 Peter Bergner <bergner@vnet.ibm.com>
18555 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18559 (vrotl<mode>3): Likewise.
18560 (vashl<mode>3): Likewise.
18561 (vlshr<mode>3): Likewise.
18562 (vashr<mode>3): Likewise.
18564 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18565 support for power8 V2DI builtins.
18567 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18568 power8 V2DI builtins.
18569 (vupkhsw): Likewise.
18570 (vupklsw): Likewise.
18571 (vaddudm): Likewise.
18572 (vminsd): Likewise.
18573 (vmaxsd): Likewise.
18574 (vminud): Likewise.
18575 (vmaxud): Likewise.
18576 (vpkudum): Likewise.
18577 (vpksdss): Likewise.
18578 (vpkudus): Likewise.
18579 (vpksdus): Likewise.
18584 (vsubudm): Likewise.
18585 (vcmpequd): Likewise.
18586 (vcmpgtsd): Likewise.
18587 (vcmpgtud): Likewise.
18588 (vcmpequd_p): Likewise.
18589 (vcmpgtsd_p): Likewise.
18590 (vcmpgtud_p): Likewise.
18591 (vupkhsw): Likewise.
18592 (vupklsw): Likewise.
18593 (vaddudm): Likewise.
18594 (vmaxsd): Likewise.
18595 (vmaxud): Likewise.
18596 (vminsd): Likewise.
18597 (vminud): Likewise.
18598 (vpksdss): Likewise.
18599 (vpksdus): Likewise.
18600 (vpkudum): Likewise.
18601 (vpkudus): Likewise.
18606 (vsubudm): Likewise.
18608 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18609 support for power8 V2DI instructions.
18611 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18612 power8 V2DI instructions. Combine pack and unpack insns to use an
18613 iterator for each mode. Check whether a particular mode supports
18614 Altivec instructions instead of just checking TARGET_ALTIVEC.
18615 (UNSPEC_VPKUWUM): Likewise.
18616 (UNSPEC_VPKSHSS): Likewise.
18617 (UNSPEC_VPKSWSS): Likewise.
18618 (UNSPEC_VPKUHUS): Likewise.
18619 (UNSPEC_VPKSHUS): Likewise.
18620 (UNSPEC_VPKUWUS): Likewise.
18621 (UNSPEC_VPKSWUS): Likewise.
18622 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18623 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18624 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18625 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18626 (UNSPEC_VUPKHSB): Likewise.
18627 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18628 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18629 (UNSPEC_VUPKHSH): Likewise.
18630 (UNSPEC_VUPKLSB): Likewise.
18631 (UNSPEC_VUPKLSH): Likewise.
18633 (VI_char): Likewise.
18634 (VI_scalar): Likewise.
18635 (VI_unit): Likewise.
18637 (VP_small): Likewise.
18638 (VP_small_lc): Likewise.
18639 (VU_char): Likewise.
18640 (add<mode>3): Likewise.
18641 (altivec_vaddcuw): Likewise.
18642 (altivec_vaddu<VI_char>s): Likewise.
18643 (altivec_vadds<VI_char>s): Likewise.
18644 (sub<mode>3): Likewise.
18645 (altivec_vsubcuw): Likewise.
18646 (altivec_vsubu<VI_char>s): Likewise.
18647 (altivec_vsubs<VI_char>s): Likewise.
18648 (altivec_vavgs<VI_char>): Likewise.
18649 (altivec_vcmpbfp): Likewise.
18650 (altivec_eq<mode>): Likewise.
18651 (altivec_gt<mode>): Likewise.
18652 (altivec_gtu<mode>): Likewise.
18653 (umax<mode>3): Likewise.
18654 (smax<mode>3): Likewise.
18655 (umin<mode>3): Likewise.
18656 (smin<mode>3): Likewise.
18657 (altivec_vpkuhum): Likewise.
18658 (altivec_vpkuwum): Likewise.
18659 (altivec_vpkshss): Likewise.
18660 (altivec_vpkswss): Likewise.
18661 (altivec_vpkuhus): Likewise.
18662 (altivec_vpkshus): Likewise.
18663 (altivec_vpkuwus): Likewise.
18664 (altivec_vpkswus): Likewise.
18665 (altivec_vpks<VI_char>ss): Likewise.
18666 (altivec_vpks<VI_char>us): Likewise.
18667 (altivec_vpku<VI_char>us): Likewise.
18668 (altivec_vpku<VI_char>um): Likewise.
18669 (altivec_vrl<VI_char>): Likewise.
18670 (altivec_vsl<VI_char>): Likewise.
18671 (altivec_vsr<VI_char>): Likewise.
18672 (altivec_vsra<VI_char>): Likewise.
18673 (altivec_vsldoi_<mode>): Likewise.
18674 (altivec_vupkhsb): Likewise.
18675 (altivec_vupkhs<VU_char>): Likewise.
18676 (altivec_vupkls<VU_char>): Likewise.
18677 (altivec_vupkhsh): Likewise.
18678 (altivec_vupklsb): Likewise.
18679 (altivec_vupklsh): Likewise.
18680 (altivec_vcmpequ<VI_char>_p): Likewise.
18681 (altivec_vcmpgts<VI_char>_p): Likewise.
18682 (altivec_vcmpgtu<VI_char>_p): Likewise.
18683 (abs<mode>2): Likewise.
18684 (vec_unpacks_hi_v16qi): Likewise.
18685 (vec_unpacks_hi_v8hi): Likewise.
18686 (vec_unpacks_lo_v16qi): Likewise.
18687 (vec_unpacks_hi_<VP_small_lc>): Likewise.
18688 (vec_unpacks_lo_v8hi): Likewise.
18689 (vec_unpacks_lo_<VP_small_lc>): Likewise.
18690 (vec_pack_trunc_v8h): Likewise.
18691 (vec_pack_trunc_v4si): Likewise.
18692 (vec_pack_trunc_<mode>): Likewise.
18694 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
18696 (vec_vmaxsd): Likewise.
18697 (vec_vmaxud): Likewise.
18698 (vec_vminsd): Likewise.
18699 (vec_vminud): Likewise.
18700 (vec_vpksdss): Likewise.
18701 (vec_vpksdus): Likewise.
18702 (vec_vpkudum): Likewise.
18703 (vec_vpkudus): Likewise.
18704 (vec_vrld): Likewise.
18705 (vec_vsld): Likewise.
18706 (vec_vsrad): Likewise.
18707 (vec_vsrd): Likewise.
18708 (vec_vsubudm): Likewise.
18709 (vec_vupkhsw): Likewise.
18710 (vec_vupklsw): Likewise.
18712 2013-05-29 Jan Hubicka <jh@suse.cz>
18714 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
18715 flags; reorder rest of fields in more consistent way.
18716 (varpool_node): Remove analyzed, finalized and alias.
18717 (cgraph_ndoe): Likewise.
18718 (symtab_alias_ultimate_target): New function.
18719 (cgraph_function_node): Move offline.
18720 (cgraph_reset_node): Declare.
18721 (cgraph_comdat_can_be_unshared_p): Remove.
18722 (varpool_remove_initializer): Declare.
18723 (varpool_first_defined_variable, varpool_next_defined_variable
18724 cgraph_first_defined_function, cgraph_next_defined_function): Update.
18725 (cgraph_function_with_gimple_body_p): Update.
18726 (varpool_all_refs_explicit_p): Update.
18727 (symtab_alias_target): New function.
18728 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
18729 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
18730 (cgraph_function_or_thunk_node): Simplify using
18731 symtab_alias_ultimate_target.
18732 (varpool_variable_node): Likewise.
18733 * cgraph.c (cgraph_create_function_alias): Update.
18734 (cgraph_add_thunk): Update.
18735 (cgraph_remove_node): Update.
18736 (dump_cgraph_node): Do not dump removed flags.
18737 (cgraph_function_body_availability): Update.
18738 (cgraph_propagate_frequency): Update.
18739 (verify_cgraph_node): Check sanity of local flag.
18740 (cgraph_function_node): Move here from cgraph.h; revamp for
18741 cgraph_function_or_thunk_node.
18742 * lto-symtab.c (lto_varpool_replace_node): Update.
18743 (lto_symtab_resolve_can_prevail_p): Update.
18744 (lto_symtab_merge_cgraph_nodes): Update.
18745 * ipa-cp.c (determine_versionability, initialize_node_lattices,
18746 propagate_constants_accross_call, devirtualization_time_bonus,
18747 ipcp_propagate_stage): Update.
18748 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
18749 * ipa-inline-transform.c (clone_inlined_nodes,
18750 preserve_function_body_p): Update.
18751 * ipa-reference.c (propagate): Update.
18752 (write_node_summary_p): Update.
18753 * toplev.c (wrapup_global_declaration_2): Update.
18754 * cgraphunit.c (cgraph_analyze_function): Rename to ...
18755 (analyze_function) ... this one.
18756 (cgraph_process_new_functions): Update.
18757 (cgraph_reset_node): Export.
18758 (cgraph_finalize_function): Update.
18759 (cgraph_add_new_function): Update.
18760 (process_function_and_variable_attributes): Update.
18761 (varpool_finalize_decl): Update.
18762 (symbol_finalized): Remove.
18763 (symbol_finalized_and_needed): Rename to ...
18764 (symbol_defined_and_needed): ... update.
18765 (cgraph_analyze_functions): Update.
18766 (handle_alias_pairs): Update.
18767 (mark_functions_to_output): Update.
18768 (assemble_thunk): Update.
18769 (output_in_order): Update.
18770 (output_weakrefs): Update.
18771 (finalize_compilation_unit): Update.
18772 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
18773 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
18774 input_node, input_varpool_node): Update.
18775 * dbxout.c (dbxout_expand_expr): Update.
18776 * cgraphclones.c (cgraph_clone_node): Update.
18777 (cgraph_copy_node_for_versioning): Update.
18778 (cgraph_materialize_clone): Update.
18779 (cgraph_materialize_all_clones): Update.
18780 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
18781 propagate_pure_const, propagate_nothrow): Update.
18782 * lto-streamer-out.c (lto_output, write_symbol): Update.
18783 * ipa-utils.c (ipa_reverse_postorder): Update.
18784 * ipa-inline.c (can_inline_edge_p): Update.
18785 (update_caller_keys, ipa_inline): Update.
18786 * dwarf2out.c (reference_to_unused,
18787 premark_types_used_by_global_vars_helper): Update.
18788 * tree-eh.c (tree_could_trap_p): Update.
18789 * ipa-split.c (consider_split, execute_split_functions): Update.
18790 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
18791 has_addr_references_p): Update; move ahead in file for better
18793 (process_references): Simplify.
18794 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
18795 bodies are removed.
18796 (cgraph_comdat_can_be_unshared_p): Make static.
18797 (cgraph_externally_visible_p): Update.
18798 (varpool_externally_visible_p): Update.
18799 (function_and_variable_visibility): Update.
18800 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
18801 ipa_tm_mark_force_output_node): Update.
18802 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
18803 estimate_edge_devirt_benefit, inline_generate_summary,
18804 inline_write_summary): Update.
18805 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
18806 * ipa-prop.c (ipa_compute_jump_functions): Update.
18807 (ipa_print_node_params, ipa_prop_read_section,
18808 ipa_update_after_lto_read, read_replacements_section): Update.
18809 * varasm.c (mark_decl_referenced): Update.
18810 (assemble_alias, dump_tm_clone_pairs): Update.
18811 * tree-inline.c (copy_bb): Update.
18812 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
18814 * symtab.c (dump_symtab_base): Print new flags.
18815 (verify_symtab_base): Verify new flags.
18816 (symtab_alias_ultimate_target): New function.
18817 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18818 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18820 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18822 * i386.c (ix86_get_function_versions_dispatcher,
18823 ix86_generate_version_dispatcher_body): Update.
18824 (fold_builtin_cpu): Use varpool_add_new_variable.
18825 * varpool.c (varpool_remove_initializer): Break out from ...
18826 (varpool_remove_node): ... this one.
18827 (dump_varpool_node, varpool_node_for_asm,
18828 cgraph_variable_initializer_availability, varpool_analyze_node,
18829 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18830 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18833 2013-05-29 Jan Hubicka <jh@suse.cz>
18835 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18837 2013-05-29 Easwaran Raman <eraman@google.com>
18839 PR tree-optimization/57442
18840 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18841 when control exits the main loop.
18843 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18845 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18847 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18848 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18849 * rx/t-rx: Add rx100 under multi library matches option for nofpu
18852 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18854 PR tree-optimization/57441
18855 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
18856 Don't limit size of incr_vec to number of candidates.
18858 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
18860 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
18861 and mips16 directories.
18862 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
18863 (MULTILIB_DIRNAMES): Ditto.
18864 (MULTILIB_EXCEPTIONS): Add new exceptions.
18865 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
18866 (MULTILIB_DIRNAMES): Ditto.
18867 (MULTILIB_EXCEPTIONS): Add new exceptions.
18869 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18870 Marcus Shawcroft <marcus.shawcroft@arm.com>
18872 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
18873 SYMBOL_TINY_ABSOLUTE.
18874 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
18875 SYMBOL_TINY_ABSOLUTE.
18876 (aarch64_expand_mov_immediate): Likewise.
18877 (aarch64_classify_symbol): Likewise.
18878 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
18879 Permit SYMBOL_TINY_ABSOLUTE.
18880 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
18882 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18883 Marcus Shawcroft <marcus.shawcroft@arm.com>
18885 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
18886 Refactor if/switch. Replace gcc_assert with if.
18888 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18890 * config/i386/i386.c (initial_ix86_tune_features): Enable
18891 FP Reassociation for AMD bdver1 and bdver2.
18893 2013-05-29 Martin Jambor <mjambor@suse.cz>
18895 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
18896 and IMAGPART_EXPR do not occur within other handled_components.
18898 2013-05-29 Richard Biener <rguenther@suse.de>
18900 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
18901 access on whether the use is in the BB we currently try to
18903 (vect_bb_vectorization_profitable_p): Pass the BB we currently
18904 vectorize to vect_bb_slp_scalar_cost.
18906 2013-05-29 Richard Biener <rguenther@suse.de>
18908 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
18909 computing scalar cost offsetted by stmts that are kept live
18911 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
18912 for computation of scalar cost.
18914 2013-05-28 Steve Ellcey <sellcey@mips.com>
18916 * config/mips/mips-cpus.def (mips32r2): Change processor type.
18918 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18920 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
18921 array notation built-in reduction functions.
18922 * doc/passes.texi (Passes): Added documentation about changes done
18924 * doc/invoke.texi (C Dialect Options): Added documentation about
18925 the -fcilkplus flag.
18926 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
18927 (BUILTINS_DEF): Depend on cilkplus.def.
18928 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
18929 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
18930 * cilkplus.def: New file.
18932 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
18934 PR rtl-optimization/57439
18935 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
18937 2013-05-28 Easwaran Raman <eraman@google.com>
18939 PR tree-optimization/57337
18940 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
18941 (find_insert_point): Correctly identify the insertion point
18942 when two statements with the same UID is compared.
18944 2013-05-28 Richard Biener <rguenther@suse.de>
18946 PR tree-optimization/56787
18947 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
18948 from the list of data references.
18949 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
18951 (vect_analyze_loop_operations): Likewise.
18952 (vect_transform_loop): Remove clobbers.
18954 2013-05-28 Martin Jambor <mjambor@suse.cz>
18956 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
18957 and REALPART_EXPRs have scalar type.
18959 2013-05-28 Richard Biener <rguenther@suse.de>
18961 PR tree-optimization/57411
18962 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
18964 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
18965 virtual operand propagation.
18967 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18969 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
18970 destination register for bmasksi_vis.
18971 (vector_init_bshuffle): Likewise.
18972 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
18974 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18976 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
18977 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
18978 mode if the instruction isn't available in the original mode.
18979 * config/sparc/sparc.opt (mfix-ut699): New option.
18980 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
18981 (divdf3): Turn into expander.
18982 (divdf3_nofix): New insn.
18983 (divdf3_fix): Likewise.
18984 (divsf3): Disable if -mfix-ut699.
18985 (sqrtdf2): Turn into expander.
18986 (sqrtdf2_nofix): New insn.
18987 (sqrtdf2_fix): Likewise.
18988 (sqrtsf2): Disable if -mfix-ut699.
18990 2013-05-27 Richard Biener <rguenther@suse.de>
18992 PR middle-end/57412
18993 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
18994 block for the new loop.
18996 2013-05-27 Richard Biener <rguenther@suse.de>
18998 PR tree-optimization/57343
18999 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
19000 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
19001 (number_of_iterations_cond): Do not build the folded tree.
19003 2013-05-27 Richard Biener <rguenther@suse.de>
19006 PR middle-end/57381
19007 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19008 OEP_CONSTANT_ADDRESS_OF retained.
19010 PR tree-optimization/57417
19011 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
19012 for unchanged base.
19013 (set_ssa_val_to): Compare addresses using
19014 get_addr_base_and_unit_offset.
19016 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19018 PR rtl-optimization/56833
19019 * postreload.c (move2add_record_mode): New function.
19020 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19021 (move2add_use_add2_insn): Use move2add_record_sym_value.
19022 (move2add_use_add3_insn): Likewise.
19023 (reload_cse_move2add): Use move2add_valid_value_p and
19024 move2add_record_mode. Invalidate call-clobbered and REG_INC
19025 affected regs by setting reg_mode to VOIDmode.
19026 (move2add_note_store): Don't pretend the inside of a SUBREG is
19027 the actual destination. Invalidate single/leading registers by
19028 setting reg_mode to VOIDmode.
19029 Use move2add_record_sym_value, move2add_valid_value_p and
19030 move2add_record_mode.
19032 2013-05-27 Richard Biener <rguenther@suse.de>
19034 PR tree-optimization/57396
19035 * tree-affine.c (double_int_constant_multiple_p): Properly
19036 return false for val == 0 and div != 0.
19038 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19040 * config/mips/mips.h: Use #elif in preprocessor conditions.
19042 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19045 * config/mips/constraints.md (kl): New constraint.
19046 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19047 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19048 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19049 from LO for MIPS16.
19050 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19052 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19055 * config/mips/mips.c (mips_can_inline_p): New function.
19056 (TARGET_CAN_INLINE_P): Define.
19058 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19060 * sched-int.h (ds_t, dw_t): Make unsigned int.
19061 Fix documentation that describes how all the ds_t bits are used.
19062 Reserve the last bit for delayed-branch scheduling.
19063 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19064 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19065 (gen_dep_weak_1): Remove prototype.
19066 * sched-deps.c (get_dep_weak_1): Make static.
19067 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19068 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19069 * doc/tm.texi: Regenerate.
19070 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19072 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19075 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19077 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19078 Sandra Loosemore <sandra@codesourcery.com>
19080 * config.gcc (powerpc-*): Allow native for with-cpu.
19082 2013-05-24 Jeff Law <law@redhat.com>
19084 PR tree-optimization/57124
19085 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19086 conversion feeding a condition if the range has an overflow
19087 if -fstrict-overflow. Add warnings for when we do make the
19090 2013-05-24 Dehao Chen <dehao@google.com>
19092 * tree-cfg.c (locus_discrim_map): Fix the typo.
19093 (locus_discrim_hasher): Likewise.
19094 (locus_discrim_hasher::hash): Likewise.
19095 (locus_discrim_hasher::equal): Likewise.
19097 2013-05-24 Martin Jambor <mjambor@suse.cz>
19099 PR tree-optimization/57294
19100 * cgraph.h (ipa_record_stmt_references): Declare.
19101 * cgraphbuild.c (ipa_record_stmt_references): New function.
19102 (build_cgraph_edges): Use ipa_record_stmt_references.
19103 (rebuild_cgraph_edges): Likewise.
19104 (cgraph_rebuild_references): Likewise.
19105 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19106 associated with the old statement and build references from the
19107 newly built statements.
19108 * ipa-ref.c (ipa_remove_stmt_references): New function.
19109 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19111 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19113 * lra-constraints.c (emit_spill_move): Use smaller mode for
19115 (check_and_process_move): Consider mem-reg moves for secondary
19117 (curr_insn_transform): Don't lose insns emitted before for
19118 secondary memory moves.
19119 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19120 reg set up in the current insn.
19122 2013-05-24 Dehao Chen <dehao@google.com>
19124 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19126 (locus_descrim_hasher::equal): Likewise.
19127 (build_gimple_cfg): New discriminator assignment algorithm.
19128 (make_edges): Likewise.
19129 (next_discriminator_for_locus): Likewise.
19130 (same_line_p): Likewise.
19131 (assign_discriminators): Likewise.
19132 (make_cond_expr_edges): Likewise.
19133 (make_gimple_switch_edges): Likewise.
19134 (make_goto_expr_edges): Likewise.
19135 (make_gimple_asm_edges): Likewise.
19137 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19139 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19140 X format specifier to only display bottom 16 bits.
19141 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19142 immediate to match for operand 2, since it will be masked.
19144 2013-05-24 Richard Biener <rguenther@suse.de>
19146 PR tree-optimization/57287
19147 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19148 all SSA names that occur in abnormal PHIs.
19150 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19153 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19154 that index is not negative.
19156 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19158 PR rtl-optimization/55177
19159 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19160 (simplify_byte_swapping_operation): New.
19161 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
19162 (simplify_relational_operation_1): Deal with BSWAP.
19164 2013-05-23 Richard Henderson <rth@redhat.com>
19167 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
19168 (ix86_reorg): Call it.
19170 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
19173 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
19174 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
19175 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
19177 2013-05-23 Christian Bruel <christian.bruel@st.com>
19180 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
19182 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19183 Marcus Shawcroft <marcus.shawcroft@arm.com>
19185 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
19186 * config/aarch64/constraints.md (Usa): Remove.
19187 * doc/md.texi (AArch64 Usa): Remove.
19189 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19190 Marcus Shawcroft <marcus.shawcroft@arm.com>
19192 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
19193 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
19194 * config/aarch64/predicates.md (aarch64_const_address): Remove.
19195 (aarch64_mov_operand): Use aarch64_mov_operand_p.
19197 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
19199 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
19200 instruction (AdvSIMD).
19201 * config/aarch64/aarch64-builtins.c
19202 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
19203 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
19205 2013-05-23 Martin Jambor <mjambor@suse.cz>
19207 PR middle-end/57347
19208 * tree.h (contains_bitfld_component_ref_p): Declare.
19209 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19210 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19212 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19213 not access a bit-field. Assert all final offsets are byte-aligned.
19215 2013-05-23 Richard Biener <rguenther@suse.de>
19217 PR tree-optimization/57380
19218 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19219 least one invariant or re-used load.
19220 * passes.c (init_optimization_passes): Move pass_phiprop before
19223 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19225 * config/aarch64/aarch64-simd.md
19226 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19228 2013-05-23 Richard Biener <rguenther@suse.de>
19230 PR middle-end/57381
19231 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19232 OEP_CONSTANT_ADDRESS_OF retained.
19234 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19236 PR middle-end/57344
19237 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19238 don't lower unit. Handle unit not being always BITS_PER_WORD.
19240 2013-05-23 Richard Biener <rguenther@suse.de>
19242 PR rtl-optimization/57341
19243 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19244 instead of true_dependence.
19246 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19248 * bb-reorder.c (branch_threshold): Make const.
19249 (exec_threshold): Ditto.
19251 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19252 Pat Haugen <pthaugen@us.ibm.com>
19253 Peter Bergner <bergner@vnet.ibm.com>
19255 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19256 documentation for the power8 crypto builtins.
19258 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19260 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19261 macros for defining power8 builtin functions.
19262 (BU_P8V_AV_2): Likewise.
19263 (BU_P8V_AV_P): Likewise.
19264 (BU_P8V_VSX_1): Likewise.
19265 (BU_P8V_OVERLOAD_1): Likewise.
19266 (BU_P8V_OVERLOAD_2): Likewise.
19267 (BU_CRYPTO_1): Likewise.
19268 (BU_CRYPTO_2): Likewise.
19269 (BU_CRYPTO_3): Likewise.
19270 (BU_CRYPTO_OVERLOAD_1): Likewise.
19271 (BU_CRYPTO_OVERLOAD_2): Likewise.
19272 (XSCVSPDP): Fix typo, point to the correct instruction.
19273 (VCIPHER): Add power8 crypto builtins.
19274 (VCIPHERLAST): Likewise.
19275 (VNCIPHER): Likewise.
19276 (VNCIPHERLAST): Likewise.
19277 (VPMSUMB): Likewise.
19278 (VPMSUMH): Likewise.
19279 (VPMSUMW): Likewise.
19280 (VPERMXOR_V2DI): Likewise.
19281 (VPERMXOR_V4SI: Likewise.
19282 (VPERMXOR_V8HI: Likewise.
19283 (VPERMXOR_V16QI: Likewise.
19284 (VSHASIGMAW): Likewise.
19285 (VSHASIGMAD): Likewise.
19286 (VPMSUM): Likewise.
19287 (VPERMXOR): Likewise.
19288 (VSHASIGMA): Likewise.
19290 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19291 __CRYPTO__ if the crypto instructions are available.
19292 (altivec_overloaded_builtins): Add support for overloaded power8
19295 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19296 support for power8 crypto builtins.
19297 (builtin_function_type): Likewise.
19298 (altivec_init_builtins): Add support for builtins that take vector
19299 long long (V2DI) arguments.
19301 * config/rs6000/crypto.md: New file, define power8 crypto
19304 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19305 Pat Haugen <pthaugen@us.ibm.com>
19306 Peter Bergner <bergner@vnet.ibm.com>
19308 * doc/invoke.texi (Option Summary): Add power8 options.
19309 (RS/6000 and PowerPC Options): Likewise.
19311 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19312 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19313 wm, wn, wr documentation.
19315 * config/rs6000/constraints.md (wm): New constraint for VSX
19316 registers if direct move instructions are enabled.
19317 (wn): New constraint for no registers.
19318 (wq): New constraint for quad word even GPR registers.
19319 (wr): New constraint if 64-bit instructions are enabled.
19320 (wv): New constraint if power8 vector instructions are enabled.
19321 (wQ): New constraint for quad word memory locations.
19323 * config/rs6000/predicates.md (const_0_to_15_operand): New
19324 constraint for 0..15 for crypto instructions.
19325 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19326 as GPR and floating point registers.
19327 (int_reg_operand): New predicate to match only GPR registers.
19328 (base_reg_operand): New predicate to match base registers.
19329 (quad_int_reg_operand): New predicate to match even GPR registers
19330 for quad memory operations.
19331 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19332 operations in both GPR and VSX registers.
19333 (quad_memory_operand): New predicate for quad memory operations.
19334 (reg_or_indexed_operand): New predicate for direct move support.
19336 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19337 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19338 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19339 (POWERPC_MASKS): Add power8 options.
19340 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19343 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19344 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19346 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19347 (-mpower8-fusion): New power8 options.
19348 (-mpower8-fusion-sign): Likewise.
19349 (-mpower8-vector): Likewise.
19350 (-mcrypto): Likewise.
19351 (-mdirect-move): Likewise.
19352 (-mquad-memory): Likewise.
19354 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19355 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19357 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19358 (rs6000_debug_vector_unit): Add p8_vector.
19359 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19360 definitions. Also print fusion state.
19361 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19362 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19363 (rs6000_option_override_internal): Add support for power8.
19364 (rs6000_common_init_builtins): Add debugging for skipped builtins
19365 if -mdebug=builtin.
19366 (rs6000_adjust_cost): Add power8 support.
19367 (rs6000_issue_rate): Likewise.
19368 (insn_must_be_first_in_group): Likewise.
19369 (insn_must_be_last_in_group): Likewise.
19370 (force_new_group): Likewise.
19371 (rs6000_register_move_cost): Likewise.
19372 (rs6000_opt_masks): Likewise.
19374 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19375 power8 capable assembler, default to power7 options.
19376 (TARGET_DIRECT_MOVE): Likewise.
19377 (TARGET_CRYPTO): Likewise.
19378 (TARGET_P8_VECTOR): Likewise.
19379 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19380 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19381 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19382 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19383 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19384 (TARGET_XSCVDPSPN): Likewise.
19385 (TARGET_XSCVSPDPN): Likewsie.
19386 (TARGET_SYNC_HI_QI): Likewise.
19387 (TARGET_SYNC_TI): Likewise.
19388 (MASK_CRYPTO): Likewise.
19389 (MASK_DIRECT_MOVE): Likewise.
19390 (MASK_P8_FUSION): Likewise.
19391 (MASK_P8_VECTOR): Likewise.
19392 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19393 temporary used by some of the direct move instructions to get two FP
19394 temporary registers does not force creation of a stack frame.
19395 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19396 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19397 that any VSX registers are tieable, even if they are also an
19398 Altivec vector mode.
19399 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19400 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19401 (RS6000_BTM_CRYPTO): Likewise.
19402 (RS6000_BTM_COMMON): Likewise.
19404 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19405 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19406 (enum rs6000_vector): Add power8 vector support.
19408 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19412 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19413 (any_sibcall_could_use_r3): this and handle indirect calls.
19414 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19416 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19418 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19420 2013-05-22 Richard Biener <rguenther@suse.de>
19422 PR middle-end/57349
19423 * profile.c (branch_prob): Do not split blocks that are
19424 abnormally receiving from ECF_RETURNS_TWICE functions.
19426 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19428 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19429 address mode. Move pointer mode initialization to the same place.
19431 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19433 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19434 while it has any effect.
19436 2013-05-21 Easwaran Raman <eraman@google.com>
19438 PR tree-optimization/57322
19439 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19440 UID of the statement added to the BB to be 1.
19442 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19444 PR tree-optimization/57331
19445 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19446 of conversion from pointer type to integral type with integer.
19448 2013-05-21 Martin Jambor <mjambor@suse.cz>
19451 * ipa-prop.c (ipa_read_node_info): Process param_used and
19452 controlled_uses in the same order as when writing.
19454 2013-05-21 Magnus Granberg <baldrick@free.fr>
19457 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19459 2013-05-21 Richard Biener <rguenther@suse.de>
19461 PR tree-optimization/57318
19462 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19463 estimate stmts with side-effects as likely eliminated.
19465 2013-05-21 Richard Biener <rguenther@suse.de>
19467 PR tree-optimization/57330
19468 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19469 preserve the call stmts fntype.
19471 2013-05-21 Richard Biener <rguenther@suse.de>
19473 PR tree-optimization/57303
19474 * tree-ssa-sink.c (statement_sink_location): Improve killing
19475 stmt detection and properly handle self-assignments.
19477 2013-05-21 Christian Bruel <christian.bruel@st.com>
19479 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19480 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19481 registers. Set register size out of the PARALLEL loop.
19483 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19486 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19487 (*fmasf4, *fmasf4_media): New insns.
19489 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19491 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19492 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19493 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19494 (mips_idiv_insns): Update the comments to say that the returned
19495 instruction counts are in units of BASE_INSN_LENGTH.
19496 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19497 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19498 using 2 rather than 4 as the length of indirect MIPS16 and
19499 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19500 length of a NOP. Don't divide MIPS16 lengths by 2.
19501 (mips16_split_long_branches): Assume a branch is long if the
19502 length is greater than 4 rather than 8.
19503 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19504 rather than multiplying them by 2. Multiply instruction counts
19505 by BASE_INSN_LENGTH rather than 4.
19506 (*jump_mips16, tls_get_tp_mips16_<mode>)
19507 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19509 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19511 * config/mips/mips.md (extended_mips16): Remove branch case.
19512 (length): Remove duplicated extended_mips16 test.
19514 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19516 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19518 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19520 * recog.h (Recog_data): Rename to...
19521 (recog_data_d): ...this.
19522 (recog_data): Update accordingly.
19523 * recog.c (recog_data): Likewise.
19524 * reload.c (save_recog_data): Likewise.
19525 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19526 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19528 2013-05-17 Julian Brown <julian@codesourcery.com>
19530 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19531 found in a REG_EQUAL note, invalidate it.
19533 2013-05-17 Easwaran Raman <eraman@google.com>
19535 * tree-ssa-reassoc.c (find_insert_point): New function.
19536 (insert_stmt_after): Likewise.
19537 (get_def_stmt): Likewise.
19538 (ensure_ops_are_available): Likewise.
19539 (not_dominated_by): Likewise.
19540 (rewrite_expr_tree): Do not move statements beyond what is
19541 necessary. Remove call to swap_ops_for_binary_stmt...
19542 (reassociate_bb): ... and move it here.
19543 (build_and_add_sum): Assign UIDs for new statements.
19544 (linearize_expr): Likewise.
19545 (do_reassoc): Renumber gimple statement UIDs.
19547 2013-05-17 Jan Hubicka <jh@suse.cz>
19549 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19551 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19552 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19554 (output_weakrefs): Update.
19556 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19557 Martin Jambor <mjambor@suse.cz>
19559 PR middle-end/57276
19560 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19561 value that corresponds to the given aggval is found in values vector.
19563 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19565 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19566 sse, sse2, sse3, ssse3 and sse4a flags to options.
19568 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19570 * gengtype-state.c: (s_expr_writer): New class, to handle
19571 prettifying of output layout of s-expressions.
19572 (state_writer): New class, to write out gtype.state.
19573 (state_written_type_count): Move this variable into member data of
19575 (s_expr_writer::s_expr_writer): New code: constructor for new class
19576 (state_writer::state_writer(): ditto
19577 (s_expr_writer::write_new_line): New function
19578 (s_expr_writer::write_any_indent): ditto
19579 (s_expr_writer::begin_s_expr): ditto
19580 (s_expr_writer::end_s_expr): ditto
19581 (write_state_fileloc): convert to method of state_writer...
19582 (state_writer:: write_state_fileloc): ...and use methods of
19583 s_expr_writer to write indentation into the gtype.state output file
19584 to visually represent the hierarchical structure of the list
19586 (write_state_fields): ditto, renaming to...
19587 (state_writer::write_state_fields)
19588 (write_state_a_string): ditto, renaming to...
19589 (state_writer::write_state_a_string)
19590 (write_state_string_option): ditto, renaming to...
19591 (state_writer::write_state_string_option)
19592 (write_state_type_option): ditto, renaming to...
19593 (state_writer::write_state_type_option)
19594 (write_state_nested_option): ditto, renaming to...
19595 (state_writer::write_state_nested_option)
19596 (write_state_option): ditto, renaming to...
19597 (state_writer::write_state_option)
19598 (write_state_options): ditto, renaming to...
19599 (state_writer::write_state_options)
19600 (write_state_lang_bitmap): ditto, renaming to...
19601 (state_writer::write_state_lang_bitmap)
19602 (write_state_version): ditto, renaming to...
19603 (state_writer::write_state_version)
19604 (write_state_scalar_type): ditto, renaming to...
19605 (state_writer::write_state_scalar_type)
19606 (write_state_string_type): ditto, renaming to...
19607 (state_writer::write_state_string_type)
19608 (write_state_undefined_type): ditto, renaming to...
19609 (state_writer::write_state_undefined_type)
19610 (write_state_struct_union_type): ditto, renaming to...
19611 (state_writer::write_state_struct_union_type)
19612 (write_state_struct_type): ditto, renaming to...
19613 (state_writer::write_state_struct_type)
19614 (write_state_user_struct_type): ditto, renaming to...
19615 (state_writer::write_state_user_struct_type)
19616 (write_state_lang_struct_type): ditto, renaming to...
19617 (state_writer::write_state_lang_struct_type)
19618 (write_state_param_struct_type): ditto, renaming to...
19619 (state_writer::write_state_param_struct_type)
19620 (write_state_pointer_type): ditto, renaming to...
19621 (state_writer::write_state_pointer_type)
19622 (write_state_array_type): ditto, renaming to...
19623 (state_writer::write_state_array_type)
19624 (write_state_gc_used): ditto, renaming to...
19625 (state_writer::write_state_gc_used)
19626 (write_state_common_type_content): ditto, renaming to...
19627 (state_writer::write_state_common_type_content)
19628 (write_state_type): ditto, renaming to...
19629 (state_writer::write_state_type)
19630 (write_state_pair_list): ditto, renaming to...
19631 (state_writer::write_state_pair_list)
19632 (write_state_pair): ditto, renaming to...
19633 (state_writer::write_state_pair)
19634 (write_state_typedefs): ditto, renaming to...
19635 (state_writer::write_state_typedefs)
19636 (write_state_structures): ditto, renaming to...
19637 (state_writer::write_state_structures)
19638 (write_state_param_structs): ditto, renaming to...
19639 (state_writer::write_state_param_structs)
19640 (write_state_variables): ditto, renaming to...
19641 (state_writer::write_state_variables)
19642 (write_state_srcdir): ditto, renaming to...
19643 (state_writer::write_state_srcdir)
19644 (write_state_files_list): ditto, renaming to...
19645 (state_writer::write_state_files_list)
19646 (write_state_languages): ditto, renaming to...
19647 (state_writer::write_state_languages)
19648 (write_state): create a state_writer instance and use it when
19649 writing out the state file
19651 2013-05-17 Mike Stump <mikestump@comcast.net>
19653 PR rtl-optimization/57304
19654 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
19655 accessing DF_REF_REAL_LOC.
19657 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19659 PR rtl-optimization/57281
19660 PR rtl-optimization/57300
19661 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
19662 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
19663 what the other splitter did if the registers are dead.
19665 2013-05-17 Richard Biener <rguenther@suse.de>
19667 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
19670 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19672 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
19675 2013-05-17 Marek Polacek <polacek@redhat.com>
19677 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
19678 length when doing non-zero store of storing '\0' to '\0'.
19680 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19682 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
19683 vect_external_def oprnd1 with loop_vinfo, try to emit
19684 optional cast, negation and and stmts on the loop preheader
19685 edge instead of into the pattern def seq.
19687 PR tree-optimization/57051
19688 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
19689 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
19691 2013-05-16 Nick Clifton <nickc@redhat.com>
19693 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
19694 (rl78_is_naked_func): New function.
19695 (rl78_expand_prologue): Skip prologue generation for naked functions.
19696 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
19697 * doc/extend.texi (naked): Add RL78 to the list of processors
19698 that supports this attribute.
19700 2013-05-16 Jeff Law <law@redhat.com>
19702 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
19704 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19706 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
19707 cache parameters using detect_caches_amd also for CYRIX,
19708 NSC and TM2 signatures.
19710 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19711 Dzianis Kahanovich <mahatma@eu.by>
19715 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
19716 VIA/Centaur processors and determine their cache parameters
19717 using detect_caches_amd.
19719 2013-05-16 Teresa Johnson <tejohnson@google.com>
19721 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
19722 (rtl_verify_edges): New function.
19723 (rtl_verify_bb_insns): Ditto.
19724 (rtl_verify_bb_pointers): Ditto.
19725 (rtl_verify_bb_insn_chain): Ditto.
19726 (rtl_verify_fallthru): Ditto.
19727 (rtl_verify_bb_layout): Ditto.
19728 (rtl_verify_flow_info_1): Outline checks into new functions.
19729 (rtl_verify_flow_info): Ditto.
19731 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
19733 * cfghooks.c (copy_bbs): Add update_dominance argument.
19734 * cfghooks.h (copy_bbs): Update prototype.
19735 * tree-cfg.c (gimple_duplicate_sese_region):
19736 Add update_dominance argument.
19737 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
19738 * tree-ssa-loop-ch.c (copy_loop_headers): Update
19739 gimple_duplicate_sese_region call.
19740 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
19741 Update copy_bbs call.
19742 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
19743 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
19745 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19747 * tree-vectorizer.h (NUM_PATTERNS): Increment.
19748 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
19749 vect_recog_rotate_pattern.
19750 (vect_recog_rotate_pattern): New function.
19752 2013-05-16 Jason Merrill <jason@redhat.com>
19754 * Makefile.in (LLINKER): New variable.
19755 (mostlyclean): Remove link mutex.
19756 * configure.ac: Handle --enable-link-mutex.
19757 * lock-and-run.sh: New script.
19759 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19762 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
19765 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19767 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
19769 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
19771 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
19772 * config/arm/arm.c (next_consecutive_mem): New function.
19773 (gen_movmem_ldrd_strd): Likewise.
19774 * config/arm/arm.md (movmemqi): Update condition and code.
19775 (unaligned_loaddi, unaligned_storedi): New patterns.
19777 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19779 * config.gcc: Obsolete *-*-solaris2.9*.
19780 * doc/install.texi (Specific, *-*-solaris2*): Document it.
19782 2013-05-16 Richard Biener <rguenther@suse.de>
19784 * passes.c (init_optimization_passes): Move pass_parallelize_loops
19785 earlier, after GRAPHITE transforms and IV canonicalization.
19787 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19789 * omp-low.c (extract_omp_for_data): For collapsed loops,
19790 if at least one of the loops is known at compile time to
19791 iterate zero times, set count to 0.
19792 (expand_omp_regimplify_p): New function.
19793 (expand_omp_for_generic): For collapsed loops, if at least
19794 one of the loops isn't known to iterate at least once,
19795 add runtime check with setting count to 0.
19796 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
19797 For unsigned types if it isn't known at compile time that
19798 the loop will iterate at least once, add runtime check to bypass
19799 the whole loop if initial condition isn't true.
19801 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
19803 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
19805 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
19807 PR middle-end/57286
19808 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
19809 transformations to avoid an infinite loop.
19811 2013-05-16 Marek Polacek <polacek@redhat.com>
19813 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
19815 2013-05-15 Leif Ekblad <leif@rdos.net>
19817 * config/i386/i386.c (ix86_decompose_address): Use
19818 DEFAULT_TLS_SEG_REG to access TLS segment register.
19819 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19820 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19821 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19823 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19826 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19827 sibling calls to functions that would normally be lazily bound,
19828 unless $gp is call-clobbered.
19830 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19832 * config/i386/i386.c (ix86_option_override_internal): Update
19833 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19834 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19835 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19836 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19838 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19840 2013-05-15 Andreas Schwab <schwab@suse.de>
19842 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19843 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19846 2013-05-15 Teresa Johnson <tejohnson@google.com>
19848 * loop-unroll.c (report_unroll_peel): Check decision before
19849 emitting unroll/peel message.
19851 2013-05-15 Teresa Johnson <tejohnson@google.com>
19853 * function.h (has_bb_partition): New rtl_data flag.
19854 (bb_reorder_complete): Ditto.
19855 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
19856 instead of flag_reorder_blocks_and_partition.
19857 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
19858 with some enhancements.
19859 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
19860 * bb-reorder.c (connect_traces): Check for has_bb_partition
19861 instead of flag_reorder_blocks_and_partition.
19862 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
19863 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
19864 verify_hot_cold_block_grouping.
19865 (partition_hot_cold_basic_blocks): Set has_bb_partition.
19867 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19870 * config/arm/predicates.md (call_insn_operand): New predicate.
19871 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
19872 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
19873 if insn is not a tail call.
19874 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
19876 * config/arm/arm.h (enum reg_class): New caller save register class.
19877 (REG_CLASS_NAMES): Likewise.
19878 (REG_CLASS_CONTENTS): Likewise.
19879 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
19882 2013-05-15 Richard Biener <rguenther@suse.de>
19884 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
19885 of MSG_OPTIMIZED_LOCATIONS.
19886 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
19887 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
19889 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
19890 of MSG_OPTIMIZED_LOCATIONS.
19891 (execute_vect_slp): Likewise.
19892 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
19893 (vect_create_cond_for_alias_checks): Likewise.
19894 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
19895 (vect_recog_widen_mult_pattern): Likewise.
19896 (vect_recog_widen_sum_pattern): Likewise.
19897 (vect_recog_over_widening_pattern): Likewise.
19898 (vect_recog_widen_shift_pattern): Likewise.
19899 (vect_recog_vector_vector_shift_pattern): Likewise.
19900 (vect_recog_divmod_pattern): Likewise.
19901 (vect_recog_mixed_size_cond_pattern): Likewise.
19902 (vect_recog_bool_pattern): Likewise.
19903 (vect_pattern_recog_1): Likewise.
19905 2013-05-15 Martin Jambor <mjambor@suse.cz>
19907 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
19908 non-functions to builtin_unreachable.
19909 * ipa-inline-transform.c (inline_call): Do not assert estimates were
19910 correct when new direct edges were discovered.
19912 2013-05-15 Martin Jambor <mjambor@suse.cz>
19914 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
19915 header, print symbol order instead of node uid, print more information
19916 about indirect edge targets.
19917 (ipa_make_edge_direct_to_target): Print symbol order instead of node
19919 (ipa_make_edge_direct_to_target): Likewise.
19920 (remove_described_reference): Likewise.
19921 (propagate_controlled_uses): Likewise.
19922 (ipa_print_node_params): Also print symbol order.
19923 (ipcp_transform_function): Print symbol order instead of node uids.
19924 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
19925 (cgraph_get_create_real_symbol_node): Likewise.
19926 * ipa-cp.c (print_lattice): Likewise.
19927 (print_all_lattices): Likewise.
19928 (determine_versionability): Likewise.
19929 (initialize_node_lattices): Likewise.
19930 (estimate_local_effects): Likewise.
19931 (update_profiling_info): Likewise.
19932 (create_specialized_node): Likewise.
19933 (perhaps_add_new_callers): Likewise.
19934 (decide_about_value): Likewise.
19935 (decide_whether_version_node): Likewise.
19936 (identify_dead_nodes): Likewise.
19937 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
19938 (dump_inline_summary): Likewise.
19939 (estimate_node_size_and_time): Likewise.
19940 (inline_analyze_function): Likewise.
19941 * ipa-inline.c (report_inline_failed_reason): Likewise.
19942 (want_early_inline_function_p): Likewise.
19943 (edge_badness): Likewise.
19944 (update_edge_key): Likewise.
19945 (inline_small_functions): Likewise. Add dumping of order to two other
19947 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
19948 instead of node uids.
19949 (propagate_pure_const): Likewise.
19950 (propagate_pure_const): Likewise.
19951 * ipa-utils.c (dump_cgraph_node_set): Likewise.
19952 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
19953 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
19955 * tree-pretty-print.c (dump_function_header): Likewise.
19956 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
19957 Print symbol order instead of node uids.
19959 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19961 * config/s390/s390.c (s390_register_move_cost): Don't impose the
19962 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
19964 2013-05-15 Richard Biener <rguenther@suse.de>
19966 PR tree-optimization/57275
19967 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
19968 return value for fail to do runtime alias checks for gather loads.
19970 2013-05-15 Jan Hubicka <jh@suse.cz>
19974 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
19975 weakrefs are not external.
19976 (lto_symtab_merge_decls): Fix thinko when dealing with
19977 non-lto_symtab decls.
19978 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
19979 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
19980 * varpool.c (dump_varpool_node): Dump more flags.
19982 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19984 * config/i386/i386.c (processor_alias_table): Add instruction
19985 FSGSBASE for AMD bdver3 architecture.
19987 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19989 * tree.c (warn_deprecated_use): Print file:line using locus color.
19990 * diagnostic.c (diagnostic_report_current_module): Print file:line
19991 and file:line:column using locus color.
19993 2013-05-14 Mike Stump <mikestump@comcast.net>
19995 * gdbinit.in: Add __null.
19997 2013-05-14 Mike Stump <mikestump@comcast.net>
19999 * recog.h: Rename struct recog_data to Recog_data.
20000 * recog.c: Likewise.
20001 * reload.c (can_reload_into): Likewise.
20002 * config/picochip/picochip.c: Likewise.
20004 2013-05-14 Mike Stump <mikestump@comcast.net>
20006 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
20008 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
20010 * resource.h (struct resources): Remove unch_memory member.
20011 (CLEAR_RESOURCE): Don't clear unch_memory.
20012 * resource.c (mark_referenced_resources): Don't set it.
20013 (mark_set_resources): Likewise.
20014 (mark_target_live_regs): Don't clear it.
20015 (init_resource_info): Likewise.
20016 * reorg.c (resource_conflicts_p): Don't compare it.
20017 (redundant_insn): Don't set it.
20019 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20021 * emit-rtl.c (next_label): Remove unused function.
20022 (skip_consecutive_labels, link_cc0_insns): Move to ...
20023 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20024 only place where these functions are used, and make them static.
20026 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20028 * fold-const.c (fold_negate_expr): Handle vectors.
20029 (fold_truth_not_expr): Make it static.
20030 (fold_invert_truthvalue): New static function.
20031 (invert_truthvalue_loc): Handle vectors. Do not call
20032 fold_truth_not_expr directly.
20033 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20034 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20035 (fold_comparison): Handle vector constants.
20036 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20037 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20038 * tree.h (fold_truth_not_expr): Remove declaration.
20040 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20042 * config/aarch64/aarch64-simd.md
20043 (aarch64_vcond_internal<mode>): Rename to...
20044 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20045 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20046 float modes. Clarify all iterator modes.
20047 (vcond<mode><mode>): Use new name for vcond expanders.
20048 (vcond<v_cmp_result><mode>): Likewise.
20049 (vcondu<mode><mode>: Likewise.
20050 * config/aarch64/iterators.md (VDQF_COND): New.
20052 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20055 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20056 variable for the shift amount. Check that we shift by non-negative
20059 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20062 * config/arm/arm.h (EPILOGUE_USES): Only return true
20063 for LR_REGNUM after epilogue_completed.
20065 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20067 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20068 is error_mark_node.
20070 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20073 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20074 and Solaris 11+/x86 with gld.
20075 * configure: Regenerate.
20077 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20079 * expmed.c (expand_shift_1): Canonicalize rotates by
20080 constant bitsize / 2 to bitsize - 1.
20081 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20082 case ROTATERT>: Likewise.
20085 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20087 * config/i386/i386.md (rotateinv): New code attr.
20088 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20089 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20090 roll $31, %eax, etc.
20092 2013-05-14 Richard Biener <rguenther@suse.de>
20094 PR middle-end/57235
20095 * tree-eh.c (sink_clobbers): Give up for successors with
20096 multiple predecessors and no virtual uses.
20098 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20100 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20101 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20103 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20105 PR middle-end/57251
20106 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20107 the case when both op0 and op1 have VOIDmode.
20109 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20111 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20112 in multiply-accumulate mode.
20114 2013-05-13 Guozhi Wei <carrot@google.com>
20116 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20118 2013-05-13 Kai Tietz <ktietz@redhat.com>
20121 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20122 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20123 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20124 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20125 (ix86_expand_prologue): Likewise.
20126 (ix86_expand_split_stack_prologue): Likewise.
20127 (legitimate_pic_address_disp_p): Likewise.
20128 (legitimize_pic_address): Likewise.
20129 (legitimize_tls_address): Likewise.
20130 (legitimize_pe_coff_symbol): Likewise.
20131 (output_pic_addr_const): Likewise.
20132 (construct_plt_address): Likewise.
20133 (ix86_expand_call): Likewise.
20134 (x86_output_mi_thunk): Likewise.
20135 (x86_function_profiler): Likewise.
20137 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20139 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20140 similar switch cases.
20141 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20142 (aarch64_simd_mov_to_<mode>low): Delete.
20143 (aarch64_simd_mov_to_<mode>high): Delete.
20144 (move_lo_quad_<mode>): Add w<-r alternative.
20145 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20146 (aarch64_simd_mov_from_*): Update type attribute.
20147 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20150 2013-05-13 Jan Hubicka <jh@suse.cz>
20152 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20153 * config/i386/i386.c (ix86_compute_frame_layout,
20154 ix86_expand_epilogue, emit_i387_cw_initialization,
20155 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20156 ix86_local_alignment): Fix use of size/speed predicates.
20158 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20160 PR tree-optimization/45216
20161 PR tree-optimization/57157
20162 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
20163 the (-Y) & (B - 1) variant if OP is |.
20164 * expmed.c (expand_shift_1): For rotations by const0_rtx just
20165 return shifted. Use (-op1) & (prec - 1) as other_amount
20166 instead of prec - op1.
20168 2013-05-13 Martin Jambor <mjambor@suse.cz>
20170 PR middle-end/42371
20171 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
20172 (ipa_constant_data): New type.
20173 (ipa_jump_func): Use ipa_constant_data to hold information about
20174 constant jump functions.
20175 (ipa_get_jf_constant): Adjust to jump function type changes.
20176 (ipa_get_jf_constant_rdesc): New function.
20177 (ipa_param_descriptor): New field controlled_uses.
20178 (ipa_get_controlled_uses): New function.
20179 (ipa_set_controlled_uses): Likewise.
20180 * ipa-ref.h (ipa_find_reference): Declare.
20181 * ipa-prop.c (ipa_cst_ref_desc): New type.
20182 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
20184 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
20185 New parameter cs. Adjust all callers.
20186 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
20187 (remove_described_reference): New function.
20188 (jfunc_rdesc_usable): Likewise.
20189 (try_make_edge_direct_simple_call): Decrement controlled use count,
20190 attempt to remove reference if it hits zero.
20191 (combine_controlled_uses_counters): New function.
20192 (propagate_controlled_uses): Likewise.
20193 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
20194 (ipa_edge_duplication_hook): Duplicate reference descriptions.
20195 (ipa_print_node_params): Print described use counter.
20196 (ipa_write_jump_function): Adjust to jump function type changes.
20197 (ipa_read_jump_function): New parameter CS, pass it to
20198 ipa_set_jf_constant. Adjust caller.
20199 (ipa_write_node_info): Stream controlled use count
20200 (ipa_read_node_info): Likewise.
20201 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
20203 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
20204 count. Remove cloning-added reference if it reaches zero.
20205 * ipa-ref.c (ipa_find_reference): New function.
20207 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20209 * config/i386/i386.c (processor_target_table): Modified default
20210 alignment values for AMD BD and BT architectures.
20212 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20214 * tree-vect-generic.c (uniform_vector_p): Move ...
20215 * tree.c (uniform_vector_p): ... here.
20216 * tree.h (uniform_vector_p): Declare it.
20217 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20220 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20222 PR tree-optimization/57230
20223 * tree-ssa-strlen.c (handle_char_store): Record length for
20224 array store from STRING_CST.
20226 PR tree-optimization/57230
20227 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20230 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20232 * config/epiphany/epiphany.c (epiphany_init): Check size of
20233 NUM_MODES_FOR_MODE_SWITCHING.
20234 (epiphany_expand_prologue):
20235 Remove CONFIG_REGNUM initial value handling code.
20236 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20237 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20238 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20239 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20240 Don't return 1 for FP_MODE_NONE.
20241 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20242 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20243 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20244 * config/epiphany/epiphany.md (save_config): New pattern.
20246 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20248 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20250 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20252 * config/i386/i386.md (memory): Handle sseishft1.
20253 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20254 (*vec_extractv2di_1): Ditto.
20256 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20258 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20261 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20263 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20264 Add mthumb/march=armv7-a multilib.
20265 Add mthumb/march=armv7-r multilib.
20266 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20268 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20270 * config/v850/t-rtems: Add more multilibs.
20272 2013-05-10 Richard Biener <rguenther@suse.de>
20274 PR tree-optimization/57214
20275 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20276 not propagate from SSA names that occur in abnormal PHI nodes.
20278 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20280 * stor-layout.c (element_precision): New function.
20281 * machmode.h (element_precision): Declare it.
20282 * tree.c (build_minus_one_cst): New function.
20283 (element_precision): Likewise.
20284 * tree.h (build_minus_one_cst): Declare new function.
20285 (element_precision): Likewise.
20286 * fold-const.c (operand_equal_p): Use element_precision.
20287 (fold_binary_loc): Handle vector types.
20288 * convert.c (convert_to_integer): Use element_precision.
20289 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20292 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20294 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20295 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20296 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20297 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20298 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20299 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20300 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20301 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20302 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20303 (Uuw8): New constraints.
20304 (Usb4): Move into alphabetical order.
20305 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20306 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20307 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20308 previously unnamed patterns.
20309 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20310 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20311 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20312 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20313 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20314 of specific lengths.
20316 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20318 * config/i386/i386.md (rotateinv): New code attr.
20319 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20320 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20321 roll $31, %eax, etc.
20323 PR tree-optimization/45216
20324 PR tree-optimization/57157
20325 * tree-ssa-forwprop.c (simplify_rotate): New function.
20326 (ssa_forward_propagate_and_combine): Call it.
20328 2013-05-10 Richard Biener <rguenther@suse.de>
20330 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20331 disable peeling when we version for aliasing.
20332 (vector_alignment_reachable_p): Honor explicit user alignment.
20333 (vect_supportable_dr_alignment): Likewise.
20334 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20335 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20336 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20337 then peeling to arrange for the cost-model check to come first.
20339 2013-05-10 Alan Modra <amodra@gmail.com>
20341 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20342 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20343 * configure: Regenerate.
20345 2013-05-10 Alan Modra <amodra@gmail.com>
20348 * varasm.c (default_elf_select_section): Move !DECL_P check..
20349 (get_named_section): ..to here before calling get_section_name.
20351 (default_section_type_flags): Add DECL_P check.
20352 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20353 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20355 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20357 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20358 When using gen_stack_adjust_str with a register offset, add a
20359 REG_FRAME_RELATED_EXPR note.
20361 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20363 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20364 (*vec_extractv4si_zext_mem): Ditto.
20365 (*vec_extractv2di): Add 0->x and x->x alternatives.
20366 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20367 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20369 2013-05-09 Jason Merrill <jason@redhat.com>
20371 N3639 C++1y VLA support
20372 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20375 * tree.c (build_constructor_va): New.
20376 * tree.h: Declare it.
20378 2013-05-09 Martin Jambor <mjambor@suse.cz>
20381 * gimple-fold.c (canonicalize_constructor_val): Call
20382 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20384 2013-05-09 Jan Hubicka <jh@suse.cz>
20385 Richard Biener <rguenther@suse.de>
20388 * symtab.c (symtab_make_decl_local): Do not add private names.
20390 2013-05-09 Jan Hubicka <jh@suse.cz>
20393 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20394 (unlink_from_assembler_name_hash): Likewise.
20395 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20396 symtab_unregister_node, symtab_initialize_asm_name_hash,
20397 change_decl_assembler_name): Update.
20399 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20401 * config/aarch64/aarch64.md: New movtf split.
20402 (*movtf_aarch64): Update.
20403 (aarch64_movdi_tilow): Handle TF modes and rename to
20404 aarch64_movdi_<mode>low.
20405 (aarch64_movdi_tihigh): Handle TF modes and rename to
20406 aarch64_movdi_<mode>high
20407 (aarch64_movtihigh_di): Handle TF modes and rename to
20408 aarch64_mov<mode>high_di
20409 (aarch64_movtilow_di): Handle TF modes and rename to
20410 aarch64_mov<mode>low_di
20411 (aarch64_movtilow_tilow): Remove spurious whitespace.
20412 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20414 (aarch64_print_operand): Update.
20416 2013-05-09 Alan Modra <amodra@gmail.com>
20418 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20420 * configure: Regenerate.
20422 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20424 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20425 splitter preparation statements.
20426 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20427 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20430 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20432 * gimple-ssa-strength-reduction.c (count_candidates): Change
20433 return value to int.
20434 (analyze_candidates_and_replace): Change type of length to int.
20436 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20438 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20439 (*vec_extract<mode>): Use VI12_128 mode iterator.
20440 (*vec_extract<mode>_mem): Ditto.
20441 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20444 2013-05-08 Diego Novillo <dnovillo@google.com>
20449 2012-08-17 Diego Novillo <dnovillo@google.com>
20452 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20453 * config.in: Regenerate.
20454 * configure: Regenerate.
20455 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20457 2013-05-08 Jan Hubicka <jh@suse.cz>
20460 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20461 * cgraph.h (symtab_node_base): Add unique_name.
20462 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20463 input_overwrite_node, input_varpool_node): Stream unique_name.
20464 * cgraphclones.c (cgraph_create_virtual_clone,
20465 cgraph_function_versioning): Set unique_name.
20466 * ipa.c (function_and_variable_visibility): Set unique_name.
20468 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20470 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20471 (alloc_cand_and_find_basis): Restrict conditional candidate
20472 processing to CAND_MULTs.
20474 2013-05-08 Jan Hubicka <jh@suse.cz>
20477 lto-symtab.c (lto_symtab_symbol_p): New function.
20478 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20479 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20480 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20481 Skip static symbols.
20483 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20485 PR tree-optimization/57200
20486 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20487 Only call inform if the preceding warning_at returns true.
20489 2013-05-07 Han Shen <shenhan@google.com>
20491 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20492 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20493 * common.opt (fstack-protector-strong): New option.
20494 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20495 * doc/invoke.texi (Optimization Options): Document
20496 "-fstack-protector-strong".
20497 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20499 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20501 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20503 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20505 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20506 it is not lhs of the insn.
20507 (setup_reg_equiv): Remove insn having equiv memory which it is not
20509 * lra-constraints.c (process_address): Try to improve generation
20510 code for address base + disp.
20511 (lra_constraints): Make correct the code for checking insn setting
20512 up backward equivalence. Remove insn only if it is in the init
20514 * lra-eliminations.c (update_reg_eliminate): Change return value.
20515 (lra_eliminate): Use the result.
20517 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20519 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20520 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20521 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20522 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20523 register target operands.
20524 (*vec_extractv8hi_sse2): New pattern.
20525 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20526 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20527 (*vec_extract<mode>_mem): New insn and split pattern.
20529 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20531 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20532 (TARGET_ASAN_SHADOW_OFFSET): Define.
20533 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20534 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20536 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20538 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20539 (incr_vec_index): Return -1 if increment not found.
20540 (create_add_on_incoming_edge): Assert if increment not found.
20541 (record_increment): Limit number of increments recorded.
20542 (all_phi_incrs_profitable): Return false if an increment not found.
20543 (replace_profitable_candidates): Don't process increments that were
20545 (analyze_candidates_and_replace): Limit size of incr_vec.
20547 2013-05-07 Richard Biener <rguenther@suse.de>
20549 * calls.c (special_function_p): setjmp-like functions are leaf.
20550 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20551 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20553 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20555 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20556 (aarch64_simd_mov<mode>): New expander.
20557 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20558 (aarch64_simd_mov_to_<mode>high): Likewise.
20559 (aarch64_simd_mov_from_<mode>low): Likewise.
20560 (aarch64_simd_mov_from_<mode>high): Likewise.
20561 (aarch64_dup_lane<mode>): Update.
20562 (aarch64_dup_lanedi): New instruction pattern.
20563 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20564 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20566 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20568 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20569 (replace_mult_candidate): Remove unnecessary argument; remove
20570 unnecessary parameter from call to introduce_cast_before_cand.
20571 (replace_unconditional_candidate): Remove unnecessary parameter
20572 from call to replace_mult_candidate.
20573 (replace_conditional_candidate): Likewise.
20574 (insert_initializers): Use make_temp_ssa_name.
20575 (introduce_cast_before_cand): Remove unnecessary argument; use
20576 make_temp_ssa_name.
20577 (replace_one_candidate): Remove unnecessary argument; remove
20578 unnecessary parameter from calls to introduce_cast_before_cand.
20579 (replace_profitable_candidates): Remove unnecessary parameters
20580 from calls to replace_one_candidate.
20582 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20584 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20585 phi def as possibly hiding a basis for a CAND_ADD whose operands
20586 have been commuted in the analysis.
20587 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20589 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20591 * config/aarch64/aarch64.md
20592 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20593 shift value between 0-4.
20595 2013-05-07 Richard Biener <rguenther@suse.de>
20597 * double-int.h (rshift): New overload.
20598 * double-int.c (rshift): New function.
20599 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20600 (create_reference_ops_from_ref): Remove.
20601 (vn_reference_insert): Use shared ops for constructing the
20602 reference and copy it.
20604 2013-05-07 Richard Biener <rguenther@suse.de>
20606 PR middle-end/57190
20607 * tree-eh.c (sink_clobbers): Properly propagate
20608 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20610 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20612 PR tree-optimization/57149
20613 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20614 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20615 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20616 uninit_undefined_value_p instead of ssa_undefined_value_p.
20619 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20620 for modifier == EXPAND_INITIALIZER.
20622 2013-05-07 Anton Blanchard <anton@samba.org>
20624 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20625 for powerpc64 little endian.
20626 * configure: Regenerate.
20628 2013-05-06 Graham Stott <grahams@btinternet.com>
20630 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
20631 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
20632 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
20633 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
20635 2013-05-06 Graham Stott <grahams@btinternet.com>
20637 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
20638 codes which allow non-lvalues.
20640 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
20642 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
20643 components are all 1s.
20644 (integer_minus_onep): New function.
20645 * tree.h (integer_minus_onep): Declare it.
20646 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
20647 integer_minus_onep instead of integer_all_onesp.
20649 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20652 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
20653 variations of these patterns.
20655 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20657 * config/i386/i386.md (isa): Add x64_sse4 member.
20658 (enabled): Handle x64_sse4.
20659 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
20660 instruction for 64bit SSE4_1 targets. Update insn attributes.
20661 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
20662 instruction for SSE4_1 targets. Update insn attributes.
20663 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20664 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
20665 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
20667 (*vec_extractv4si): Rename from *sse4_1_pextrd.
20668 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
20669 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
20671 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20674 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
20676 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
20678 * final.c (do_assembler_dialects): Don't handle curly braces and
20679 vertical bar escaped by % as dialect delimiters.
20680 (output_asm_insn): Print curly braces and vertical bar if escaped
20681 by % and ASSEMBLER_DIALECT defined.
20682 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
20683 * doc/tm.texi: Regenerated.
20685 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20687 * config/mips/mips.c: Include tree-pass.h.
20688 (mips_reorg): Split in pre- and post-dbr_schedule parts.
20689 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
20690 (pass_mips_machine_reorg2): New machine specific pass.
20691 (insert_pass_mips_machine_reorg2): New pass plugin definition.
20692 (mips_option_override): Register the new pass.
20693 * rtl.h (cleanup_barriers): Remove prototype.
20694 (dbr_schedule): Likewise.
20695 * jump.c (cleanup_barriers): Make static.
20696 * reorg.c (dbr_schedule): Likewise.
20698 2013-05-06 Richard Biener <rguenther@suse.de>
20700 PR tree-optimization/57185
20701 * tree-parloops.c (add_field_for_reduction): Handle anonymous
20702 SSA names properly.
20704 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20707 * config/i386/i386.c (add_parameter_dependencies): Add dependence
20708 between "first_arg" and "insn", not "last" and "insn".
20710 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20712 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
20713 (find_candidates_in_block): Re-enable slsr_process_phi.
20714 (create_phi_basis): Fix double counting of candidate adjustment.
20716 2013-05-06 Richard Biener <rguenther@suse.de>
20718 PR middle-end/57147
20719 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
20720 the edge is also fallthru, preserve it and just clear the
20722 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
20723 also complex, preserve that and just clear the fallthru flag.
20724 * tree-inline.c (update_ssa_across_abnormal_edges): Also
20725 update virtual operands.
20727 2013-05-06 Alan Modra <amodra@gmail.com>
20729 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
20730 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
20731 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
20732 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
20733 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20734 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
20736 2013-05-06 Alan Modra <amodra@gmail.com>
20738 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
20739 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
20740 (DEFAULT_ASM_ENDIAN): Define.
20741 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20742 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
20743 Update -K PIC clause from sysv4.h.
20744 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
20745 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
20747 2013-05-06 Alan Modra <amodra@gmail.com>
20749 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
20750 twice for little-endian.
20751 (ashrdi3_no_power, ashrdi3): Support little-endian.
20753 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20756 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
20757 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
20759 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
20760 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
20761 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
20764 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
20765 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20767 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
20768 * config/pa/pa.opt: Make mbig-switch a no-op.
20769 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
20770 (CASE_VECTOR_MODE): Always return SImode.
20771 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
20772 for the !TARGET_BIG_SWITCH case.
20773 * config/pa/pa-linux.h: Likewise.
20774 * config/pa/pa-openbsd.h: Likewise.
20775 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
20776 * config/pa/pa.md (short_jump): Remove define_insn.
20777 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
20778 (casesi0): Remove define_insn.
20779 (type): Remove btable_branch.
20780 (pa_combine_type): Likewise.
20781 (in_nullified_branch_delay): Likewise.
20782 (in_call_delay): Likewise.
20783 (define_delay): Likewise.
20784 (define_insn_reservation "Z3"): Likewise.
20785 (define_insn_reservation "Z4"): Likewise.
20786 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
20787 (pa_adjust_insn_length): Remove adjustment for btable branches.
20788 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
20791 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
20793 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20794 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
20795 Add m->r,x alternatives.
20796 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
20797 splitters using SWI48x mode iterator.
20798 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
20799 TARGET_64BIT. Add m->x alternative.
20800 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
20801 Add o->x alternative. Enable for TARGET_SSE.
20802 (sse_storeq): Remove expander.
20803 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
20804 with memory input operand.
20805 (*vec_extractv2di_1 splitter): New.
20806 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
20807 * config/i386/i386.md (ssevecmodelower): New mode attribute.
20809 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20811 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
20812 (INT_LOWPART): Delete.
20813 (extract_MB): Adjust.
20814 (extract_ME): Adjust.
20815 (print_operand): Adjust.
20817 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20819 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20820 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20821 (reg_or_logical_cint_operand, easy_fp_constant,
20822 logical_const_operand): Delete "CONST_DOUBLE" case.
20823 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20824 "HOST_BITS_PER_WIDE_INT == 64" test.
20825 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20826 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20827 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20828 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20830 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20831 CONST_DOUBLE DImode/VOIDmode case.
20832 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20833 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20834 CONST_DOUBLE VOIDmode case.
20835 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20836 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20837 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20838 Delete CONST_DOUBLE case.
20839 (splitters for mov FMOVE64 const_double): Delete
20840 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20841 "HOST_BITS_PER_WIDE_INT >= 64" test.
20842 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20844 (mov DI const_double): Delete.
20846 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20848 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20849 on op shows all bits zero in mode of a lowpart subreg, return zero.
20851 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
20854 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
20855 to save TFmode registers and DImode to save TImode registers for
20856 caller save operations.
20857 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
20858 mark being partially clobbered since they only use the first
20861 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
20862 and TDmode only use the upper 64-bits of each VSX register.
20864 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20866 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
20867 (find_candidates_in_block): Disable slsr_process_phi.
20869 2013-05-03 Guozhi Wei <carrot@google.com>
20871 * coverage.c (coverage_obj_init): Move the construction of gcov
20873 (build_init_ctor): ... here.
20875 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20877 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
20878 (slsr_cand_d): Redefine def_phi.
20879 (stride_status, phi_adjust_status, count_phis_status): New enums.
20880 (find_phi_def): New.
20881 (find_basis_for_base_expr): New.
20882 (find_basis_for_candidate): Handle hidden bases.
20883 (alloc_cand_and_find_basis): Handle phi candidates.
20884 (slsr_process_phi): New.
20885 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
20886 (create_mul_imm_cand): Likewise.
20887 (create_add_ssa_cand): Exclude phi base candidates.
20888 (create_add_imm_cand): Likewise.
20889 (slsr_process_cast): Likewise.
20890 (slsr_process_copy): Likewise.
20891 (find_candidates_in_block): Handle phi candidates.
20892 (dump_candidate): Likewise.
20893 (unconditional_cands): Delete.
20894 (unconditional_cands_with_known_stride_p): Delete.
20895 (phi_dependent_cand_p): New.
20896 (cand_increment): Handle phi-dependent candidates.
20897 (replace_dependent): Delete.
20898 (replace_mult_candidate): New.
20899 (replace_unconditional_candidate): New.
20900 (incr_vec_index): Move to avoid forward reference.
20901 (create_add_on_incoming_edge): New.
20902 (create_phi_basis): New.
20903 (replace_dependents): Delete.
20904 (replace_conditional_candidate): New.
20905 (phi_add_costs): New.
20906 (replace_uncond_cands_and_profitable_phis): New.
20907 (record_increment): Handle phi adjustments.
20908 (record_phi_increments): New.
20909 (record_increments): Handle phi adjustments.
20910 (phi_incr_cost): New.
20911 (lowest_cost_path): Handle phis.
20912 (total_savings): Likewise.
20913 (analyze_increments): Likewise.
20914 (ncd_with_phi): New.
20915 (ncd_of_cand_and_phis): New.
20916 (nearest_common_dominator_for_cands): Handle phi increments.
20917 (all_phi_incrs_profitable): New.
20918 (replace_profitable_candidates): Handle phi-dependent candidates.
20919 (analyze_candidates_and_replace): Likewise.
20921 2013-05-03 Teresa Johnson <tejohnson@google.com>
20924 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
20925 do not exceed REG_BR_PROB_BASE.
20927 2013-05-03 Jeff Law <law@redhat.com>
20929 PR tree-optimization/57144
20930 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
20931 operand of the condition will bit into the new type when eliminating
20932 a cast feeding a condition.
20934 2013-05-03 Jakub Jelinek <jakub@redhat.com>
20936 PR rtl-optimization/57130
20937 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
20938 of COMPARE as in_code to the recursive call if needed.
20940 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
20942 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
20943 (enabled): Handle new members.
20944 * config/i386/sse.md (*vec_concatv2si): Merge from
20945 *vec_concatv2si_sse2 and vec_concatv2si_sse.
20946 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
20948 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
20950 PR tree-optimization/57027
20951 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
20952 for fnms opportunity, check we got the prerequisite kind
20953 of tree / gimple before using accessor functions.
20955 2013-05-03 Richard Biener <rguenther@suse.de>
20957 * double-int.h (lshift): New overload without precision
20958 and arith argument.
20959 (operator *=, operator +=, operator -=): Move ...
20960 * double-int.c (operator *=, operator +=, operator -=): ... here
20961 and implement more efficiently.
20962 (mul_double_with_sign): Remove.
20963 (lshift_double): Adjust to take unsinged shift argument, push
20964 dispatching code to callers.
20965 (mul_double_wide_with_sign): Add early out for callers that
20966 are not interested in high parts or overflow.
20967 (lshift): New function.
20968 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
20969 dispatch code here.
20970 (lrotate, rrotate): Use logical shifts.
20971 * expr.c (get_inner_reference): Use lshift.
20972 * fixed-value.c (do_fixed_divide): Likewise.
20973 * tree-dfa.c (get_ref_base_and_extent): Likewise.
20974 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
20975 (indirect_refs_may_alias_p): Likewise.
20976 (stmt_kills_ref_p_1): Likewise.
20978 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20980 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
20982 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20984 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
20985 scalar form of FABD instruction.
20987 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
20989 * lra-constraints.c (process_alt_operands): Add checking alt
20990 number to choose the best alternative.
20992 2013-05-02 Richard Biener <rguenther@suse.de>
20994 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
20995 bitmap and its handling.
20996 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
20998 2013-05-02 Richard Biener <rguenther@suse.de>
21000 PR middle-end/57140
21001 * tree-inline.c (copy_loops): Properly handle removed loops.
21002 (copy_cfg_body): Mark destination loops for fixup if source
21003 loops needed fixup.
21005 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
21008 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
21009 generating simple_return for naked functions.
21011 2013-05-02 Martin Jambor <mjambor@suse.cz>
21013 PR middle-end/56988
21014 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
21015 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
21017 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21018 ipa_agg_replacement_value structures.
21019 (known_aggs_to_agg_replacement_list): Likewise.
21020 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21021 (read_agg_replacement_chain): Likewise.
21022 (ipcp_transform_function): Also check that by_ref flags match.
21024 2013-05-02 Richard Biener <rguenther@suse.de>
21026 * graphds.h (struct graph): Add obstack member.
21027 * graphds.c (new_graph): Initialize obstack and allocate
21029 (add_edge): Allocate edge from the obstack.
21030 (free_graph): Free the obstack instead of all edges and vertices.
21032 2013-05-02 Teresa Johnson <tejohnson@google.com>
21034 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21036 * cfg.c (update_bb_profile_for_threading): Ditto.
21037 * tree-inline.c (copy_bb): Ditto.
21038 (copy_edges_for_bb): Ditto.
21039 (initialize_cfun): Ditto.
21040 (copy_cfg_body): Ditto.
21041 (expand_call_inline): Ditto.
21042 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21043 (estimate_node_size_and_time): Ditto.
21044 (inline_merge_summary): Ditto.
21045 * cgraphclones.c (cgraph_clone_edge): Ditto.
21046 (cgraph_clone_node): Ditto.
21047 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21048 (compute_trg_info): Ditto.
21050 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21052 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21053 S reg when fp attribute set.
21054 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21056 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21058 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21060 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21061 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21062 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21064 2013-05-02 Richard Biener <rguenther@suse.de>
21066 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21067 (struct instantiate_cache_entry): New type.
21068 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21069 (struct instantiate_cache_type): New type.
21070 (set_instantiated_value, get_instantiated_value): Remove.
21071 (get_instantiated_value_entry): New function.
21072 (instantiate_scev_name): Use the new cache and adjust.
21073 (instantiate_scev_poly): Adjust.
21074 (instantiate_scev_binary): Likewise.
21075 (instantiate_array_ref): Likewise.
21076 (instantiate_scev_convert): Likewise.
21077 (instantiate_scev_not): Likewise.
21078 (instantiate_scev_3): Likewise.
21079 (instantiate_scev_2): Likewise.
21080 (instantiate_scev_r): Likewise.
21081 (instantiate_scev): Likewise.
21082 (resolve_mixers): Likewise.
21084 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21087 * lra-constraints.c (best_small_class_operands_num): Remove.
21088 (process_alt_operands): Remove small_class_operands_num. Take
21089 small classes operands into losers and only if the operand is not
21090 matched. Modify debugging output.
21091 (curr_insn_transform): Remove best_small_class_operands_num.
21094 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21096 * config/aarch64/aarch64-builtins.c
21097 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21098 * config/aarch64/aarch64-simd-builtins.def
21099 (reduc_splus_): Add new modes.
21100 (reduc_uplus_): New.
21101 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21102 (reduc_uplus_v4sf): Likewise.
21103 (reduc_splus_v4sf): Likewise.
21104 (aarch64_addv<mode>): Likewise.
21105 (reduc_uplus_<mode>): Likewise.
21106 (reduc_splus_<mode>): Likewise.
21107 (aarch64_addvv2di): Likewise.
21108 (reduc_uplus_v2di): Likewise.
21109 (reduc_splus_v2di): Likewise.
21110 (aarch64_addvv2si): Likewise.
21111 (reduc_uplus_v2si): Likewise.
21112 (reduc_splus_v2si): Likewise.
21113 (reduc_<sur>plus_<mode>): New.
21114 (reduc_<sur>plus_v2di): Likewise.
21115 (reduc_<sur>plus_v2si): Likewise.
21116 (reduc_<sur>plus_v4sf): Likewise.
21117 (aarch64_addpv4sf): Likewise.
21118 * config/aarch64/arm_neon.h
21119 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21120 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21121 add UNSPEC_SADDV, UNSPEC_UADDV.
21123 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21125 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21127 * config/aarch64/arm_neon.h
21128 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21130 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21132 * config/aarch64/aarch64-builtins
21133 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21135 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21137 * config/aarch64/aarch64-simd-builtins.def
21138 (reduc_smax_): New.
21139 (reduc_smin_): Likewise.
21140 (reduc_umax_): Likewise.
21141 (reduc_umin_): Likewise.
21142 (reduc_smax_nan_): Likewise.
21143 (reduc_smin_nan_): Likewise.
21146 (smax): Update for V2SF, V4SF and V2DF modes.
21149 (smin_nan): Likewise.
21150 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21151 (<su><maxmin><mode>3): ...This, refactor.
21152 (s<maxmin><mode>3): New.
21153 (<maxmin_uns><mode>3): Likewise.
21154 (reduc_<maxmin_uns>_<mode>): Refactor.
21155 (reduc_<maxmin_uns>_v4sf): Likewise.
21156 (reduc_<maxmin_uns>_v2si): Likewise.
21157 (aarch64_<fmaxmin><mode>: Remove.
21158 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21160 (vmin<q>_f<32,64>): Likewise.
21161 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21163 (su): Add mappings for smax, smin, umax, umin.
21165 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21166 (FMAXMIN): Rename as...
21167 (FMAXMIN_UNS): ...This.
21169 (fmaxminv): Likewise.
21170 (fmaxmin): Likewise.
21172 (maxmin_uns_op): Likewise.
21174 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21176 * config/aarch64/arm_neon.h
21177 (vac<ge, gt><sd>_f<32, 64>): Rename to...
21178 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
21179 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
21181 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21183 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
21184 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
21186 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21188 * config/aarch64/aarch64-simd.md
21189 (vcond<mode>_internal): Handle special cases for constant masks.
21190 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
21191 (vcondu<mode><mode>): Likewise.
21192 (vcond<v_cmp_result><mode>): New.
21194 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21196 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
21197 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
21198 * config/aarch64/aarch64-simd-builtins.def
21199 (cmeq): Update to BUILTIN_VALLDI.
21204 * config/aarch64/arm_neon.h
21205 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
21206 to builtins or C as appropriate.
21208 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21210 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21212 (cmhi): Rename to...
21214 * config/aarch64/aarch64-simd.md
21215 (simd_mode): Add SF.
21216 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21217 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21218 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21219 (cstore<mode>_neg): ...This.
21220 * config/aarch64/iterators.md
21222 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21223 (COMPARISONS): New.
21224 (UCOMPARISONS): Likewise.
21225 (optab): Add missing comparisons.
21231 (VCMP_S): Likewise.
21232 (VCMP_U): Likewise.
21233 (V_cmp_result): Add DF, SF modes.
21234 (v_cmp_result): Likewise.
21236 (vmtype): Likewise.
21237 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21239 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21241 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21242 define_insn to define_insn_and_split.
21243 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21244 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21245 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21246 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21247 (thumb2_negscc): Likewise.
21249 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21251 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21253 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21255 * config/arm/thumb2.md: Remove trailing whitespaces.
21257 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21259 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21260 Use gen_int_mode rather than GEN_INT.
21262 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21264 * value-prof.c (stream_in_histogram_value): Remove the strayed
21267 2013-04-30 Richard Biener <rguenther@suse.de>
21269 PR middle-end/57122
21270 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21272 2013-04-30 Richard Biener <rguenther@suse.de>
21274 PR middle-end/57107
21275 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21277 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21279 PR rtl-optimization/56957
21280 PR rtl-optimization/57105
21281 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21282 variable. Use just INSN_UID for determining whether an insn
21283 should be only disconnected from the insn stream.
21284 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21286 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21288 PR tree-optimization/57104
21289 * tsan.c (instrument_expr): Don't instrument accesses to
21290 DECL_HARD_REGISTER VAR_DECLs.
21292 2013-04-30 Richard Biener <rguenther@suse.de>
21294 * function.h (loops_for_fn): New inline function.
21295 (set_loops_for_fn): Likewise.
21296 * cfgloop.h (place_new_loop): Add struct function parameter.
21297 (get_loop): Likewise.
21298 (get_loops): Likewise.
21299 (number_of_loops): Likewise.
21300 (fel_next): Adjust.
21301 (fel_init): Likewise.
21302 * cfg.c (get_loop_copy): Adjust.
21303 * cfgloop.c (flow_loops_dump): Likewise.
21304 (record_loop_exits): Likewise.
21305 (verify_loop_structure): Likewise.
21306 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21307 (estimate_reg_pressure_cost): Likewise.
21308 (mark_loop_exit_edges): Likewise.
21309 * cfgloopmanip.c (place_new_loop): Likewise.
21310 (add_loop): Likewise.
21311 (duplicate_loop): Likewise.
21312 * graph.c (draw_cfg_nodes): Likewise.
21313 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21314 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21315 (extract_affine_chrec): Likewise.
21316 (build_scop_iteration_domain): Likewise.
21317 * graphite.c (graphite_initialize): Likewise.
21318 * ira-build.c (create_loop_tree_nodes): Likewise.
21319 (more_one_region_p): Likewise.
21320 (rebuild_regno_allocno_maps): Likewise.
21321 (mark_loops_for_removal): Likewise.
21322 (mark_all_loops_for_removal): Likewise.
21323 (remove_unnecessary_regions): Likewise.
21324 (ira_build): Likewise.
21325 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21326 * loop-init.c (fix_loop_structure): Likewise.
21327 (gate_rtl_move_loop_invariants): Likewise.
21328 (gate_rtl_unswitch): Likewise.
21329 (gate_rtl_unroll_and_peel_loops): Likewise.
21330 (rtl_doloop): Likewise.
21331 * lto-streamer-in.c (input_cfg): Likewise.
21332 * lto-streamer-out.c (output_cfg): Likewise.
21333 * modulo-sched.c (sms_schedule): Likewise.
21334 * predict.c (tree_estimate_probability): Likewise.
21335 (tree_estimate_probability_driver): Likewise.
21336 (estimate_loops): Likewise.
21337 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21338 (move_sese_region_to_fn): Likewise.
21339 (debug_loop_num): Likewise.
21340 * tree-chrec.c (chrec_evaluate): Likewise.
21341 (hide_evolution_in_other_loops_than_loop): Likewise.
21342 (chrec_component_in_loop_num): Likewise.
21343 (reset_evolution_in_loop): Likewise.
21344 (evolution_function_is_invariant_rec_p): Likewise.
21345 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21346 * tree-inline.c (copy_loops): Likewise.
21347 (copy_cfg_body): Likewise.
21348 (tree_function_versioning): Likewise.
21349 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21350 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21352 (add_to_evolution_1): Likewise.
21353 (scev_const_prop): Likewise.
21354 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21355 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21356 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21357 (tree_ssa_lim_initialize): Likewise.
21358 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21359 (verify_loop_closed_ssa): Likewise.
21360 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21361 (tree_ssa_loop_im): Likewise.
21362 (tree_ssa_loop_unswitch): Likewise.
21363 (tree_vectorize): Likewise.
21364 (check_data_deps): Likewise.
21365 (tree_ssa_loop_ivcanon): Likewise.
21366 (tree_ssa_loop_bounds): Likewise.
21367 (tree_complete_unroll): Likewise.
21368 (tree_complete_unroll_inner): Likewise.
21369 (tree_parallelize_loops): Likewise.
21370 (tree_ssa_loop_prefetch): Likewise.
21371 (tree_ssa_loop_ivopts): Likewise.
21372 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21373 * tree-vectorizer.c (vectorize_loops): Likewise.
21375 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21377 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21378 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21379 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21380 with EABI_LINK_SPEC.
21382 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21385 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21388 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21391 * lra-constraints.c (process_alt_operands): Discourage a bit more
21392 using memory for pseudos. Print cost dump for alternatives.
21393 Modify cost values for conflicts with early clobbers.
21394 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21396 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21399 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21401 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21403 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21404 from/to S register.
21405 (movdi_aarch64): Support LDR/STR from/to D register.
21407 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21409 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21410 or higher by default.
21412 2013-04-29 Richard Biener <rguenther@suse.de>
21414 PR middle-end/57075
21415 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21416 even if not adding abnormal edges for calls that can make
21419 2013-04-29 Richard Biener <rguenther@suse.de>
21421 PR middle-end/57103
21422 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21424 (move_stmt_r): Remove redundant checking.
21426 2013-04-29 Teresa Johnson <tejohnson@google.com>
21429 * basic-block.h (apply_scale): New function.
21430 (apply_probability): Use apply_scale.
21431 * gimple-streamer-in.c (input_bb): Ditto.
21432 * lto-streamer-in.c (input_cfg): Ditto.
21433 * lto-cgraph.c (merge_profile_summaries): Ditto.
21434 * tree-optimize.c (execute_fixup_cfg): Ditto.
21435 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21436 (copy_edges_for_bb): Ditto.
21437 (copy_cfg_body): Ditto.
21439 2013-04-29 Tom de Vries <tom@codesourcery.com>
21441 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21442 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21443 (tail_merge_optimize): Handle current_loops == NULL.
21445 2013-04-26 Jeff Law <law@redhat.com>
21447 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21448 (simplify_cond_using_ranges): Generalize code to simplify
21449 COND_EXPRs where one argument is a constant and the other
21450 is an SSA_NAME created by an integral type conversion.
21452 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21454 * config/arm/arm.md (store_minmaxsi): Use only when
21455 optimize_insn_for_size_p.
21457 2013-04-29 Christian Bruel <christian.bruel@st.com>
21460 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21462 2013-04-29 Richard Biener <rguenther@suse.de>
21464 PR middle-end/57089
21465 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21466 loop tree make sure to schedule a fixup for the child as well.
21467 (expand_omp_for_generic): Properly add loops.
21468 (expand_omp_for_static_nochunk): Likewise.
21469 (expand_omp_for_static_chunk): Likewise.
21470 (expand_omp_for): For the degenerate case fixup loops.
21471 (expand_omp_sections): Fix default bb placement in loops.
21472 (expand_omp_atomic_pipeline): Properly add loops.
21474 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21476 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21478 2013-04-29 Tom de Vries <tom@codesourcery.com>
21480 * tree-ssa-tail-merge.c: Update header comment.
21482 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21484 * config/aarch64/arm_neon.h
21485 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21486 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21487 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21488 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21489 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21490 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21491 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21492 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21494 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21496 * config/aarch64/aarch64-simd.md
21497 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21498 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21499 fix_trunc, fixuns_trunc.
21500 (ftrunc<VDQF:mode>2): New.
21501 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21502 (fix_trunc_optab): New.
21504 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21506 * config/aarch64/aarch64-builtins.c
21507 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21508 iceilf, lround, iroundf.
21510 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21513 * config/i386/i386.h (enum ix86_tune_indices)
21514 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21515 New, split from X86_TUNE_INTER_UNIT_MOVES.
21516 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21517 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21518 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21519 (TARGET_INTER_UNIT_MOVES): Remove.
21520 * config/i386/i386.c (initial_ix86_tune_features): Update.
21521 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21522 (ix86_expand_convert_uns_didf_sse): Use
21523 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21524 (ix86_expand_vector_init_one_nonzero): Ditto.
21525 (ix86_expand_vector_init_interleave): Ditto.
21526 (inline_secondary_memory_needed): Return true for moves from SSE class
21527 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21528 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21529 * config/i386/constraints.md (Yi, Ym): Depend on
21530 TARGET_INTER_UNIT_MOVES_TO_VEC.
21531 (Yj, Yn): New constraints.
21532 * config/i386/i386.md (*movdi_internal): Change constraints of
21533 operand 1 from Yi to Yj and from Ym to Yn.
21534 (*movsi_internal): Ditto.
21535 (*movdf_internal): Ditto.
21536 (*movsf_internal): Ditto.
21537 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21538 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21539 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21540 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21541 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21542 * config/i386/sse.md (movdi_to_sse): Ditto.
21543 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21544 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21545 TARGET_INTER_UNIT_MOVES.
21546 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21547 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21548 instead of TARGET_INTER_UNIT_MOVES.
21549 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21550 operand 1 from Yi to Yj and from Ym to Yn.
21552 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21554 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21555 (float_truncate_hi_): Likewise.
21556 (float_extend_lo_): Likewise.
21557 (float_truncate_lo_): Likewise.
21558 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21559 (aarch64_float_extend_lo_v2df): Likewise.
21560 (vec_unpacks_hi_v4sf): Likewise.
21561 (aarch64_float_truncate_lo_v2sf): Likewise.
21562 (aarch64_float_truncate_hi_v4sf): Likewise.
21563 (vec_pack_trunc_v2df): Likewise.
21564 (vec_pack_trunc_df): Likewise.
21566 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21568 * config/aarch64/aarch64-builtins.c
21569 (aarch64_fold_builtin): Fold float conversions.
21570 * config/aarch64/aarch64-simd-builtins.def
21571 (floatv2si, floatv4si, floatv2di): New.
21572 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21573 * config/aarch64/aarch64-simd.md
21574 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21575 * config/aarch64/iterators.md (FLOATUORS): New.
21576 (optab): Add float, floatuns.
21577 (su_optab): Likewise.
21579 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21581 * config/aarch64/aarch64-builtins.c
21582 (aarch64_builtin_vectorized_function): Use new names for
21584 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21585 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21586 (fcvtzu): Split as...
21587 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21588 (fcvtas): Split as...
21589 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21590 (fcvtau): Split as...
21591 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21592 (fcvtps): Split as...
21593 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21594 (fcvtpu): Split as...
21595 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21596 (fcvtms): Split as...
21597 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21598 (fcvtmu): Split as...
21599 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21600 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21601 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21602 (lfrintnusf, lfrintnudf): Likewise.
21603 * config/aarch64/aarch64-simd.md
21604 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21606 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21607 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21608 (fcvt_pattern): Likewise.
21610 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21612 * config/aarch64/aarch64-simd.md
21613 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21614 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21616 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21618 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21619 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21620 (vrnd<a,m,n,p>_f32): Implement using builtins.
21621 (vrnd<i,x><q>_f<32, 64>): New.
21623 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21625 * config/aarch64/aarch64-builtins.c
21626 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21627 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21628 (frintz): Rename to...
21630 (frintp): Rename to...
21632 (frintm): Rename to...
21634 (frinti): Rename to...
21635 (nearbyint): ...this.
21636 (frintx): Rename to...
21638 (frinta): Rename to...
21640 * config/aarch64/aarch64-simd.md
21641 (aarch64_frint<frint_suffix><mode>): Delete.
21642 (<frint_pattern><mode>2): Convert to insn.
21643 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
21644 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
21645 (frint_pattern): Likewise.
21646 (frint_suffix): Likewise.
21648 2013-04-29 Richard Biener <rguenther@suse.de>
21650 PR tree-optimization/57081
21651 * loop-init.c: Include tree-flow.h.
21652 (loop_optimizer_finalize): Free number of iteration estimates.
21653 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
21655 2013-04-29 Jakub Jelinek <jakub@redhat.com>
21657 PR tree-optimization/57083
21658 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
21659 non-singleton shift count range, zero extend low_bound for uns case.
21661 * config/i386/predicates.md (general_vector_operand): New predicate.
21662 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
21663 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
21664 if they aren't nonimmediate operands. If their original values
21665 satisfy const_vector_equal_evenodd_p, don't shift them.
21666 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
21667 predicates. For the SSE4.1 case force operands[{1,2}] into registers
21668 if not nonimmediate_operand.
21669 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
21670 instead of register_operand.
21671 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
21673 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
21675 * stor-layout.c (finalize_size_functions): Allocate a structure and
21676 reset cfun before dumping the functions.
21678 2013-04-27 Jakub Jelinek <jakub@redhat.com>
21680 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
21683 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
21684 use xop_pmacsdqh if uns_p.
21685 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
21686 the immediate rotate count.
21688 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
21690 * rtl.h (struct rtx_def): Add comment for field jump.
21691 (LRA_SUBREG_P): New macro.
21692 * recog.c (register_operand): Check LRA_SUBREG_P.
21693 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21695 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21696 (lra_final_code_change): Skip subreg change for operators.
21697 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21698 if there are no operand changes.
21699 * lra-constraints.c (curr_insn_set): New.
21700 (match_reload): Set LRA_SUBREG_P.
21701 (emit_spill_move): Ditto.
21702 (check_and_process_move): Use curr_insn_set. Process only single
21703 set insns. Don't initialize sec_mem_p and change_p.
21704 (simplify_operand_subreg): Use LRA_SUBREG_P.
21705 (reg_in_class_p): New function.
21706 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21707 of #ifdef. Add code to remove cycling.
21708 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21709 non-null disp. Reload inner instead of disp when base and index
21710 are null. Try to put lo_sum into register.
21711 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21712 (check_and_process_move): Move code for move cost check to
21713 simple_move_p. Remove equiv_substitution.
21714 (simple_move_p): New function.
21715 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21716 curr_insn_set. Call check_and_process_move only for single set
21717 insns. Use the new function. Move call of check_and_process_move
21718 after operand equiv substitution and address process.
21720 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21723 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
21724 with nonlocal goto receivers or returns twice calls, ignore
21725 unininitialized values from abnormal edges to nl goto receiver
21726 or returns twice call.
21728 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21730 PR tree-optimization/57051
21731 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
21732 and VEC_RSHIFT_EXPR if shift count is a multiple of element
21735 2013-04-26 Richard Biener <rguenther@suse.de>
21737 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
21738 (expand_omp_taskreg): Likewise. Mark loops for fixup.
21739 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
21740 (fixup_loop_arrays_after_move): New function.
21741 (move_sese_region_to_fn): Properly outline the loop tree parts
21742 of the SESE region.
21744 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
21746 * config/i386/i386.md (type, unit): Fix long lines.
21748 2013-04-26 Richard Biener <rguenther@suse.de>
21750 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
21751 (lto-streamer-out.o): Likewise.
21752 * cfgloop.c (init_loops_structure): Export, add struct function
21753 argument and adjust.
21754 (flow_loops_find): Adjust.
21755 * cfgloop.h (enum loop_estimation): Add EST_LAST.
21756 (init_loops_structure): Declare.
21757 * lto-streamer-in.c: Include cfgloop.h.
21758 (input_cfg): Input the loop tree.
21759 * lto-streamer-out.c: Include cfgloop.h.
21760 (output_cfg): Output the loop tree.
21761 (output_struct_function_base): Do not drop PROP_loops.
21763 2013-03-26 Richard Biener <rguenther@suse.de>
21765 * tree-cfg.c (execute_build_cfg): Build the loop tree.
21766 (pass_build_cfg): Provide PROP_loops.
21767 (move_sese_region_to_fn): Remove loops that are outlined into fn
21769 * tree-inline.c: Include cfgloop.h.
21770 (initialize_cfun): Do not drop PROP_loops.
21771 (copy_loops): New function.
21772 (copy_cfg_body): Copy loop structure.
21773 (tree_function_versioning): Initialize destination loop tree.
21774 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
21775 (pass_parallelize_loops): Do IL verification.
21776 * loop-init.c (loop_optimizer_init): Fixup loops if required.
21777 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
21778 the CFG make sure we fixup loops as well.
21779 * tree-ssa-tail-merge.c: Include cfgloop.h.
21780 (replace_block_by): When merging loop latches mark loops for fixup.
21781 * lto-streamer-out.c (output_struct_function_base): Drop
21782 PROP_loops for now.
21783 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
21784 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
21785 * ipa-split.c: Include cfgloop.h.
21786 (split_function): Add the new return block to the loop tree root.
21787 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
21788 whether we have removed the forwarder block.
21789 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
21790 * cfgloop.h (place_new_loop): Declare.
21791 * cfgloopmanip.c (place_new_loop): Export.
21792 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
21793 (tree-switch-conversion.o): Likewise.
21794 (tree-complex.o): Likewise.
21795 (tree-inline.o): Likewise.
21796 (tree-ssa-tailmerge.o): Likewise.
21797 (ipa-split.o): Likewise.
21798 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
21799 (tree-ssa-copy.o): Likewise.
21800 * tree-switch-conversion.c: Include cfgloop.h
21801 (process_switch): If we emit a bit-test cascade, schedule loops
21803 * tree-complex.c: Include cfgloop.h.
21804 (expand_complex_div_wide): Properly add new basic-blocks to loops.
21805 * asan.c: Include cfgloop.h.
21806 (create_cond_insert_point): Properly add new basic-blocks to
21807 loops, schedule loop fixup.
21808 * cfgloop.c (verify_loop_structure): Check that looks are not
21810 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
21812 (expand_omp_for_generic): Likewise.
21813 (expand_omp_sections): Likewise.
21814 (expand_omp_atomic_pipeline): Schedule loops for fixup.
21815 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
21816 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
21817 is initialized, not when loops are present.
21818 * tree-parloops.c (parallelize_loops): Remove checking here.
21819 * passes.c (init_optimization_passes): Schedule a copy-propagation
21820 pass before complete unrolling of inner loops.
21822 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21824 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21825 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21826 (colorize_init): Add argument to _WIN32 version.
21827 * toplev.c: Include diagnostic-color.h.
21828 (process_options): Default to -fdiagnostics-color=auto if
21829 GCC_COLORS env var is in the environment.
21830 * common.opt (fdiagnostics-color=): Add Var and Init.
21831 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21832 env var is in the environment, the default is auto rather than never.
21834 * diagnostic.h (file_name_as_prefix): Add context argument.
21835 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21836 the string as locus.
21837 * langhooks.c (lhd_print_error_function): Adjust caller.
21839 2013-04-25 Lawrence Crowl <crowl@google.com>
21841 * var-tracking.c (shared_hash_def::htab):
21842 Change type to hash_table. Update dependent calls and types.
21844 2013-04-25 Lawrence Crowl <crowl@google.com>
21846 * Makefile.in: Update as needed below.
21848 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21849 Move declaration to after the type's method definitons.
21851 * attribs.c (htab_t scoped_attributes::attribute_hash):
21852 Change type to hash_table. Update dependent calls and types.
21854 * bitmap.c (htab_t bitmap_desc_hash):
21855 Change type to hash_table. Update dependent calls and types.
21857 * cselib.c (htab_t cselib_hash_table):
21858 Change type to hash_table. Update dependent calls and types.
21860 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
21861 (hash_string_slot_node): Move implementation into lto-streamer.h
21862 struct string_slot_hasher.
21863 (eq_string_slot_node): Likewise.
21865 * data-streamer-out.c: Update output_block::string_hash_table
21866 dependent calls and types.
21868 * dwarf2cfi.c (htab_t trace_index):
21869 Change type to hash_table. Update dependent calls and types.
21871 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
21872 Change type to hash_table. Update dependent calls and types.
21873 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
21874 (htab_t optimize_external_refs::map): Likewise.
21875 (htab_t output_comp_unit::extern_map): Likewise.
21876 (htab_t output_comdat_type_unit::extern_map): Likewise.
21877 (htab_t output_macinfo::macinfo_htab): Likewise.
21878 (htab_t optimize_location_lists::htab): Likewise.
21879 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
21881 * except.c (htab_t ehspec_hash_type):
21882 Change type to hash_table. Update dependent calls and types.
21883 (assign_filter_values::ttypes): Likewise.
21884 (assign_filter_values::ehspec): Likewise.
21885 (sjlj_assign_call_site_values::ar_hash): Likewise.
21886 (convert_to_eh_region_ranges::ar_hash): Likewise.
21888 * gcse.c (htab_t pre_ldst_table):
21889 Change type to hash_table. Update dependent calls and types.
21891 * ggc-common.c (htab_t saving_htab):
21892 Change type to hash_table. Update dependent calls and types.
21893 (htab_t loc_hash): Likewise.
21894 (htab_t ptr_hash): Likewise.
21895 (call_count): Rename ggc_call_count.
21896 (call_alloc): Rename ggc_call_alloc.
21897 (loc_descriptor): Rename make_loc_descriptor.
21898 (add_statistics): Rename ggc_add_statistics.
21900 * ggc-common.c (saving_htab):
21901 Change type to hash_table. Update dependent calls and types.
21903 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
21904 (push_gimplify_context): Likewise.
21905 (pop_gimplify_context): Likewise.
21906 (struct gimple_temp_hash_elt): Added.
21907 (struct gimplify_hasher): Likewise.
21908 (struct gimplify_ctx.temp_htab):
21909 Change type to hash_table. Update dependent calls and types.
21911 * gimple-fold.c: Include gimplify-ctx.h.
21913 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
21914 Change type to hash_table. Update dependent calls and types.
21915 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
21916 avoid potential global name collision.
21918 * gimplify.c: Include gimplify-ctx.h.
21919 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
21920 (htab_t gimplify_ctx::temp_htab):
21921 Update dependent calls and types for new type hash_table.
21922 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
21923 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
21925 * gimplify-ctx.h: New.
21926 (struct gimple_temp_hash_elt): Move from gimplify.c.
21927 (class gimplify_hasher): New.
21928 (struct gimplify_ctx): Move from gimple.h.
21929 (htab_t gimplify_ctx::temp_htab):
21930 Change type to hash_table. Update dependent calls and types.
21932 * graphite-clast-to-gimple.c: Include graphite-htab.h.
21933 (htab_t ivs_params::newivs_index):
21934 Change type to hash_table. Update dependent calls and types.
21935 (htab_t ivs_params::params_index): Likewise.
21936 (htab_t print_generated_program::params_index): Likewise.
21937 (htab_t gloog::newivs_index): Likewise.
21938 (htab_t gloog::params_index): Likewise.
21940 * graphite.c: Include graphite-htab.h.
21941 4htab_t graphite_transform_loops::bb_pbb_mapping):
21942 Change type to hash_table. Update dependent calls and types.
21944 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
21945 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
21946 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
21948 * graphite-dependences.c: Include graphite-htab.h.
21949 (loop_is_parallel_p): Change hash table type of parameter.
21951 * graphite-htab.h: New.
21952 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
21953 (extern find_pbb_via_hash): Move from graphite-poly.h.
21954 (extern loop_is_parallel_p): Move from graphite-poly.h.
21955 (extern get_loop_body_pbbs): Move from graphite-poly.h.
21957 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
21958 (extern loop_is_parallel_p): Move to graphite-htab.h.
21959 (extern get_loop_body_pbbs): Move to graphite-htab.h.
21961 * haifa-sched.c (htab_t delay_htab):
21962 Change type to hash_table. Update dependent calls and types.
21963 (htab_t delay_htab_i2): Likewise.
21965 * ira-color.c (htab_t allocno_hard_regs_htab):
21966 Change type to hash_table. Update dependent calls and types.
21968 * ira-costs.c (htab_t cost_classes_htab):
21969 Change type to hash_table. Update dependent calls and types.
21971 * loop-invariant.c (htab_t merge_identical_invariants::eq):
21972 Change type to hash_table. Update dependent calls and types.
21974 * loop-iv.c (htab_t bivs):
21975 Change type to hash_table. Update dependent calls and types.
21977 * loop-unroll.c (htab_t opt_info::insns_to_split):
21978 Change type to hash_table. Update dependent calls and types.
21979 (htab_t opt_info::insns_with_var_to_expand): Likewise.
21981 * lto-streamer.h (struct string_slot): Move from data-streamer.h
21982 (struct string_slot_hasher): New.
21983 (htab_t output_block::string_hash_table):
21984 Change type to hash_table. Update dependent calls and types.
21986 * lto-streamer-in.c (freeing_string_slot_hasher): New.
21987 (htab_t file_name_hash_table):
21988 Change type to hash_table. Update dependent calls and types.
21990 * lto-streamer-out.c: Update output_block::string_hash_table dependent
21993 * lto-streamer.c (htab_t tree_htab):
21994 Change type to hash_table. Update dependent calls and types.
21996 * omp-low.c: Include gimplify-ctx.h.
21998 * passes.c (htab_t name_to_pass_map):
21999 Change type to hash_table. Update dependent calls and types.
22000 (pass_traverse): Rename to passes_pass_traverse.
22002 * plugin.c (htab_t event_tab):
22003 Change type to hash_table. Update dependent calls and types.
22005 * postreload-gcse.c (htab_t expr_table):
22006 Change type to hash_table. Update dependent calls and types.
22007 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
22009 * sese.c (debug_rename_map_1): Make extern.
22010 (htab_t copy_bb_and_scalar_dependences::rename_map):
22011 Change type to hash_table. Update dependent calls and types.
22013 * sese.h (extern debug_rename_map): Move to .c file.
22015 * store-motion.c (htab_t store_motion_mems_table):
22016 Change type to hash_table. Update dependent calls and types.
22018 * trans-mem.c (htab_t tm_new_mem_hash):
22019 Change type to hash_table. Update dependent calls and types.
22021 * tree-browser.c (htab_t TB_up_ht):
22022 Change type to hash_table. Update dependent calls and types.
22024 * tree-cfg.c (htab_t discriminator_per_locus):
22025 Change type to hash_table. Update dependent calls and types.
22027 * tree-complex.c: Include tree-hasher.h
22028 (htab_t complex_variable_components):
22029 Change type to hash_table. Update dependent calls and types.
22031 * tree-eh.c (htab_t finally_tree):
22032 Change type to hash_table. Update dependent calls and types.
22034 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22035 struct int_tree_hasher.
22036 (extern int_tree_map_eq): Likewise.
22037 (uid_decl_map_hash): Removed.
22038 (extern decl_tree_map_eq): Likewise.
22040 * tree-hasher.h: New.
22041 (struct int_tree_hasher): New.
22042 (typedef int_tree_htab_type): New.
22044 * tree-inline.c: Include gimplify-ctx.h.
22046 * tree-mudflap.c: Include gimplify-ctx.h.
22048 * tree-parloops.c: Include tree-hasher.h.
22049 (htab_t eliminate_local_variables_stmt::decl_address):
22050 Change type to hash_table. Update dependent calls and types.
22051 (htab_t separate_decls_in_region::decl_copies): Likewise.
22053 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22054 Change type to hash_table. Update dependent calls and types.
22056 * tree-sra.c (candidates):
22057 Change type to hash_table. Update dependent calls and types.
22059 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22061 (int_tree_map_hash): Likewise.
22063 * tree-ssa-dom.c (htab_t avail_exprs):
22064 Change type to hash_table. Update dependent calls and types.
22066 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22067 Change type to hash_table. Update dependent calls and types.
22069 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22070 Change type to hash_table. Update dependent calls and types.
22072 * tree-ssa-phiopt.c (seen_ssa_names):
22073 Change type to hash_table. Update dependent calls and types.
22075 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22076 Change type to hash_table. Update dependent calls and types.
22078 * tree-ssa-uncprop.c (equiv):
22079 Change type to hash_table. Update dependent calls and types.
22081 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22083 PR rtl-optimization/57003
22084 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22085 call note_stores with kill_clobbered_value callback again after
22086 killing regs_invalidated_by_call.
22088 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22090 * config/aarch64/aarch64-simd.md
22091 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22092 (aarch64_simd_bsl<mode>): Likewise.
22093 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22095 2013-04-25 Marek Polacek <polacek@redhat.com>
22097 PR tree-optimization/57066
22098 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22100 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22102 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22104 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22106 * config/aarch64/aarch64-builtins.c
22107 (aarch64_fold_builtin): New.
22108 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22109 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22110 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22111 * config/aarch64/arm_neon.h
22112 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22114 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22115 Tejas Belagod <tejas.belagod@arm.com>
22117 * config/aarch64/aarch64-builtins.c
22118 (aarch64_gimple_fold_builtin): New.
22119 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22120 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22121 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22122 (addvv4sf): Update.
22123 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22125 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22127 * config/aarch64/aarch64.md
22128 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22130 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22132 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22133 (*ngcsi_uxtw): New pattern.
22135 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22136 Julian Brown <julian@codesourcery.com>
22138 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22139 (TB_DREG): Add T_V4HF.
22140 (v4hf_UP): New macro.
22141 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22142 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22143 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22145 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22146 (arm_vector_mode_supported_p): Handle V4HF.
22147 (arm_mangle_map): Handle V4HFmode.
22148 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22149 * config/arm/arm_neon_builtins.def: Add entries for
22150 vcvtv4hfv4sf, vcvtv4sfv4hf.
22151 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22152 (neon_vcvtv4hfv4sf): Likewise.
22153 * config/arm/neon-gen.ml: Handle half-precision floating point
22155 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22156 * config/arm/arm_neon.h: Regenerate.
22157 * config/arm/neon.ml (type elts): Add F16.
22158 (type vectype): Add T_float16x4, T_floatHF.
22159 (type vecmode): Add V4HF.
22160 (type features): Add Requires_FP_bit feature.
22161 (elt_width): Handle F16.
22162 (elt_class): Likewise.
22163 (elt_of_class_width): Likewise.
22164 (mode_of_elt): Refactor.
22165 (type_for_elt): Handle F16, fix error messages.
22166 (vectype_size): Handle T_float16x4.
22167 (vcvt_sh): New function.
22168 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
22169 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
22170 (string_of_mode): Handle V4HF.
22171 * doc/arm-neon-intrinsics.texi: Regenerate.
22173 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22175 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
22176 format specifier in 'X' case.
22178 2013-04-25 Alan Modra <amodra@gmail.com>
22181 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
22182 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
22183 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
22184 Repeat for many other rotate/shift and mask patterns using subregs.
22185 Name lshiftrt insns.
22186 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
22187 on WORDS_BIG_ENDIAN.
22189 2013-04-25 Alan Modra <amodra@gmail.com>
22191 * config.gcc: Support little-endian powerpc-linux targets.
22192 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
22193 (LINK_OS_LINUX_SPEC): Define.
22194 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
22195 Preserve MASK_LITTLE_ENDIAN.
22196 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
22197 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
22198 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
22199 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
22200 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
22201 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
22202 Correct fp word order for little-endian. Don't shift toc entries
22203 smaller than a word for little-endian.
22204 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
22205 (bswapdi2 splits): Correct low-part subreg for little-endian.
22206 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
22207 low/high where such is correct only for be.
22208 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
22209 little-endian for -mcall-aixdesc.
22211 2013-04-25 Alan Modra <amodra@gmail.com>
22213 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22214 replace_equiv_address_nv.
22216 2013-04-25 Alan Modra <amodra@gmail.com>
22218 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22220 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22223 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22224 * rtl.h (struct rtx_def): ...
22226 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22228 PR rtl-optimizations/57046
22229 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22230 for multi-reg splits.
22232 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22234 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22236 2013-04-24 Sterling Augustine <saugustine@google.com>
22238 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22239 (comp_dir_string, debug_str_dwo_section): New.
22240 (DEBUG_STR_DWO_SECTION): Rename to ...
22241 (DEBUG_DWO_STR_SECTION): ... this.
22242 (DEBUG_NORM_STR_SECTION): Delete.
22243 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22244 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22245 (find_AT_string): Move most logic to ...
22246 (find_AT_string_in_table): ... here. New.
22247 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22248 add_skeleton_AT_string. Delete logic.
22249 (output_skeleton_debug_sections): Remove call to
22250 add_top_level_skeleton_die_attrs.
22251 (add_comp_dir_attribute): Move logic to comp_dir_string.
22252 (dwarf2out_init): Initialize debug_str_dwo_section.
22253 (output_indirect_string): Call find_string_form.
22254 (output_indirect_strings): Rewrite.
22255 (prune_unused_types): Empty skeleton_debug_str_hash.
22256 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22257 (dwarf2out_finish): Call output_indirect_strings.
22259 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22261 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22263 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22265 * rtl.h (struct rtx_def): Add comment for field jump.
22266 (LRA_SUBREG_P): New macro.
22267 * recog.c (register_operand): Check LRA_SUBREG_P.
22268 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22270 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22271 (lra_final_code_change): Skip subreg change for operators.
22272 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22273 if there are no operand changes.
22274 * lra-constraints.c (curr_insn_set): New.
22275 (match_reload): Set LRA_SUBREG_P.
22276 (emit_spill_move): Ditto.
22277 (check_and_process_move): Use curr_insn_set. Process only single
22278 set insns. Don't initialize sec_mem_p and change_p.
22279 (simplify_operand_subreg): Use LRA_SUBREG_P.
22280 (reg_in_class_p): New function.
22281 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22282 of #ifdef. Add code to remove cycling.
22283 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22284 non-null disp. Reload inner instead of disp when base and index
22285 are null. Try to put lo_sum into register.
22286 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22287 (check_and_process_move): Move code for move cost check to
22288 simple_move_p. Remove equiv_substitution.
22289 (simple_move_p): New function.
22290 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22291 curr_insn_set. Call check_and_process_move only for single set
22292 insns. Use the new function. Move call of check_and_process_move
22293 after operand equiv substitution and address process.
22295 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22297 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22298 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22299 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22301 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22303 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22305 2013-04-24 Marek Polacek <polacek@redhat.com>
22307 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22308 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22309 (select_loops_exit_conditions): Likewise.
22310 (number_of_iterations_for_all_loops): Likewise.
22311 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22312 (scev_analysis): Likewise.
22314 2013-04-02 Catherine Moore <clm@codesourcery.com>
22315 Chao-ying Fu <fu@mips.com>
22317 * config/mips/micromips.md (jraddiusp): New pattern.
22318 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22319 instruction if possible.
22321 2013-04-24 Alan Modra <amodra@gmail.com>
22323 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22325 2013-04-24 Julian Brown <julian@codesourcery.com>
22326 Chung-Lin Tang <cltang@codesourcery.com>
22328 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22329 dependency behavior in enumeration type DIE generation. Add TODO note
22330 to comments about future DW_FORM_sdata/udata re-work of related code.
22332 2013-04-23 Lawrence Crowl <crowl@google.com>
22334 * Makefile.in: Update as needed below.
22336 * hash-table.h (class hash_table):
22337 Correct many methods with parameter types compare_type to the correct
22338 value_type. (Correct code was unlikely to notice the change.)
22339 (hash_table::elements_with_deleted) New.
22340 (class hashtable::iterator): New.
22341 (hashtable::begin()): New.
22342 (hashtable::end()): New.
22343 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22345 * statistics.c (statistics_hashes):
22346 Change type to hash_table. Update dependent calls and types.
22348 * tree-into-ssa.c (var_infos):
22349 Change type to hash_table. Update dependent calls and types.
22351 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22352 Change type to hash_table. Update dependent calls and types.
22354 * tree-ssa-loop-im.c (struct mem_ref.refs):
22355 Change type to hash_table. Update dependent calls and types.
22357 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22358 Change type to hash_table. Update dependent calls and types.
22360 * tree-ssa-sccvn.c (vn_tables_s::nary):
22361 Change type to hash_table. Update dependent calls and types.
22362 (vn_tables_s::phis): Likewise.
22363 (vn_tables_s::references): Likewise.
22365 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22366 (vn_reference_eq): Update parameter and return types.
22368 * tree-ssa-structalias.c (pointer_equiv_class_table):
22369 Change type to hash_table. Update dependent calls and types.
22370 (location_equiv_class_table): Likewise.
22372 * tree-vect-data-refs.c: Consequential changes for making
22373 peeling a hash_table.
22375 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22376 (destroy_loop_vec_info): Dependent hash_table update.
22378 * tree-vectorizer.h (peeling_htab):
22379 Change type to hash_table. Update dependent calls and types.
22381 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22383 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22384 to check the register content is equal or not.
22385 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22386 to assign register content record.
22387 * lra-eliminations.c (update_reg_eliminate): Use
22388 lra_update_reg_val_offset to update register content offset.
22389 * lra-int.h (struct lra_reg): Add offset member.
22390 (lra_reg_val_equal_p): New static inline function.
22391 (lra_update_reg_val_offset): New static inline function.
22392 (lra_assign_reg_val): New static inline function.
22393 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22394 to assign register content record.
22395 (initialize_lra_reg_info_element): Initial offset to zero.
22397 2013-04-23 Catherine Moore <clm@codesourcery.com>
22399 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22400 operands. Record compression.
22402 2013-04-23 Xinliang David Li <davidxl@google.com>
22404 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22406 2013-04-23 Richard Biener <rguenther@suse.de>
22408 PR middle-end/57036
22409 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22410 parameter, only add abnormal goto edges from the copied body
22411 if the call could perform abnormal gotos.
22412 (copy_cfg_body): Adjust.
22414 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22416 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22418 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22420 * coretypes.h (gimple_stmt_iterator): Add struct to make
22423 2013-04-23 Richard Biener <rguenther@suse.de>
22425 PR tree-optimization/57026
22426 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22427 from SSA names occuring in abnormal PHI nodes.
22429 2013-04-22 Andi Kleen <ak@linux.intel.com>
22431 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22433 2013-04-22 Andi Kleen <ak@linux.intel.com>
22435 * lto/lto.c (print_lto_report_1): Declare early.
22436 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22438 2013-04-22 Andi Kleen <ak@linux.intel.com>
22440 * common.opt (-flto-report-wpa): Add.
22441 * doc/invoke.texi (-flto-report-wpa): Add.
22442 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22445 2013-04-22 Xinliang David Li <davidxl@google.com>
22447 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22448 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22449 * Makefile.in: New dependency
22451 David Daney <ddaney.cavm@gmail.com>
22453 * configure.ac (gcc_cv_as_micromips_support): Use the
22454 --fatal-warnings option.
22455 * configure: Regenerate.
22457 2013-04-22 Marek Polacek <polacek@redhat.com>
22460 * tsan.c (instrument_expr): Don't instrument expression
22461 in case its size is zero.
22463 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22467 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22469 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22471 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22473 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22474 (gimple_stmt_iterator): New typedef.
22475 * gimple.h (gimple_stmt_iterator): Rename to...
22476 (gimple_stmt_iterator_d): ... This.
22477 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22478 trees be valid for GIMPLE and GENERIC.
22479 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22480 * gimple-fold.c (gimple_fold_call): Call target hook
22481 gimple_fold_builtin.
22482 * hooks.c (hook_bool_gsiptr_false): New.
22483 * hooks.h (hook_bool_gsiptr_false): New.
22484 * target.def (fold_stmt): New.
22485 * doc/tm.texi: Regenerate.
22487 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22490 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22491 a set sp if no stack realignment.
22493 2013-04-22 Nick Clifton <nickc@redhat.com>
22495 * config.gcc (tilegx-linux): Extend extra_objs rather than
22497 (tilepro-linux): Likewise.
22499 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22501 * config/aarch64/aarch64-builtins.c
22503 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22504 (VAR<1-12>): Add MAP parameter.
22505 (BUILTIN_*): Likewise.
22506 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22507 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22508 (aarch64_ushl_n<mode>): Likewise.
22509 (aarch64_sshr_n<mode>): Likewise.
22510 (aarch64_ushr_n<mode>): Likewise.
22511 (aarch64_<maxmin><mode>): Likewise.
22512 (aarch64_sqrt<mode>): Likewise.
22513 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22514 (vshr<q>_n_*): Likewise.
22516 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22518 * config/aarch64/aarch64-builtins.c
22519 (aarch64_simd_builtin_type_mode): Handle SF types.
22521 (BUILTIN_GPF): Define.
22522 (aarch64_init_simd_builtins): Handle SF types.
22523 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22524 (frecps): Likewise.
22525 (frecpx): Likewise.
22526 * config/aarch64/aarch64-simd.md
22527 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22528 (aarch64_frecpe<mode>): New.
22529 (aarch64_frecps<mode>): Likewise.
22530 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22531 (v8type): Add frecp<esx>.
22532 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22533 (aarch64_frecps<mode>): Likewise.
22534 * config/aarch64/iterators.md (FRECP): New.
22535 (frecp_suffix): Likewise.
22536 * config/aarch64/arm_neon.h
22537 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22539 2013-04-22 Christian Bruel <christian.bruel@st.com>
22542 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22543 (REG_CLASS_NAMES): Idem.
22544 (REG_CLASS_CONTENTS): Idem.
22545 (REGCLASS_HAS_FP_REG): Idem.
22546 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22547 (sh_conditional_register_usage): Idem.
22549 2013-04-21 Jeff Law <law@redhat.com>
22551 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22552 (ssa_forward_propagate_and_combine): Use it.
22554 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22556 * lra.c: Update the flow chart diagram.
22558 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22560 PR rtl-optimization/56847
22561 * lra-constraints.c (process_alt_operands): Discourage alternative
22562 with non-matche doffsettable memory constraint fro memory with
22565 2013-04-19 Richard Biener <rguenther@suse.de>
22567 PR tree-optimization/56982
22568 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22570 * gimplify.c (gimplify_call_expr): Notice special calls.
22571 (gimplify_modify_expr): Likewise.
22572 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22573 abnormal control flow receivers.
22574 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22575 in the same way as cfun->has_nonlocal_labels.
22576 (gimple_purge_dead_abnormal_call_edges): Likewise.
22577 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22578 receivers start a basic-block.
22580 2013-04-19 Richard Biener <rguenther@suse.de>
22582 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22584 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22585 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22586 (SLP_TREE_LOAD_PERMUTATION): Add.
22587 (vect_transform_slp_perm_load): Adjust prototype.
22588 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22589 (vect_free_slp_instance): Likewise.
22590 (vect_create_new_slp_node): Likewise.
22591 (vect_supported_slp_permutation_p): Remove.
22592 (vect_slp_rearrange_stmts): Adjust.
22593 (vect_supported_load_permutation_p): Likewise. Inline
22594 vect_supported_slp_permutation_p here.
22595 (vect_analyze_slp_instance): Compute load permutations per
22596 slp node instead of per instance.
22597 (vect_get_slp_defs): Adjust.
22598 (vect_transform_slp_perm_load): Likewise.
22599 (vect_schedule_slp_instance): Remove redundant code.
22600 (vect_schedule_slp): Remove hack for PR56270, add it ...
22601 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22602 CSE loads for SLP. Adjust.
22604 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22606 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22607 spelling in two comments.
22609 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22612 * config/arm/arm.c (load_multiple_sequence): Require SP
22613 as base register for loads if SP is in the register list.
22615 2013-04-19 Martin Jambor <mjambor@suse.cz>
22617 PR tree-optimization/56718
22618 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22619 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22620 and made public. Adjusted all callers.
22621 (ipa_intraprocedural_devirtualization): New function.
22622 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22623 (ipa_intraprocedural_devirtualization): Likewise.
22624 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22626 2013-04-19 Richard Biener <rguenther@suse.de>
22628 PR tree-optimization/57000
22629 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
22631 2013-04-19 Terry Guo <terry.guo@arm.com>
22633 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
22635 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
22636 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
22637 (cortex_m4_fmacs): Use new reservations.
22638 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
22640 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
22642 PR rtl-optimization/56999
22643 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
22645 (lra_coalesce): Remove split_origin_bitmap and related code.
22646 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
22647 ranges if necessary.
22649 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
22651 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
22653 (ix86_expand_call): Remove clobbered_registers array and use
22654 x86_64_ms_sysv_extra_clobbered_registers instead.
22655 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
22657 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
22659 * config/i386/i386.md (*call_rex64_ms_sysv): Use
22660 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
22661 (*call_value_rex64_ms_sysv): Ditto.
22663 2013-04-18 Cary Coutant <ccoutant@google.com>
22665 * dwarf2out.c (output_pubnames): Check die_perennial_p of
22666 parent instead of die_mark.
22668 2013-04-18 Diego Novillo <dnovillo@google.com>
22670 * gimple.c (create_gimple_tmp): New.
22671 (get_expr_type): New.
22672 (build_assign): New.
22673 (build_type_cast): New.
22674 * gimple.h (enum ssa_mode): Define.
22675 (gimple_seq_set_location): New.
22676 * asan.c (build_check_stmt): Change some gimple_build_* calls
22677 to use build_assign and build_type_cast.
22679 2013-04-18 Richard Biener <rguenther@suse.de>
22681 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
22682 handle negative step. Remove redundant checks.
22683 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
22684 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
22685 for negative step and grouped loads fail to vectorize.
22687 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22689 * emit-rtl.c (reset_insn_used_flags): New function.
22690 (reset_all_used_flags): Use it.
22691 (verify_insn_sharing): New function.
22692 (verify_rtl_sharing): Fix verification for SEQUENCEs.
22694 2013-04-18 Jakub Jelinek <jakub@redhat.com>
22696 PR tree-optimization/56984
22697 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
22698 and (x >> M) >= N don't register any assertion if N << M is the
22701 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22703 * lower-subreg.c (resolve_simple_move): If called self-recursive,
22704 do not delete_insn insns that have not yet been emitted, only
22705 unlink them with remove_insn.
22706 * df-scan.c (df_insn_delete): Revert r197492.
22708 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
22710 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
22711 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
22713 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22715 * config/arm/arm.md (movsicc_insn): Convert define_insn into
22716 define_insn_and_split.
22717 (and_scc,ior_scc,negscc): Likewise.
22718 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
22720 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22722 * config/arm/arm.c (use_return_insn): Return 0 for targets that
22723 can benefit from using a sequence of LDRD instructions in epilogue
22724 instead of a single LDM instruction.
22726 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
22729 * doc/extend.texi: Fix typo.
22731 2013-04-17 Richard Biener <rguenther@suse.de>
22733 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
22734 (vect_build_slp_tree): ... here.
22735 (vect_build_slp_tree_1): Compute which stmts of the SLP group
22736 match. Remove special-casing of mismatched complex loads.
22737 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
22738 re-try the match with swapped commutative operands.
22739 (vect_supported_load_permutation_p): Remove special-casing of
22740 mismatched complex loads.
22741 (vect_analyze_slp_instance): Adjust.
22743 2013-04-17 Richard Biener <rguenther@suse.de>
22745 PR rtl-optimization/56921
22746 * cfgloop.h (struct loop): Add simple_loop_desc member.
22747 (struct niter_desc): Mark with GTY(()).
22748 (simple_loop_desc): Do not use aux field but simple_loop_desc.
22749 * loop-iv.c (get_simple_loop_desc): Likewise.
22750 (free_simple_loop_desc): Likewise.
22753 2013-04-16 Richard Biener <rguenther@suse.de>
22755 PR rtl-optimization/56921
22756 * loop-init.c (pass_rtl_move_loop_invariants): Add
22757 TODO_do_not_ggc_collect to todo_flags_finish.
22758 (pass_rtl_unswitch): Same.
22759 (pass_rtl_unroll_and_peel_loops): Same.
22760 (pass_rtl_doloop): Same.
22762 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
22764 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
22765 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
22766 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
22768 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
22769 * tree-streamer.c (record_common_node): Adjust reference in comment.
22771 2013-04-17 Terry Guo <terry.guo@arm.com>
22773 * config/arm/cortex-m4.md: Add a new bypass.
22775 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22777 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
22779 (*subs_<optab><mode>_multp2): New pattern.
22780 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22781 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22783 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22785 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
22786 (*subs_mul_imm_<mode>): New pattern.
22788 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
22791 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
22792 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
22793 (vsx_movti_32bit): Change j->wa to O->wa.
22795 2013-04-16 Richard Biener <rguenther@suse.de>
22797 PR rtl-optimization/56921
22798 * loop-init.c (pass_rtl_move_loop_invariants): Add
22799 TODO_do_not_ggc_collect to todo_flags_finish.
22800 (pass_rtl_unswitch): Same.
22801 (pass_rtl_unroll_and_peel_loops): Same.
22802 (pass_rtl_doloop): Same.
22804 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
22806 * config/arm/arm.c (emit_multi_reg_push): New declaration
22807 for an existing function.
22808 (arm_emit_strd_push): New function.
22809 (arm_expand_prologue): Used here.
22810 (arm_emit_ldrd_pop): New function.
22811 (arm_expand_epilogue): Used here.
22812 (arm_get_frame_offsets): Update condition.
22813 (arm_emit_multi_reg_pop): Add a special case for load of a single
22814 register with writeback.
22816 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22818 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22821 2013-04-16 Richard Biener <rguenther@suse.de>
22823 PR tree-optimization/56756
22824 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22825 (first_mem_ref_loc): New.
22826 (execute_sm): Place the load temporarily before a previous
22827 access instead of in the latch edge to ensure its SSA dependencies
22828 are defined at points dominating the load.
22830 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22832 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22833 correct fix by moving header and footer insn to the footer of
22834 the merged basic block. Clear BB_END of the merged-away block.
22836 PR middle-end/43631
22837 * emit-rtl.c (make_note_raw): New function.
22838 (link_insn_into_chain): New static inline function.
22839 (add_insn): Use it.
22840 (add_insn_before, add_insn_after): Factor insn chain linking code...
22841 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22842 using link_insn_into_chain.
22843 (note_outside_basic_block_p): New helper function for emit_note_after
22844 and emit_note_before.
22845 (emit_note_after): Use nobb variant of add_insn_after if the note
22846 should not be contained in a basic block.
22847 (emit_note_before): Use nobb variant of add_insn_before if the note
22848 should not be contained in a basic block.
22849 (emit_note_copy): Use make_note_raw.
22850 (emit_note): Likewise.
22851 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
22852 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
22853 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
22854 the moved barrier the tail of the basic block it follows.
22855 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
22857 2013-04-15 Jakub Jelinek <jakub@redhat.com>
22859 PR tree-optimization/56962
22860 * gimple-ssa-strength-reduction.c (record_increment): Only set
22861 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
22862 either rhs1 or rhs2 is equal to c->base_expr.
22864 2013-04-15 Richard Biener <rguenther@suse.de>
22866 PR tree-optimization/56933
22867 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
22869 (GROUP_READ_WRITE_DEPENDENCE): Remove.
22870 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
22871 * tree-vect-data-refs.c (vect_analyze_group_access): Move
22872 dependence check ...
22873 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
22875 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
22876 GROUP_READ_WRITE_DEPENDENCE.
22878 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22880 * emit-rtl.c (reset_all_used_flags): New function.
22881 (verify_rtl_sharing): Call reset_all_used_flags before and after
22882 performing the checks.
22884 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22886 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
22887 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
22888 * config/arm/constraints.md (De): New constraint.
22889 * config/arm/neon.md (anddi3_neon): Delete.
22890 (neon_vand<mode>): Expand to standard anddi3 pattern.
22891 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
22892 Move earlier in the file.
22893 (neon_inv_logic_op2): Likewise.
22894 (arm_anddi_operand_neon): New predicate.
22896 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22898 * configure.ac (gcc_cv_ld_as_needed): Set
22899 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
22900 Use -z ignore, -z record on *-*-solaris2*.
22901 (HAVE_LD_AS_NEEDED): Update comment.
22902 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
22903 * configure: Regenerate.
22904 * config.in: Regenerate.
22905 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
22906 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
22907 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
22908 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
22909 equivalents. Fix markup.
22910 * doc/tm.texi: Regenerate.
22912 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
22914 * config/i386/i386.opt: New option mstack-protector-guard=.
22915 * config/i386/i386-opts.h: Add enum stack_protector_guard.
22916 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
22917 TARGET_SSP_TLS_GUARD.
22918 * config/i386/i386.c (ix86_option_override_internal): Set
22919 ix86_stack_protector_guard.
22920 * config/i386/i386.md (stack_protect_set): Enable for
22921 TARGET_SSP_TLS_GUARD only.
22922 (stack_protect_set_<mode>): Ditto.
22923 (stack_protect_test): Ditto.
22924 (stack_protect_test_<mode>): Ditto.
22925 * doc/invoke.texi (i386 Option): Document.
22927 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
22930 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
22931 (S_MODES): Set H_MODE bit.
22932 (SF_MODES): Set only S_MODE and SF_MODE bits.
22933 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
22934 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
22935 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
22936 <MODE_FLOAT>: Likewise.
22938 2013-04-15 Joey Ye <joey.ye@arm.com>
22940 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
22942 2013-04-15 Joey Ye <joey.ye@arm.com>
22944 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
22946 (thumb_far_jump_used_p): Count instruction size and set
22949 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
22951 * reorg.c (fill_simple_delay_slots): Reindent block of code.
22952 * resource.c (mark_target_live_regs): Reformat conditional block.
22954 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22956 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
22957 notes, they are emitted only just before final.
22958 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
22960 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22962 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
22963 * cfgrtl.c (delete_insn): Call it here instead.
22964 * lra-spills.c (lra_final_code_change): Use delete_insn.
22965 * haifa-sched.c (sched_remove_insn): Likewise.
22966 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
22967 returning to the nop pool.
22968 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
22969 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
22971 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22973 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
22974 * doc/tm.texi: Regenerated.
22976 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
22978 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
22981 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22983 * df-core.c (df_find_def): Compare register numbers.
22984 (df_find_use): Likewise.
22986 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
22989 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
22990 lra_in_progress for return.
22992 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22994 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
22995 define_insn into define_insn_and_split and emit movsicc patterns.
22997 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22999 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
23001 2013-04-12 Richard Biener <rguenther@suse.de>
23003 * tree-pass.h (TODO_do_not_ggc_collect): New.
23004 * passes.c (execute_one_ipa_transform_pass): Honor
23005 TODO_do_not_ggc_collect.
23006 (execute_one_pass): Likewise.
23009 2013-04-10 Richard Biener <rguenther@suse.de>
23011 * passes.c (init_optimization_passes): Remove reload pass.
23012 * ira.c (do_reload): Merge into ...
23014 (rest_of_handle_reload): Remove.
23015 (pass_reload): Likewise.
23016 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23017 of reload for vzeroupper pass placement.
23019 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23021 PR tree-optimization/56918
23022 PR tree-optimization/56920
23023 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23024 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23025 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23026 use wide_mul_with_sign method.
23028 2013-04-12 Richard Biener <rguenther@suse.de>
23030 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23031 not be considered a gimple constant.
23033 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23035 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23036 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23039 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23040 Jakub Jelinek <jakub@redhat.com>
23042 * opts.c: Include diagnostic-color.h.
23043 (common_handle_option): Handle OPT_fdiagnostics_color_.
23044 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23045 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23046 (diagnostic-color.o): New.
23047 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23048 (diagnostic_color_rule): New enum.
23049 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23050 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23051 the location string.
23052 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23053 either NULL, or color kind.
23054 * diagnostic-color.c: New file.
23055 * diagnostic-color.h: New file.
23056 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23058 * doc/invoke.texi (-fdiagnostics-color): Document.
23059 * pretty-print.h (pp_show_color): Define.
23060 (struct pretty_print_info): Add show_color field.
23061 * diagnostic.c: Include diagnostic-color.h.
23062 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23063 macros. Colorize error:, warning: etc. strings and also the location
23065 (diagnostic_show_locus): Colorize the caret line.
23066 * pretty-print.c: Include diagnostic-color.h.
23067 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23068 inside of %< %> quotes or quoted through q format modifier.
23070 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23072 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23074 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23076 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23077 code in CC_NZ mode.
23078 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23081 2013-04-11 Marek Polacek <polacek@redhat.com>
23083 PR tree-optimization/48184
23084 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23086 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23088 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23089 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23090 (skip_simple_arithmetic): Tidy up.
23091 * tree.h (skip_simple_constant_arithmetic): Declare.
23093 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23095 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23097 2013-04-11 Richard Biener <rguenther@suse.de>
23099 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23100 generate vector constants.
23102 2013-04-11 Richard Biener <rguenther@suse.de>
23104 PR tree-optimization/56878
23105 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23106 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23108 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23109 Prefer to align the DR with the most invariant base address.
23111 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23113 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23116 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23118 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23119 floating-point vector comparisons against 0.
23121 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23123 PR tree-optimization/56899
23124 * fold-const.c (extract_muldiv_1): Apply distributive law
23125 only if TYPE_OVERFLOW_WRAPS (ctype).
23127 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23130 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23131 loading has side effect.
23133 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23135 * config/sparc/sparc.c: Include tree-pass.h.
23136 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23137 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23138 head of file. Change return type. Split off gate function.
23139 (sparc_gate_work_around_errata): New function.
23140 (pass_work_around_errata): New pass definition.
23141 (insert_pass_work_around_errata) New pass insert definition to
23142 insert pass_work_around_errata just after delayed-branch scheduling.
23143 (sparc_option_override): Insert the pass.
23144 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23146 2013-04-10 David S. Miller <davem@davemloft.net>
23148 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23149 or -mcpu=hypersparc.
23151 * target.def (cstore_mode): New hook.
23152 * target.h: Include insn-codes.h
23153 * targhooks.c: Likewise.
23154 (default_cstore_mode): New function.
23155 * targhooks.h: Declare it.
23156 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23157 * doc/tm.texi: Rebuild.
23158 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23159 target hook, rather than inspecting the insn_data.
23160 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23161 (TARGET_CSTORE_MODE): Redefine.
23162 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
23164 * config/sparc/predicates.md (cstore_result_operand): New special
23166 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
23167 Use it for operand 0.
23168 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
23169 (*snesi_special): Likewise.
23170 (*snesi_zero): Likewise.
23171 (*seqsi_zero): Likewise.
23172 (*sltu_insn): Likewise.
23173 (*sgeu_insn): Likewise.
23174 (*seqdi_special): Make operand 0 and comparison operation be of
23176 (*snedi_special): Likewise.
23177 (*snedi_special_vis3): Likewise.
23178 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
23179 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
23180 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
23181 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
23182 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
23183 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
23184 (*sltu_extend_sp64): Likewise.
23185 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
23186 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
23187 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
23188 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
23189 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
23191 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
23193 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
23194 (aarch64_start_file): Use the new function.
23196 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23197 Jason Merrill <jason@redhat.com>
23199 * common.opt: Add -gdwarf.
23200 * opts.c (common_handle_option): Handle it.
23201 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
23203 2013-04-10 Richard Biener <rguenther@suse.de>
23205 * passes.c (execute_todo): Do not call ggc_collect conditional here.
23206 (execute_one_ipa_transform_pass): But unconditionally here.
23207 (execute_one_pass): And here.
23208 (init_optimization_passes): Remove reload pass.
23209 * tree-pass.h (TODO_ggc_collect): Remove.
23210 (pass_reload): Likewise.
23211 * ira.c (do_reload): Merge into ...
23213 (rest_of_handle_reload): Remove.
23214 (pass_reload): Likewise.
23215 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23216 of reload for vzeroupper pass placement.
23217 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23218 and todo_flags_finish of all passes.
23220 2013-04-10 Richard Biener <rguenther@suse.de>
23222 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23223 first_const_oprnd field, rename first_def_type to first_op_type.
23224 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23225 (vect_get_and_check_slp_defs): Always use the type of the
23226 operand. Allow mixed vect_external_def, vect_constant_def types.
23227 (vect_get_constant_vectors): Handle mixed vect_external_def,
23228 vect_constant_def types.
23230 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23232 PR tree-optimization/55524
23233 * tree-ssa-math-opts.c
23234 (convert_mult_to_fma): Don't use an fms construct
23235 when we don't have an fms operation, but fnma, and it looks
23236 likely that we'll be able to use the latter.
23238 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23240 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23242 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23243 inline fail caused by overwritable functions.
23245 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23247 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23248 unnecessary bits in the constant power of two case.
23250 2013-04-10 Richard Biener <rguenther@suse.de>
23252 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23253 broken code swapping operands.
23254 (vect_build_slp_tree): Do not compute load permutations here.
23255 (vect_analyze_slp_instance): Compute load permutations here,
23256 after building the SLP tree.
23258 2013-04-09 Christian Bruel <christian.bruel@st.com>
23260 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23261 of next/prev_real_insn.
23263 2013-04-09 Jan Hubicka <jh@suse.cz>
23265 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23266 Drop aliased parameter.
23267 (function_and_variable_visibility): Do not handle alias pairs.
23268 * cgraph.c (varpool_externally_visible_p): Update prototype.
23269 * varpool.c (varpool_add_new_variable): Update.
23271 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23273 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23275 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23277 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23279 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23281 2013-04-09 Marek Polacek <polacek@redhat.com>
23283 PR tree-optimization/48762
23284 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23286 2013-04-09 Richard Biener <rguenther@suse.de>
23288 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23290 (vect_build_slp_tree): Likewise.
23291 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23292 calculating the cost of a SLP instance.
23293 (vect_analyze_slp_instance): Use it from here, after building
23296 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23298 PR middle-end/56883
23299 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23300 expand_omp_for_static_chunk): Use simple_p = true in
23301 force_gimple_operand_gsi calls when assigning to addressable decls.
23303 2013-04-09 Jeff Law <law@redhat.com>
23305 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23306 when the boolean was created by converting a wider object which
23307 had a boolean range.
23309 2013-04-09 Richard Biener <rguenther@suse.de>
23311 * tree-vectorizer.h (slp_void_p): Remove.
23312 (slp_tree): Typedef before _slp_tree declaration.
23313 (struct _slp_tree): Use a vector of slp_tree as children.
23314 (vect_get_place_in_interleaving_chain): Remove.
23315 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23317 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23319 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23320 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23321 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23322 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23323 Use slp_node instead of slp_void_p and adjust.
23325 2013-04-09 Richard Biener <rguenther@suse.de>
23327 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23328 work that is not necessary.
23330 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23332 PR tree-optimization/56854
23333 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23334 forward into clobber stmts if it would change MEM_REF lhs into
23337 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23339 * tree.c (type_hash_lookup, type_hash_add): Make static.
23340 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23342 2013-04-09 Richard Biener <rguenther@suse.de>
23344 * tree.h (unsave_expr_now): Remove.
23345 * tree-inline.c (mark_local_for_remap_r): Remove.
23346 (unsave_expr_1): Likewise.
23347 (unsave_r): Likewise.
23348 (unsave_expr_now): Likewise.
23349 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23350 (propagate_tree_value): Likewise.
23352 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23354 * doc/rtl.texi (sequence): Rewrite documentation to match the
23355 current use of SEQUENCE rtl objects.
23356 * rtl.def (SEQUENCE): Likewise.
23358 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23359 Update documentation.
23360 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23361 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23363 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23365 2013-04-08 Teresa Johnson <tejohnson@google.com>
23367 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23368 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23370 (estimate_edge_size_and_time): Add comment to suggest using rounding
23372 (estimate_node_size_and_time): Ditto.
23373 (remap_edge_change_prob): Use helper rounding divide methods.
23374 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23375 (gimple_mod_pow2_value_transform): Ditto.
23376 (gimple_mod_subtract_transform): Ditto.
23377 (gimple_ic_transform): Ditto.
23378 (gimple_stringops_transform): Ditto.
23379 * stmt.c (conditional_probability): Ditto.
23380 (emit_case_dispatch_table): Ditto.
23381 * lto-cgraph.c (merge_profile_summaries): Ditto.
23382 * tree-optimize.c (execute_fixup_cfg): Ditto.
23383 * cfgcleanup.c (try_forward_edges): Ditto.
23384 * cfgloopmanip.c (scale_loop_profile): Ditto.
23386 (duplicate_loop_to_header_edge): Ditto.
23387 (lv_adjust_loop_entry_edge): Ditto.
23388 * tree-vect-loop.c (vect_transform_loop): Ditto.
23389 * profile.c (compute_branch_probabilities): Ditto.
23390 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23391 * lto-streamer-in.c (input_cfg): Ditto.
23392 * gimple-streamer-in.c (input_bb): Ditto.
23393 * ipa-cp.c (update_profiling_info): Ditto.
23394 (update_specialized_profile): Ditto.
23395 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23396 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23398 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23399 (compute_trg_info): Ditto.
23400 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23401 (purge_dead_edges): Ditto.
23402 * loop-unswitch.c (unswitch_loop): Ditto.
23403 * cgraphclones.c (cgraph_clone_edge): Ditto.
23404 (cgraph_clone_node): Ditto.
23405 * tree-inline.c (copy_bb): Ditto.
23406 (copy_edges_for_bb): Ditto.
23407 (initialize_cfun): Ditto.
23408 (copy_cfg_body): Ditto.
23409 (expand_call_inline): Ditto.
23411 2013-04-08 Kai Tietz <ktietz@redhat.com>
23413 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23414 TARGET_CYGWIN64 by TARGET_64BIT.
23416 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23418 * config/epiphany/epiphany.md (GPR_1): New constant.
23419 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23420 * config/epiphany/epiphany.c (gen_compare_reg):
23421 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23422 is already in place.
23423 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23424 Don't require being called during rtl expansion; If y operlaps r0,
23426 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23427 (epiphany_expand_epilogue): Likewise.
23429 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23430 Don't use CC_FPmode for ORDERED / UNORDERED.
23431 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23433 * config/epiphany/constraints.md (CnL): New constraint.
23434 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23435 * config/epiphany/predicates.md (add_operand): Allow 1024.
23437 * config/epiphany/epiphany.md (logical_op): New code iterator.
23438 (op_mnc): New code attribute.
23439 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23440 (mov_f+1, mov_f+2): New peephole2 patterns.
23442 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23443 (cstoresi4): Also allow re-use of zero result when doing a NE
23444 comparison to a non-zero operand.
23445 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23447 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23448 Use gen_addsi3_i / gen_subsi3_i.
23450 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23454 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23455 contain anything but clobbers, at most one __builtin_stack_restore,
23456 optionally debug stmts and final resx, and if it has at least one
23457 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23458 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23459 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23460 which isn't defaut definition, remove them.
23461 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23462 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23463 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23464 with MEM_REF LHS with SSA_NAME address.
23466 2013-04-08 Jeff Law <law@redhat.com>
23468 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23470 2013-04-08 Richard Biener <rguenther@suse.de>
23472 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23474 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23475 determined vector type.
23476 (vect_analyze_data_refs): Likewise.
23477 (vect_get_new_vect_var): Adjust.
23478 (vect_create_destination_var): Preserve SSA name versions.
23479 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23480 not dump anything here.
23482 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23484 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23485 Add member lr_slot_known.
23486 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23488 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23489 Remove code that sets lr_slot_offset according to what a previous
23490 version of epiphany_emit_save_restore used to do.
23491 (epiphany_emit_save_restore): When doing an lr save or restore,
23492 set/verify lr_slot_known and lr_slot_offset.
23494 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23497 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23500 2013-04-08 Richard Biener <rguenther@suse.de>
23502 * alias.c (find_base_term): Fix thinko in previous change.
23504 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23506 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23507 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23508 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23509 if possible to compute val.
23510 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23511 For QImode integers don't require anything about precision. Use
23512 const_with_all_bytes_same to find out if the constant doesn't have
23513 repeated bytes in it.
23515 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23517 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23520 2013-04-08 Marek Polacek <polacek@redhat.com>
23522 PR rtl-optimization/48182
23523 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23526 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23529 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23530 nuses, make sure we have a label.
23532 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23535 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23536 (rs6000_emit_swdiv_low_precision): Remove.
23537 (rs6000_emit_swdiv): Rewrite to handle between one and four
23538 iterations of Newton-Raphson generally; modify required number of
23539 iterations for some cases.
23540 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23542 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23544 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23545 set-but-unused variable.
23547 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23548 basic blocks of released function bodies garbage-collectable.
23550 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23551 (struct rtl_opt_pass): Add TODO_df_finish.
23553 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23555 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23557 * config/arm/constraints.md (q): New constraint.
23558 * config/arm/ldrdstrd.md: New file.
23559 * config/arm/arm.md (ldrdstrd.md) New include.
23560 (arm_movdi): Use "q" instead of "r" constraint
23561 for double-word memory access.
23562 (movdf_soft_insn): Likewise.
23563 * config/arm/vfp.md (movdi_vfp): Likewise.
23564 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23565 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23566 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23567 (mem_ok_for_ldrd_strd): Likewise.
23568 (output_move_double): Update assertion.
23570 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23572 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23574 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23576 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23577 define_insn_and_split.
23578 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23580 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23582 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23583 define_insn_and_split.
23584 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23585 (shiftsi3_compare): New pattern.
23586 (rrx): New pattern.
23587 * config/arm/unspecs.md (UNSPEC_RRX): New.
23589 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23591 * config/arm/arm.md (negdi_extendsidi): New pattern.
23592 (negdi_zero_extendsidi): Likewise.
23594 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23596 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23597 define_insn_and_split.
23598 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23599 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23601 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23603 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23604 define_insn_and_split.
23605 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23606 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23608 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23610 * config/arm/arm.md (subsi3_carryin): New pattern.
23611 (subsi3_carryin_const): Likewise.
23612 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23613 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23615 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23617 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23619 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23621 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23622 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23624 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23626 * config/arm/arm.c (arm_expand_builtin): Change fcode
23627 type to unsigned int.
23629 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23631 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
23633 2013-04-04 Ian Lance Taylor <iant@google.com>
23635 * doc/standards.texi (Standards): The Go frontend supports the Go 1
23638 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
23640 PR middle-end/56729
23641 * df-scan.c (df_insn_delete): Disable failing assert.
23643 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23645 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
23646 New function prototype.
23647 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
23648 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
23649 (arm_builtin_vectorized_function): New function.
23651 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23653 * config/arm/arm_neon_builtins.def: New file.
23654 * config/arm/arm.c (neon_builtin_data): Move contents to
23655 arm_neon_builtins.def.
23656 (enum arm_builtins): Include neon builtin definitions.
23657 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
23658 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
23660 2013-04-04 Marek Polacek <polacek@redhat.com>
23662 PR tree-optimization/48186
23663 * predict.c (maybe_hot_frequency_p): Return false if
23664 HOT_BB_FREQUENCY_FRACTION is 0.
23665 (cgraph_maybe_hot_edge_p): Likewise.
23667 2013-04-04 Richard Biener <rguenther@suse.de>
23669 PR tree-optimization/56826
23670 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
23673 2013-04-04 Richard Biener <rguenther@suse.de>
23675 PR tree-optimization/56213
23676 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
23677 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
23679 2013-04-04 Richard Biener <rguenther@suse.de>
23681 PR tree-optimization/56837
23682 * tree-loop-distribution.c (classify_partition): For non-zero
23683 values require that the value has the same precision as its
23684 mode to be useful as memset value.
23686 2013-04-03 Nick Clifton <nickc@redhat.com>
23688 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
23689 (fmssf4): Use fmsf.s on E3V5 architectures.
23690 (fnmasf4): Use fnmaf.s on E3V5 architectures.
23691 (fnmssf4): Use fnmsf.s on E3V5 architectures.
23693 2013-04-03 Jeff Law <law@redhat.com>
23695 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
23696 (lra-eliminations.o): Likewise.
23698 2013-04-03 Teresa Johnson <tejohnson@google.com>
23700 * gcov-io.c (compute_working_sets): Moved most of body of old
23701 compute_working_sets here from profile.c.
23702 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
23703 (gcov_working_set_t): Moved typedef here from basic-block.h
23704 (compute_working_set): Declare.
23705 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
23706 (get_working_sets): Renamed from compute_working_set,
23707 replace most of body with call to new compute_working_sets.
23708 (get_exec_counts): Replace call to compute_working_sets
23709 to get_working_sets.
23710 * profile.h (get_working_sets): Renamed from compute_working_set.
23711 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
23712 to get_working_sets.
23713 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
23714 * gcov-dump.c (dump_working_sets): New function.
23716 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
23718 * hwint.c (sext_hwi, zext_hwi): New functions.
23719 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
23720 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
23721 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
23722 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
23723 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
23724 (sext_hwi, zext_hwi): New functions.
23726 2013-04-03 Jeff Law <law@redhat.com>
23728 PR tree-optimization/56799
23729 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
23730 back test for widening conversion erroneously dropped in prior change.
23732 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23735 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
23736 instead of next_real_insn.
23738 2013-04-03 Marek Polacek <polacek@redhat.com>
23741 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
23743 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23746 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
23748 (thumb1_output_casesi): Likewise.
23749 (thumb2_output_casesi): Likewise.
23751 2013-04-03 Richard Biener <rguenther@suse.de>
23753 PR tree-optimization/56817
23754 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
23756 (tree_unroll_loops_completely_1): ... new function to manually
23757 walk the loop tree, properly defering outer loops of unrolled
23758 loops to later iterations.
23760 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23762 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
23763 (vectorizable_load): Likewise.
23764 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
23765 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
23767 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23769 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
23772 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
23774 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
23776 2013-04-03 Bin Cheng <bin.cheng@arm.com>
23778 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
23780 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23782 PR tree-optimization/56790
23783 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
23786 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23788 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
23790 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
23791 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
23794 2013-04-03 Jakub Jelinek <jakub@redhat.com>
23797 * tree.h (force_folding_builtin_constant_p): New decl.
23798 * builtins.c (force_folding_builtin_constant_p): New variable.
23799 (fold_builtin_constant_p): Fold immediately also if
23800 force_folding_builtin_constant_p.
23802 2013-04-03 Richard Biener <rguenther@suse.de>
23804 PR tree-optimization/56812
23805 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
23806 DRs of the same interleaving chain are independent.
23808 2013-04-02 Jason Merrill <jason@redhat.com>
23810 * gdbinit.in (pbb): Use debug fn.
23812 2013-04-02 Lawrence Crowl <crowl@google.com>
23814 * sese.h (struct ivtype_map_elt_s): Remove unused.
23815 (extern debug_ivtype_map): Remove unused.
23816 (extern eq_ivtype_map_elts): Remove unused.
23817 * sese.c (debug_ivtype_map): Removed unused.
23818 (debug_ivtype_map_1): Removed unused.
23819 (debug_ivtype_elt): Remove unused.
23820 (eq_ivtype_map_elts): Remove unused.
23823 2013-04-02 Kai Tietz <ktietz@redhat.com>
23826 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23827 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23828 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23830 (legitimize_pe_coff_symbol): Likewise.
23831 (is_imported_p): New helper-function.
23832 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23833 for Windows x64 targets.
23834 (ix86_expand_prologue): Optimize for pe-coff targets.
23835 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23836 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23837 medium/large code-model.
23838 (legitimize_pic_address): Likewise.
23839 (legitimize_tls_address): Likewise.
23840 (ix86_expand_call): Likewise.
23841 (x86_output_mi_thunk): Likewise.
23842 (get_dllimport_decl): Add new beimport argument.
23843 (construct_plt_address): Don't assert for x64 pe-coff targets.
23844 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23846 (SYMBOL_FLAG_STUBVAR): New macro.
23847 (SYMBOL_REF_STUBVAR_P): Likewise.
23848 * config/i386/winnt.c (stub_list): New structure.
23849 (stub_head): New local variable.
23850 (i386_pe_record_stub): New function.
23851 (i386_pe_file_end): Emit refptr-stubs.
23853 2013-04-02 Jakub Jelinek <jakub@redhat.com>
23855 PR rtl-optimization/56745
23856 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
23857 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
23860 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
23861 and both of them are MEM_REFs, just compare first argument for
23862 equality and attempt to deal even with differing offsets.
23865 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
23866 of gimple_clobber_p to be MEM_REF.
23867 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
23868 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
23869 after gimplification.
23870 * asan.c (get_mem_ref_of_assignment): Don't instrument
23871 gimple_clobber_p stmts.
23872 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
23873 gimple_clobber_p stmt if they have MEM_REF lhs and
23874 are dead because of another gimple_clobber_p stmt.
23875 * tree-ssa-live.c (clear_unused_block_pointer): Treat
23876 gimple_clobber_p stmts like debug stmts.
23877 (remove_unused_locals): Remove clobbers with MEM_REF lhs
23878 that refer to unused VAR_DECLs or uninitialized values.
23879 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
23880 gimple_clobber_p stmts if they refer to removed parameters.
23881 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
23884 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
23886 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
23887 using SWI48 mode attribute.
23889 2013-04-02 Wei Mi <wmi@google.com>
23891 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
23892 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
23893 *<rotate_insn><mode>3_mask in i386.md.
23895 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
23897 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
23899 2013-04-02 Richard Biener <rguenther@suse.de>
23901 PR tree-optimization/56778
23902 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
23903 Runtime alias tests are not supported for gather loads.
23904 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
23905 stmts referenced from SSA operands before updating SSA form.
23907 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
23908 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23910 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
23911 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
23912 * config/arm/cortex-a53.md: New file.
23913 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
23914 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
23915 * config/arm/arm.c (arm_issue_rate): Likewise.
23916 * config/arm/arm-tune.md: Regenerate
23917 * config/arm/arm-tables.opt: Regenerate.
23918 * config/arm/arm-cores.def: Add cortex-a53.
23920 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
23922 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
23925 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23927 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
23928 scalar load/store operations using B/H registers.
23929 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
23931 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23933 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
23935 * config/aarch64/aarch64.c
23936 (aarch64_simd_scalar_immediate_valid_for_move): New.
23937 * config/aarch64/aarch64-protos.h
23938 (aarch64_simd_scalar_immediate_valid_for_move): New.
23939 * config/aarch64/constraints.md (Dh, Dq): New.
23940 * config/aarch64/iterators.md (hq): New.
23942 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
23944 * reorg.c (get_branch_condition): Deal with conditional returns.
23945 (fill_simple_delay_slots): Remove dead code dealing with jumps.
23947 2013-04-01 Wei Mi <wmi@google.com>
23949 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
23950 Truncate operand 2 using %b asm operand modifier.
23951 (*<shift_insn><mode>3_mask): Ditto.
23952 (*<rotate_insn><mode>3_mask): Ditto.
23954 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
23956 PR middle-end/56798
23957 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
23959 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
23961 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
23963 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
23965 2013-03-30 Lawrence Crowl <crowl@google.com>
23967 * dse.c (clear_alias_sets): Remove never set.
23968 (disqualified_clear_alias_sets): Remove never set.
23969 (clear_alias_mode_pool): Remove never set.
23970 (dse_step0): Remove condition that is never true.
23971 (canon_address): Remove condition that is never true.
23972 (dse_step7): Remove condition that is never true.
23973 (rest_of_handle_dse): Remove condition that is never true.
23974 (rest_of_handle_dse::did_global): Remove never read from above.
23975 (dse_step2_spill): Remove never called from above.
23976 (dse_step5_spill): Remove never called from above.
23978 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
23980 * doc/md.texi (Standard Names) <casesi>: Update documentation for
23981 JUMP_TABLE_DATA changes.
23982 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
23983 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
23984 (Insns) <jump_table_data>: New entry.
23985 * doc/tm.texi: Regenerate.
23987 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
23989 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
23990 for table jump at the end of a basic block using tablejump_p.
23991 * targhooks.c (default_invalid_within_doloop): Likewise.
23992 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
23993 target hook implementation that is identical to the default hook.
23994 (rs6000_invalid_within_doloop): Remove.
23996 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
23997 unused variable from tablejump_p call.
23999 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
24000 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
24001 (INSN_DELETED_P): Likewise.
24002 (emit_jump_table_data): New prototype.
24003 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
24004 after 4th as unused.
24005 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
24006 * sched-vis.c (print_insn): Likewise.
24007 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
24008 insn for compatibility with back ends that use next_active_insn to
24009 identify jump table data.
24010 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
24011 (remove_insn): Likewise.
24012 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
24014 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
24015 (emit_jump_table_data): New function.
24017 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24018 basic block, a JUMP_TABLE_DATA never is.
24019 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24020 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24021 off from code handling real insns.
24022 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24023 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24024 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24025 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24026 is not a NONDEBUG_INSN_P.
24027 * ira-costs.c (scan_one_insn): Likewise.
24028 * jump.c (mark_all_labels): Likewise.
24029 (mark_jump_label_1): Likewise.
24030 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24031 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24033 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24034 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24035 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24036 JUMP_TABLE_DATA_P insns.
24037 (calculate_elim_costs_all_insns): Likewise.
24038 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24039 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24040 (delete_output_reload): Code style fixups.
24041 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24042 insn flags on this non-insn.
24043 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24044 as scheduling barriers, for pre-change compatibility.
24045 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24046 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24048 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24049 redundant JUMP_TABLE_DATA_P test.
24050 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24051 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24052 (frv_for_each_packet): Likewise.
24053 * config/i386/i386.c (min_insn_size): Likewise.
24054 (ix86_avoid_jump_mispredicts): Likewise.
24055 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24056 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24057 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24058 (mips16_insn_length): Robustify.
24059 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24060 (mips16_split_long_branches): Likewise.
24061 * config/pa/pa.c (pa_combine_instructions): Likewise.
24062 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24063 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24064 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24065 as contributing to pool range lengths.
24066 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24067 Remove redundant JUMP_TABLE_DATA_P test.
24068 (sh_loop_align): Likewise.
24069 (split_branches): Likewise.
24070 (sh_insn_length_adjustment): Likewise.
24071 * config/spu/spu.c (get_branch_target): Likewise.
24073 2013-03-29 Jan Hubicka <jh@suse.cz>
24075 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24076 gcov streaming; stream hot bb threshold to ltrans.
24077 * predict.c (get_hot_bb_threshold): Break out from ....
24078 (maybe_hot_count_p): ... here.
24079 (set_hot_bb_threshold): New function.
24080 * lto-section-in.c (lto_section_name): Add profile.
24081 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24082 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24083 and data-streamer.h
24084 (histogram_entry): New structure.
24085 (histogram, histogram_pool): New global vars.
24086 (histogram_hash): New structure.
24087 (histogram_hash::hash): New method.
24088 (histogram_hash::equal): Likewise.
24089 (account_time_size): New function.
24090 (cmp_counts): New function.
24091 (dump_histogram): New function.
24092 (ipa_profile_generate_summary): New function.
24093 (ipa_profile_write_summary): New function.
24094 (ipa_profile_read_summary): New function.
24095 (ipa_profile): Decide on threshold.
24096 (pass_ipa_profile): Add ipa_profile_write_summary and
24097 ipa_profile_read_summary.
24098 * Makefile.in (ipa.o): Update dependencies.
24099 * lto-streamer.h (LTO_section_ipa_profile): New section.
24101 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24103 * tree.h (VAR_P): New.
24105 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24108 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24110 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24112 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24113 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24114 (control_flow_insn_p): Likewise.
24115 * cfgrtl.c (duplicate_insn_chain): Likewise.
24116 * final.c (get_attr_length_1): Likewise.
24117 (shorten_branches): Likewise.
24118 (final_scan_insn): Likewise.
24119 * function.c (instantiate_virtual_regs): Likewise.
24120 * gcse.c (insert_insn_end_basic_block): Likewise.
24121 * ira-costs.c (scan_one_insn): Likewise.
24122 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24123 * lra.c (check_rtl): Likewise.
24124 * reload1.c (elimination_costs_in_insn): Likewise.
24125 * reorg.c (follow_jumps): Likewise.
24127 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24128 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24129 (thumb_far_jump_used_p): Likewise.
24130 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24131 (workaround_speculation): Likewise.
24132 (add_sched_insns_for_speculation): Likewise.
24133 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24134 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24135 (frv_for_each_packet): Likewise.
24136 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24137 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24138 (final_emit_insn_group_barriers): Likewise.
24139 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24140 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24141 (mips16_insn_length): Likewise.
24142 * config/pa/pa.c (pa_reorg): Likewise.
24143 (pa_combine_instructions): Likewise.
24144 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24145 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24146 (sh_reorg): Likewise.
24147 (split_branches): Likewise.
24148 * config/spu/spu.c (get_branch_target): Likewise.
24150 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24153 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24155 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24156 Fix declaration name.
24158 2013-03-28 Lawrence Crowl <crowl@google.com>
24160 * graphds.h (struct graph.indicies): Remove unused.
24161 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24162 (SCOP_ORIGINAL_PDDRS): Remove unused.
24163 * sese.h (extern insert_loop_close_phis): Removed unused.
24164 (extern insert_guard_phis): Removed unused.
24165 (extern ivtype_map_elt_info): Removed unused.
24166 (new_ivtype_map_elt): Removed unused.
24167 * sese.c (ivtype_map_elt_info): Removed unused.
24169 2013-03-28 Lawrence Crowl <crowl@google.com>
24171 * Makefile.in: Add several missing include dependences.
24173 (test-dump.o): New. This object is not added to any executable,
24174 but is present for ad-hoc testing.
24176 (debug (const bitmap_head_def &)): New.
24177 (debug (const bitmap_head_def *)): New.
24179 (extern debug (const bitmap_head_def &)): New.
24180 (extern debug (const bitmap_head_def *)): New.
24182 (debug (edge_def &)): New.
24183 (debug (edge_def *)): New.
24185 (debug (basic_block_def &)): New.
24186 (debug (basic_block_def *)): New.
24188 (dump_node (const_tree, int, FILE *)): Correct source file.
24190 (debug (die_struct &)): New.
24191 (debug (die_struct *)): New.
24193 (extern debug (die_struct &)): New.
24194 (extern debug (die_struct *)): New.
24195 * gimple-pretty-print.c
24196 (debug (gimple_statement_d &)): New.
24197 (debug (gimple_statement_d *)): New.
24198 * gimple-pretty-print.h
24199 (extern debug (gimple_statement_d &)): New.
24200 (extern debug (gimple_statement_d *)): New.
24202 (debug (ira_allocno_copy &)): New.
24203 (debug (ira_allocno_copy *)): New.
24204 (debug (ira_allocno &)): New.
24205 (debug (ira_allocno *)): New.
24207 (extern debug (ira_allocno_copy &)): New.
24208 (extern debug (ira_allocno_copy *)): New.
24209 (extern debug (ira_allocno &)): New.
24210 (extern debug (ira_allocno *)): New.
24212 (debug (live_range &)): New.
24213 (debug (live_range *)): New.
24215 (debug (lra_live_range &)): New.
24216 (debug (lra_live_range *)): New.
24218 (debug (lra_live_range &)): New.
24219 (debug (lra_live_range *)): New.
24221 (debug (omega_pb_d &)): New.
24222 (debug (omega_pb_d *)): New.
24224 (extern debug (omega_pb_d &)): New.
24225 (extern debug (omega_pb_d *)): New.
24227 (debug (const rtx_def &)): New.
24228 (debug (const rtx_def *)): New.
24230 (debug_tree (tree): Move within file.
24231 (debug_raw (const tree_node &)): New.
24232 (debug_raw (const tree_node *)): New.
24233 (dump_tree_via_hooks (const tree_node *, int)): New.
24234 (debug (const tree_node &)): New.
24235 (debug (const tree_node *)): New.
24236 (debug_verbose (const tree_node &)): New.
24237 (debug_verbose (const tree_node *)): New.
24238 (debug_head (const tree_node &)): New.
24239 (debug_head (const tree_node *)): New.
24240 (debug_body (const tree_node &)): New.
24241 (debug_body (const tree_node *)): New.
24242 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24243 (debug (vec<tree, va_gc> &)): New.
24244 (debug (vec<tree, va_gc> *)): New.
24246 (extern debug (const rtx_def &)): New.
24247 (extern debug (const rtx_def *)): New.
24249 (debug_raw (simple_bitmap_def &)): New.
24250 (debug_raw (simple_bitmap_def *)): New.
24251 (debug (simple_bitmap_def &)): New.
24252 (debug (simple_bitmap_def *)): New.
24254 (extern debug (simple_bitmap_def &)): New.
24255 (extern debug (simple_bitmap_def *)): New.
24256 (extern debug_raw (simple_bitmap_def &)): New.
24257 (extern debug_raw (simple_bitmap_def *)): New.
24259 (debug (vinsn_def &)): New.
24260 (debug (vinsn_def *)): New.
24261 (debug_verbose (vinsn_def &)): New.
24262 (debug_verbose (vinsn_def *)): New.
24263 (debug (expr_def &)): New.
24264 (debug (expr_def *)): New.
24265 (debug_verbose (expr_def &)): New.
24266 (debug_verbose (expr_def *)): New.
24267 (debug (vec<rtx> &)): New.
24268 (debug (vec<rtx> *)): New.
24270 (extern debug (vinsn_def &)): New.
24271 (extern debug (vinsn_def *)): New.
24272 (extern debug_verbose (vinsn_def &)): New.
24273 (extern debug_verbose (vinsn_def *)): New.
24274 (extern debug (expr_def &)): New.
24275 (extern debug (expr_def *)): New.
24276 (extern debug_verbose (expr_def &)): New.
24277 (extern debug_verbose (expr_def *)): New.
24278 (extern debug (vec<rtx> &)): New.
24279 (extern debug (vec<rtx> *)): New.
24281 (_list_iter_cond_expr): Make inline instead of static.
24283 (debug (sreal &)): New.
24284 (debug (sreal *)): New.
24286 (extern debug (sreal &)): New.
24287 (extern debug (sreal *)): New.
24289 (extern debug_raw (const tree_node &)): New.
24290 (extern debug_raw (const tree_node *)): New.
24291 (extern debug (const tree_node &)): New.
24292 (extern debug (const tree_node *)): New.
24293 (extern debug_verbose (const tree_node &)): New.
24294 (extern debug_verbose (const tree_node *)): New.
24295 (extern debug_head (const tree_node &)): New.
24296 (extern debug_head (const tree_node *)): New.
24297 (extern debug_body (const tree_node &)): New.
24298 (extern debug_body (const tree_node *)): New.
24299 (extern debug (vec<tree, va_gc> &)): New.
24300 (extern debug (vec<tree, va_gc> *)): New.
24302 (debug (struct loop &)): New.
24303 (debug (struct loop *)): New.
24304 (debug_verbose (struct loop &)): New.
24305 (debug_verbose (struct loop *)): New.
24306 * tree-dump.c: Add header dependence.
24308 (extern debug (struct loop &)): New.
24309 (extern debug (struct loop *)): New.
24310 (extern debug_verbose (struct loop &)): New.
24311 (extern debug_verbose (struct loop *)): New.
24313 (debug (data_reference &)): New.
24314 (debug (data_reference *)): New.
24315 (debug (vec<data_reference_p> &)): New.
24316 (debug (vec<data_reference_p> *)): New.
24317 (debug (vec<ddr_p> &)): New.
24318 (debug (vec<ddr_p> *)): New.
24320 (extern debug (data_reference &)): New.
24321 (extern debug (data_reference *)): New.
24322 (extern debug (vec<data_reference_p> &)): New.
24323 (extern debug (vec<data_reference_p> *)): New.
24324 (extern debug (vec<ddr_p> &)): New.
24325 (extern debug (vec<ddr_p> *)): New.
24327 (debug (pt_solution &)): New.
24328 (debug (pt_solution *)): New.
24330 (extern debug (pt_solution &)): New.
24331 (extern debug (pt_solution *)): New.
24333 (debug (_var_map &)): New.
24334 (debug (_var_map *)): New.
24335 (debug (tree_live_info_d &)): New.
24336 (debug (tree_live_info_d *)): New.
24338 (extern debug (_var_map &)): New.
24339 (extern debug (_var_map *)): New.
24340 (extern debug (tree_live_info_d &)): New.
24341 (extern debug (tree_live_info_d *)): New.
24343 2013-03-28 Jan Hubicka <jh@suse.cz>
24345 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24347 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24349 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24350 record only when desired or required.
24352 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24354 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24355 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24357 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24359 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24360 (*andsi3_compare0_uxtw): New pattern.
24361 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24362 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24364 2013-03-28 Jan Hubicka <jh@suse.cz>
24366 * data-streamer-in.c (streamer_read_gcov_count): New function.
24367 * gimple-streamer-out.c: Include value-prof.h.
24368 (output_gimple_stmt): Output histogram.
24369 (output_bb): Use streamer_write_gcov_count.
24370 * value-prof.c: Include data-streamer.h
24371 (dump_histogram_value): Add HIST_TYPE_MAX.
24372 (stream_out_histogram_value): New function.
24373 (stream_in_histogram_value): New function.
24374 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24375 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24376 * data-streamer-out.c (streamer_write_gcov_count): New function.
24377 (streamer_write_gcov_count_stream): New function.
24378 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24379 (lto_output_node): Likewise.
24380 (input_node, input_edge): Likewise.
24381 * lto-streamer-out.c (output_cfg): Update streaming.
24382 * lto-streamer-in.c (input_cfg): Likewise.
24383 * data-streamer.h (streamer_write_gcov_count,
24384 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24385 * gimple-streamer-in.c: Include value-prof.h
24386 (input_gimple_stmt): Input histograms.
24387 (input_bb): Update profile streaming.
24389 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24391 * genmodes.c (emit_max_int): New function.
24392 (emit_insn_modes_h): Added call to emit_max_function.
24393 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24395 * machmode.def: Fixed comment.
24397 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24399 * combine.c (try_combine): Removed useless assert.
24400 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24402 2013-03-28 Marek Polacek <polacek@redhat.com>
24403 Richard Biener <rguenther@suse.de>
24405 PR tree-optimization/56695
24406 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24407 build signed result of a vector comparison.
24408 * tree-cfg.c (verify_gimple_comparison): Check that a result
24409 of a vector comparison has signed type.
24411 2013-03-28 Richard Biener <rguenther@suse.de>
24413 PR tree-optimization/37021
24414 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24415 do not restrict gaps between groups.
24416 * tree-vect-stmts.c (vectorizable_load): Properly account for
24417 a gap between groups.
24419 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24421 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24422 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24425 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24427 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24428 * doc/extend.texi (Named Address Spaces): Ditto.
24429 (Variable Attributes): Ditto.
24431 2013-03-27 Kai Tietz <ktietz@redhat.com>
24433 * config.build: Add support for cygwin x64 target.
24434 * config.gcc: Likewise.
24435 * config.host: Likewise.
24436 * configure.ac: Likewise
24437 * configure: Regenerated.
24439 2013-03-27 Kai Tietz <ktietz@redhat.com>
24441 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24442 * config/i386/t-cygwin-w64: New file.
24443 * config/i386/cygwin-w64.h: New file.
24444 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24445 and add support for x64-cygwin target.
24446 (CPP_SPEC): Likewise.
24447 (CXX_WRAP_SPEC_LIST): Undefine before define.
24448 (LIBGCJ_SONAME): Use 15 as version.
24450 2013-03-27 Richard Biener <rguenther@suse.de>
24452 PR tree-optimization/56716
24453 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24454 dumping for ref nodes.
24456 2013-03-27 Martin Jambor <mjambor@suse.cz>
24458 PR tree-optimization/55334
24459 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24460 restricted pointers to arrays.
24462 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24464 * Makefile.in (.SUFFIXES): Add .cc.
24465 (.c.o): Apply same recipe for implicit rule .cc.o.
24467 2013-03-27 Richard Biener <rguenther@suse.de>
24469 PR tree-optimization/37021
24470 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24471 REALPART/IMAGPART_EXPRs around the supported refs.
24472 * tree-ssa-structalias.c (find_func_aliases): Assume that
24473 floating-point values are not used to transfer pointers.
24475 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24477 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24478 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24479 * doc/tm.texi: Regenerate.
24480 * targhooks.h (default_has_ifunc_p): New.
24481 * targhooks.c (default_has_ifunc_p): Ditto.
24482 * config/linux-protos.h: New file.
24483 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24484 hook for linux which disables support of indirect functions in android.
24485 * config/linux-android.c: New file.
24486 * config/t-linux-android.c: Ditto.
24487 * config.gcc: Added new object file linux-android.o.
24488 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24489 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24490 * varasm.c (do_assemble_alias): Likewise.
24491 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24492 doesn't support indirect functions.
24493 * configure: Regenerate.
24495 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24498 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24499 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24502 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24504 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24506 2013-03-27 Terry Guo <terry.guo@arm.com>
24508 * config/arm/arm-cores.def: Added core cortex-r7.
24509 * config/arm/arm-tune.md: Regenerated.
24510 * config/arm/arm-tables.opt: Regenerated.
24511 * doc/invoke.texi: Added entry for core cortex-r7.
24513 2013-03-27 Walter Lee <walt@tilera.com>
24515 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24516 double-decrement of next_scratch_regno.
24518 2013-03-27 Walter Lee <walt@tilera.com>
24520 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24522 (insn_v1mulus): Ditto.
24523 (insn_v2muls): Ditto.
24525 2013-03-27 Walter Lee <walt@tilera.com>
24527 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24528 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24530 2013-03-27 Walter Lee <walt@tilera.com>
24532 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24533 (*sibcall_value): Ditto.
24535 2013-03-27 Walter Lee <walt@tilera.com>
24537 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24538 (insn_mnz_v8qi): ... this ...
24539 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24541 (insn_v<n>mnz): Replaced by ...
24542 (insn_v1mnz): ... this ...
24543 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24545 (insn_mz_<mode>): Replaced by ...
24546 (insn_mz_v8qi): ... this ...
24547 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24549 (insn_v<n>mz): Replaced by ...
24550 (insn_v1mz): ... this ...
24551 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24554 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24556 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24558 2013-03-26 Roland McGrath <mcgrathr@google.com>
24560 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24561 than fprintf with a non-constant, non-format string.
24563 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24565 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24566 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24567 operand 0 predicate.
24568 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24569 attribute. Use general_x64nomem_operand as operand 1 predicate.
24570 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24571 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24572 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24573 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24574 (mov<mode>_insv_1): Remove expander. Merge insn with
24575 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24576 Use general_x64nomem_operand as operand 1 predicate.
24577 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24578 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24579 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24580 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24581 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24582 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24583 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24584 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24585 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24586 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24587 (general_x64nomem_operand): Ditto.
24589 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24591 * config/rtems.opt: Add -pthread option.
24593 2013-03-26 Richard Biener <rguenther@suse.de>
24595 * alias.c (find_base_term): Avoid redundant and not used recursion.
24596 (base_alias_check): Get the initial base term from the caller.
24597 (true_dependence_1): Compute and pass base terms to base_alias_check.
24598 (write_dependence_p): Likewise.
24599 (may_alias_p): Likewise.
24601 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24603 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24604 PC-relative load in SI modes and above only.
24606 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24608 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24609 * config/arm/iwmmxt.md (WCGR0): Update.
24610 (WCGR1, WCGR2, WCGR3): Likewise.
24612 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24614 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24615 Use x64 and nox64 isa attributes.
24617 2013-03-26 Richard Biener <rguenther@suse.de>
24619 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24620 alignment computations and rely on get_object_alignment_1
24621 for the !TYPE_P case.
24622 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24624 2013-03-26 Walter Lee <walt@tilera.com>
24626 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24627 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24629 2013-03-25 Jeff Law <law@redhat.com>
24631 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
24632 check for INTEGRAL_TYPE_P that was missing due to checking in
24633 wrong version of prior patch.
24635 2013-03-25 Walter Lee <walt@tilera.com>
24637 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
24638 TILEGX_INSN_SHUFFLEBYTES1.
24639 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
24641 (tilegx_builtins): Ditto.
24642 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
24644 2013-03-25 Walter Lee <walt@tilera.com>
24646 * config/tilegx/tilegx.md (floatsisf2): New pattern.
24647 (floatunssisf2): New pattern.
24648 (floatsidf2): New pattern.
24649 (floatunssidf2): New pattern.
24651 2013-03-25 Walter Lee <walt@tilera.com>
24653 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
24654 tests for constraint J, K, N, P.
24656 2013-03-25 Walter Lee <walt@tilera.com>
24658 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
24659 Use indirect/pcrel encoding.
24660 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
24663 2013-03-25 Steve Ellcey <sellcey@mips.com>
24665 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
24666 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
24667 * config/mips/mips.c (mips_option_override): Set IMADD default.
24668 * config/mips/mips.h (PTF_AVOID_IMADD): New.
24669 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
24670 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
24671 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
24672 * doc/invoke.texi (-mimadd/-mno-imadd): New.
24674 2013-03-25 Jeff Law <law@redhat.com>
24676 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
24677 slightly to avoid creating and folding useless trees. Simplify
24678 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
24680 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
24682 * config/i386/i386.md (*zero_extendsidi2): Merge with
24683 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
24684 * config/i386/predicates.md (x86_64_zext_operand): Rename from
24685 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
24686 targets. Clarify comment.
24688 2013-03-25 Martin Jambor <mjambor@suse.cz>
24690 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
24691 pass-through jump functions differently.
24692 (ipa_read_jump_function): Likewise. Also use setter functions to set
24695 2013-03-25 Martin Jambor <mjambor@suse.cz>
24697 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
24698 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
24700 (ipa_get_indirect_edge_target): New function.
24701 (devirtualization_time_bonus): New parameter known_aggs, pass it to
24702 ipa_get_indirect_edge_target. Update all callers.
24703 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
24704 ipa_get_indirect_edge_target_1 instead of calling
24705 ipa_get_indirect_edge_target.
24706 (create_specialized_node): Pass aggvlas to
24707 ipcp_discover_new_direct_edges.
24709 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24711 * config/arm/arm.md (f_sels, f_seld): New types.
24712 (*cmov<mode>): New pattern.
24713 * config/arm/predicates.md (arm_vsel_comparison_operator): New
24716 2013-03-25 Kai Tietz <ktietz@redhat.com>
24718 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
24719 POSIX-printf for mingw-hosted builds.
24721 2013-03-25 Richard Biener <rguenther@suse.de>
24723 PR middle-end/56694
24724 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
24725 must-not-throw stmt location.
24727 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24729 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
24730 Emit load-acquire versions when acq is true.
24731 (arm_emit_store_exclusive): Add rel parameter.
24732 Emit store-release versions when rel is true.
24733 (arm_split_compare_and_swap): Use acquire-release instructions
24735 of barriers when appropriate.
24736 (arm_split_atomic_op): Likewise.
24737 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
24738 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
24739 (VUNSPEC_SLX): Likewise.
24740 (VUNSPEC_LDA): Likewise.
24741 (VUNSPEC_STL): Likewise.
24742 * config/arm/sync.md (atomic_load<mode>): New pattern.
24743 (atomic_store<mode>): Likewise.
24744 (arm_load_acquire_exclusive<mode>): Likewise.
24745 (arm_load_acquire_exclusivesi): Likewise.
24746 (arm_load_acquire_exclusivedi): Likewise.
24747 (arm_store_release_exclusive<mode>): Likewise.
24749 2013-03-25 Catherine Moore <clm@codesourcery.com>
24751 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
24752 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
24753 * config/mip/predicates.md (lwsp_swsp_operand,
24754 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
24755 sb16_operand, db4_operand, db7_operand, ib3_operand,
24756 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
24757 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
24758 andi16_operand): New predicates.
24759 * config/mips/mips.md (compression): New attribute.
24760 (enabled): New attribute.
24761 (length): Consider compression in computing length.
24762 (shift_compression): New code attribute.
24763 (*add<mode>3): New operands. Record compression.
24764 (sub<mode>3): Likewise.
24765 (one_cmpl<mode>2): Likewise.
24766 (*and<mode>3): Likewise.
24767 (*ior<mode>3): Likewise.
24768 (unnamed pattern for xor): Likewise.
24769 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
24770 (*<optab><mode>3): Likewise.
24771 (*mov<mode>_internal: Likewise.
24772 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
24773 (mips_unsigned_immediate_p): New.
24774 (umips_lwsp_swsp_address_p): New.
24775 (m16_based_address_p): New.
24776 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
24777 (mips_unsigned_immediate_p): New prototype.
24778 (lwsp_swsp_address_p): New prototype.
24779 (m16_based_address_p): New prototype.
24780 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
24781 (mips_signed_immediate_p): New function.
24782 (m16_based_address_p): New function.
24783 (lwsp_swsp_address_p): New function.
24784 (mips_print_operand_punctuation): Recognize short delay slot insns
24785 for microMIPS.add<mode>3"
24787 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24790 * config/arm/iterators.md (v_cmp_result): New mode attribute.
24791 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
24793 2013-03-25 Richard Biener <rguenther@suse.de>
24795 PR tree-optimization/56689
24796 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
24799 2013-03-25 Richard Biener <rguenther@suse.de>
24801 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
24803 (memory_references): Likewise.
24804 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
24805 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
24806 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
24807 (gather_mem_refs_in_loops): Fold into ...
24808 (analyze_memory_references): ... this. Move initialization
24809 to tree_ssa_lim_initialize.
24810 (fill_always_executed_in): Rename to ...
24811 (fill_always_executed_in_1): ... this.
24812 (fill_always_executed_in): Move contains_call computation to
24813 this new function from ...
24814 (tree_ssa_lim_initialize): ... here.
24815 (tree_ssa_lim): Call fill_always_executed_in.
24817 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24819 * postreload.c (reload_combine): Fix code detecting returns.
24821 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24823 * function.c (emit_use_return_register_into_block): On cc0 targets,
24824 do not emit the sequence between cc0 setter and user.
24826 2013-03-25 Kai Tietz <ktietz@redhat.com>
24828 * config/i386/predicates.md (local_symbolic_operand): Interpret
24829 dll-imported symbols as none-local.
24831 2013-03-25 Richard Biener <rguenther@suse.de>
24833 * tree-ssa-loop-im.c (struct depend): Remove.
24834 (struct lim_aux_data): Make depends a vec of gimples.
24835 (free_lim_aux_data): Adjust.
24836 (add_dependency): Likewise.
24837 (set_level): Likewise.
24839 2013-03-25 Richard Biener <rguenther@suse.de>
24841 PR middle-end/56434
24842 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24843 the pointer returned by calls with ECF_MALLOC set.
24845 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24847 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24849 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24851 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24852 using MMXMODE mode iterator.
24853 (*move<mode>_internal): Merge with *movv2sf_internal and
24854 *movv2sf_internal_rex64 using MMXMODE mode iterator.
24856 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24858 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
24859 (record_last_mem_set_info): Likewise.
24861 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
24862 of XNEWVEC followed by memset.
24863 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
24865 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24867 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
24868 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
24869 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
24870 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
24871 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
24872 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
24873 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
24874 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
24875 BARRIER_P instead of GET_CODE.
24877 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24879 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
24880 inaccuracy in the probing code.
24882 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
24883 (ctrapdi4): Likewise.
24885 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24887 * calls.c (expand_call): Add missing guard to code handling return
24888 of non-BLKmode structures in MSB.
24889 * function.c (expand_function_end): Likewise.
24891 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24893 * combine.c (try_combine): Adjust comment. Do not add the set of
24894 insn #0 if the destination indirectly is set or dies in insn #2.
24895 Tidy up code to distribute a new note.
24897 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24899 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
24900 also for alternatives 16 and 17.
24902 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24904 * config/i386/sse.md (*mov<mode>_internal): Merge with
24905 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
24906 Emit insn template depending on type attribute. Use
24907 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24908 movd instead of movq mnemonic for interunit moves. Rewrite mode
24909 attribute calculation. Remove unit attribute calculation.
24910 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
24911 Set prefix_data16 attribute for DImode ssemov types.
24912 Use Ym instead of y for SSE-MMX conversion alternatives.
24913 Reorder operand constraints.
24915 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
24917 * df.h (df_insn_delete): Adjust prototype.
24918 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
24919 and let it decide whether mark the basic block dirty.
24920 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
24921 * df-scan.c (df_insn_info_delete): New helper function, split
24922 off from df_insn_delete.
24923 (df_scan_free_bb_info): Use it.
24924 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
24926 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
24927 that the insn is actually an insn and it has a non-NULL basic block.
24928 Do not mark basic block dirty if only deleting a DEBUG_INSN.
24930 2013-03-22 Richard Biener <rguenther@suse.de>
24932 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
24934 (mem_ref_alloc): Do not allocate them.
24935 (refs_independent_p): Do not query or maintain a cache.
24937 2013-03-22 Richard Biener <rguenther@suse.de>
24939 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
24940 (gather_mem_refs_in_loops): Do not compute it.
24941 (analyze_memory_references): Do not allocate it.
24942 (tree_ssa_lim_finalize): Do not free it.
24943 (for_all_locs_in_loop): Do not query all_refs_in_loop.
24945 2013-03-22 Richard Biener <rguenther@suse.de>
24947 * is-a.h (as_a): Use gcc_checking_assert.
24949 2013-03-22 Ian Bolton <ian.bolton@arm.com>
24951 * config/aarch64/aarch64.c (aarch64_print_operand): New
24952 format specifier for printing a constant in hex.
24953 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
24954 format specifier for printing second operand.
24956 2013-03-22 Richard Biener <rguenther@suse.de>
24958 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
24960 (gather_mem_refs_in_loops): Perform store accumulation here.
24961 (create_vop_ref_mapping_loop): Remove.
24962 (create_vop_ref_mapping): Likewise.
24963 (analyze_memory_references): Initialize refs_stored_in_loop.
24964 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
24965 (record_indep_loop): Remove.
24966 (record_dep_loop): New function.
24967 (ref_indep_loop_p_1): Adjust to only walk over references
24968 in the loop, not its subloops.
24969 (ref_indep_loop_p): Rename to ...
24970 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
24971 maintaining a more fine-grained cache.
24972 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
24973 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
24975 2013-03-22 Richard Biener <rguenther@suse.de>
24977 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
24978 (struct mem_ref): Make accesses_in_loop a vec of a vec of
24979 aggregate mem_ref_loc.
24980 (free_mem_ref_locs): Inline into ...
24981 (memref_free): ... this and adjust.
24982 (mem_ref_alloc): Adjust.
24983 (mem_ref_locs_alloc): Remove.
24984 (record_mem_ref_loc): Adjust.
24985 (get_all_locs_in_loop): Rewrite into ...
24986 (for_all_locs_in_loop): ... this iterator.
24987 (rewrite_mem_ref_loc): New functor.
24988 (rewrite_mem_refs): Use for_all_locs_in_loop.
24989 (sm_set_flag_if_changed): New functor.
24990 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
24991 (ref_always_accessed): New functor.
24992 (ref_always_accessed_p): Use for_all_locs_in_loop.
24994 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24996 * tree-pass.h (PROP_gimple_lvec): New.
24997 * passes.c (dump_properties): Handle PROP_gimple_lvec.
24998 (init_optimization_passes): Move pass_lower_vector.
24999 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
25001 (pass_lower_vector): Provide PROP_gimple_lvec.
25002 (pass_lower_vector_ssa): Likewise.
25003 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
25005 2013-03-21 Mark Wielaard <mjw@redhat.com>
25007 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
25009 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25011 * config/i386/i386.md (*movdi_internal): Disparage slightly
25012 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
25013 conversion alternatives.
25015 2013-03-21 Jakub Jelinek <jakub@redhat.com>
25017 PR middle-end/48087
25018 * diagnostic.def (DK_WERROR): New kind.
25019 * diagnostic.h (werrorcount): Define.
25020 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25021 promoted to DK_ERROR, increment DK_WERROR counter instead of
25023 * toplev.c (toplev_main): Call print_ignored_options even if
25024 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25025 even if just werrorcount is non-zero.
25028 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25030 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25031 (string_cst_pool_decl): New function.
25032 (optimize_one_addr_into_implicit_ptr): New function.
25033 (resolve_addr_in_expr): Optimize DWARF location expression
25034 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25035 which doesn't live in memory, but has DW_AT_location or
25036 DW_AT_const_value, or refers to a string literal, into
25037 DW_OP_GNU_implicit_pointer.
25038 (optimize_location_into_implicit_ptr): New function.
25039 (resolve_addr): If removing DW_AT_location of a variable because
25040 it was DW_OP_addr of address of the variable, but the variable doesn't
25041 live in memory, try to emit const value attribute for the initializer.
25043 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25045 * tree.h (VECTOR_TYPE_P): New macro.
25046 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25047 TYPE_MODE): Use it.
25048 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25049 VEC_COND_EXPR cannot be lvalues.
25050 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25052 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25054 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25055 Restrict the transformation to equal modes.
25057 2013-03-21 Richard Biener <rguenther@suse.de>
25059 PR tree-optimization/39326
25060 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25061 (MEM_ANALYZABLE): Adjust.
25062 (record_mem_ref_loc): Move bitmap ops ...
25063 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25064 unanalyzable refs, do not record locations for it.
25065 (analyze_memory_references): Allocate ref zero as shared
25067 (refs_independent_p): Do not test for unanalyzed mems here.
25068 (ref_indep_loop_p_1): Special-case disambiguation against
25069 the unanalyzed ref.
25070 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25072 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25074 * config/arm/arm-protos.h (tune_params): Add
25075 prefer_neon_for_64bits field.
25076 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25077 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25078 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25079 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25080 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25081 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25082 (arm_option_override): Handle -mneon-for-64bits new option.
25083 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25084 (prefer_neon_for_64bits): Declare new variable.
25085 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25086 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25087 (arch_enabled): Handle new arch types. Remove support for onlya8
25089 (one_cmpldi2): Use new arch names.
25090 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25091 * config/arm/arm.opt (mneon-for-64bits): Add option.
25092 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25093 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25094 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25096 * doc/invoke.texi (-mneon-for-64bits): Document.
25098 2013-03-21 Richard Biener <rguenther@suse.de>
25100 PR tree-optimization/39326
25101 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25102 (sort_bbs_in_loop_postorder_cmp): New function.
25103 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25105 2013-03-21 Richard Biener <rguenther@suse.de>
25107 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25108 (vect_insert_into_interleaving_chain): Likewise.
25109 (vect_drs_dependent_in_basic_block): Inline ...
25110 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25112 (vect_analyze_data_ref_dependence): ... here. Simplify.
25113 (vect_check_interleaving): Simplify.
25114 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25115 (vect_slp_analyze_data_ref_dependences): ... this new function.
25116 (dr_group_sort_cmp): New function.
25117 (vect_analyze_data_ref_accesses): Compute data-reference groups
25118 here instead of in vect_analyze_data_ref_dependence. Use
25119 a more efficient algorithm.
25120 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25121 vect_slp_analyze_data_ref_dependences. Call
25122 vect_analyze_data_ref_accesses earlier.
25123 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25124 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25125 (vect_slp_analyze_data_ref_dependences): New prototype.
25127 2013-03-21 Richard Biener <rguenther@suse.de>
25129 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25130 ref is stored in the loop.
25131 (find_refs_for_sm): Walk only over all stores.
25132 (store_motion_loop): Allocate from lim_bitmap_obstack.
25133 (store_motion): Likewise.
25135 2013-03-21 Richard Biener <rguenther@suse.de>
25137 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25138 Update virtual SSA form.
25140 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25142 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25143 * configure: Regenerate.
25144 * config.in: Regenerate.
25145 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25146 if !HAVE_LD_EH_FRAME_CIEV3.
25148 2013-03-21 Richard Biener <rguenther@suse.de>
25150 * tree-cfg.c (verify_expr_no_block): New function.
25151 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25152 nor DECL_VALUE_EXPR have locations with associated blocks.
25153 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25154 (clear_unused_block_pointer): Remove code dealing with
25155 blocks in DECL_DEBUG_EXPR locations.
25157 2013-03-21 Richard Biener <rguenther@suse.de>
25159 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25160 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25161 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25162 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
25163 instead of DECL_DEBUG_EXPR_IS_FROM.
25164 * gimplify.c (gimplify_modify_expr): Likewise.
25165 * tree-cfg.c (verify_expr_location_1): Likewise.
25166 * tree-complex.c (create_one_component_var): Likewise.
25167 * tree-sra.c (create_access_replacement): Likewise.
25168 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
25169 (clear_unused_block_pointer): Likewise.
25170 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
25171 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
25172 * var-tracking.c (var_debug_decl): Likewise.
25173 (track_expr_p): Likewise.
25174 * tree-inline.c (add_local_variables): Likewise. Set
25175 DECL_HAS_DEBUG_EXPR_P after copying it.
25176 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
25177 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
25179 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25182 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
25183 * configure: Regenerate.
25184 * config.in: Regenerate.
25185 * config/i386/i386.md (*movdf_internal): Use
25186 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25187 movd instead of movq mnemonic for interunit moves.
25188 (*movdi_internal): Ditto.
25190 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25192 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
25193 (abd<mode>_3): New pattern.
25194 (aba<mode>_3): New pattern.
25195 (fabd<mode>_3): New pattern.
25197 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25199 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
25200 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
25201 occurrence of REGISTER_PREFIX as its empty string.
25203 2013-03-20 Jeff Law <law@redhat.com>
25205 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
25206 addititional equivalences for equality comparisons between an SSA_NAME
25207 and a constant where the SSA_NAME was set from a widening conversion.
25209 2013-03-20 Walter Lee <walt@tilera.com>
25211 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25213 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25215 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25216 depending on type attribute.
25217 (*movti_internal): Ditto.
25218 (*movtf_internal): Ditto.
25219 (*movxf_internal): Ditto.
25220 (*movdf_internal): Ditto.
25221 (*movsf_internal): Ditto.
25223 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25225 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25226 maybe_vex for sselog1 and ssemov types.
25227 (*movdi_internal): Reorder operand constraints.
25228 (*movsi_internal): Ditto. Set prefix attribute to
25229 maybe_vex for sselog1 and ssemov types.
25230 (*movtf_internal): Set prefix attribute to maybe_vex
25231 for sselog1 and ssemov types.
25232 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25233 DImode ssemov types. Reorder operand constraints.
25234 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25235 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25236 attribute for SImode ssemov types. Reorder operand constraints.
25238 2013-03-20 Martin Jambor <mjambor@suse.cz>
25240 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25241 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25243 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25245 * config/rs6000/predicates.md (indexed_address, update_address_mem
25246 update_indexed_address_mem): New predicates.
25247 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25248 attribute for load/store instructions.
25249 * config/rs6000/dfp.md (movsd_store): Likewise.
25250 (movsd_load): Likewise.
25251 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25252 (unnamed HI->DI extend define_insn): Likewise.
25253 (unnamed SI->DI extend define_insn): Likewise.
25254 (unnamed QI->SI extend define_insn): Likewise.
25255 (unnamed QI->HI extend define_insn): Likewise.
25256 (unnamed HI->SI extend define_insn): Likewise.
25257 (unnamed HI->SI extend define_insn): Likewise.
25258 (extendsfdf2_fpr): Likewise.
25259 (movsi_internal1): Likewise.
25260 (movsi_internal1_single): Likewise.
25261 (movhi_internal): Likewise.
25262 (movqi_internal): Likewise.
25263 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25264 attribute for load/store instructions.
25265 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25267 (mov<mode>_softfloat): Likewise.
25268 (mov<mode>_hardfloat32): Likewise.
25269 (mov<mode>_hardfloat64): Likewise.
25270 (mov<mode>_softfloat64): Likewise.
25271 (movdi_internal32): Likewise.
25272 (movdi_internal64): Likewise.
25273 (probe_stack_<mode>): Likewise.
25275 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25277 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25278 floating point, and decimal floating point to reload iterator.
25280 * config/rs6000/constraints.md (wl constraint): New constraints to
25281 return FLOAT_REGS if certain options are used to reduce the number
25282 of separate patterns that exist in the file.
25283 (wx constraint): Likewise.
25284 (wz constraint): Likewise.
25286 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25287 -mdebug=reg, print wg, wl, wx, and wz constraints.
25288 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25289 Initialize the reload functions for 64-bit binary/decimal floating
25291 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25292 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25293 create the buffer on the stack to overcome not having a 32-bit
25295 (rs6000_emit_move): Likewise.
25296 (rs6000_secondary_memory_needed_rtx): Likewise.
25297 (rs6000_alloc_sdmode_stack_slot): Likewise.
25298 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25299 via xxlxor, just like DFmode 0.0.
25301 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25302 define as 1 if we are running on a power7 or newer.
25303 (enum r6000_reg_class_enum): Add new constraints.
25305 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25306 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25307 with other moves by using conditional constraits (wg). Use LFIWZX
25308 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25309 (movsd splitter): Likewise.
25310 (movsd_hardfloat): Likewise.
25311 (movsd_softfloat): Likewise.
25313 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25314 binary and decimal floating point moves.
25315 (fmove_ok): New attributes to combine binary and decimal floating
25316 point moves, and to combine power6x (mfpgpr) moves along normal
25318 (real_value_to_target): Likewise.
25319 (f32_lr): Likewise.
25320 (f32_lm): Likewise.
25321 (f32_li): Likewise.
25322 (f32_sr): Likewise.
25323 (f32_sm): Likewise.
25324 (f32_si): Likewise.
25325 (movsf): Combine binary and decimal floating point moves. Combine
25326 power6x (mfpgpr) moves with other moves by using conditional
25327 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25328 (mov<mode> for SFmode/SDmode); Likewise.
25329 (SFmode/SDmode splitters): Likewise.
25330 (movsf_hardfloat): Likewise.
25331 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25332 (movsf_softfloat): Likewise.
25333 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25335 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25336 wx and wz constraints.
25338 * config/rs6000/constraints.md (wg constraint): New constraint to
25339 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25341 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25344 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25345 -mdebug=reg, print wg, wl, wx, and wz constraints.
25346 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25347 Initialize the reload functions for 64-bit binary/decimal floating
25349 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25350 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25351 create the buffer on the stack to overcome not having a 32-bit
25353 (rs6000_emit_move): Likewise.
25354 (rs6000_secondary_memory_needed_rtx): Likewise.
25355 (rs6000_alloc_sdmode_stack_slot): Likewise.
25356 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25357 via xxlxor, just like DFmode 0.0.
25359 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25360 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25361 with other moves by using conditional constraits (wg). Use LFIWZX
25362 and STFIWX for loading SDmode on power7.
25363 (movdd splitters): Likewise.
25364 (movdd_hardfloat32): Likewise.
25365 (movdd_softfloat32): Likewise.
25366 (movdd_hardfloat64_mfpgpr): Likewise.
25367 (movdd_hardfloat64): Likewise.
25368 (movdd_softfloat64): Likewise.
25370 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25371 64-bit binary and decimal floating point moves.
25372 (FMOVE64X): Likewise.
25373 (movdf): Combine 64-bit binary and decimal floating point moves.
25374 Combine power6x (mfpgpr) moves with other moves by using
25375 conditional constraits (wg).
25376 (mov<mode> for DFmode/DDmode): Likewise.
25377 (DFmode/DDmode splitters): Likewise.
25378 (movdf_hardfloat32): Likewise.
25379 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25380 (movdf_softfloat32): Likewise.
25381 (movdf_hardfloat64_mfpgpr): Likewise.
25382 (movdf_hardfloat64): Likewise.
25383 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25384 (movdf_softfloat64): Likewise.
25385 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25386 (reload_<mode>_load): Move to later in the file so they aren't in
25387 the middle of the floating point move insns.
25388 (reload_<mode>_store): Likewise.
25390 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25393 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25394 constraint if -mdebug=reg.
25395 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25396 Enable using dd reload support if needed.
25398 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25399 binary and decimal floating point moves in rs6000.md.
25400 (movtd_internal): Likewise.
25402 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25403 decimal floating point moves.
25405 (movtf_internal): Likewise.
25406 (mov<mode>_internal, TDmode/TFmode): Likewise.
25407 (movtf_softfloat): Likewise.
25408 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25410 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25411 movdi_internal64, using wg constraint for move direct operations.
25412 (movdi_internal64): Likewise.
25414 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25415 MODES_TIEABLE_P for selected modes. Print the numerical value of
25416 the various virtual registers. Use GPR/FPR first/last values,
25417 instead of hard coding the register numbers. Print which modes
25418 have reload functions registered.
25419 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25420 settings before/after setting cpu, target and subtarget settings.
25421 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25422 and for secondary reload failures in rs6000_secondary_reload_inner.
25423 (rs6000_secondary_reload_fail): Likewise.
25424 (rs6000_secondary_reload_inner): Likewise.
25426 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25427 macros for first/last GPR and FPR registers.
25428 (LAST_GPR_REGNO): Likewise.
25429 (FIRST_FPR_REGNO): Likewise.
25430 (LAST_FPR_REGNO): Likewise.
25432 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25433 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25434 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25435 (vcond<mode><mode>): Likewise.
25436 (vcondu<mode><mode>): Likewise.
25437 (vector_gtu<mode>): Likewise.
25438 (vector_gte<mode>): Likewise.
25439 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25440 to prevent the compiler from converting DImode operations to TImode.
25441 (ior<mode>3): Likewise.
25442 (and<mode>3): Likewise.
25443 (one_cmpl<mode>2): Likewise.
25444 (nor<mode>3): Likewise.
25445 (andc<mode>3): Likewise.
25447 * config/rs6000/constraints.md (wt constraint): New constraint
25448 that returns VSX_REGS if TImode is allowed in VSX registers.
25450 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25451 constant under VSX.
25453 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25454 similar to TImode, but it is restricted to being in the GPRs.
25456 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25457 TImode to occupy a single VSX register.
25459 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25460 -mvsx-timode for power7/power8.
25461 (power7 cpu): Likewise.
25462 (power8 cpu): Likewise.
25464 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25465 sure that TFmode/TDmode take up two registers if they are ever
25466 allowed in the upper VSX registers.
25467 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25469 (rs6000_init_hard_regno_mode_ok): Likewise.
25470 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25471 constraint. Print if LRA is turned on.
25472 (rs6000_option_override_internal): Give an error if -mvsx-timode
25473 and VSX is not enabled.
25474 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25475 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25476 to reg+offset addressing. Use PTImode when checking offset
25477 addresses for validity.
25478 (reg_offset_addressing_ok_p): Likewise.
25479 (rs6000_legitimate_offset_address_p): Likewise.
25480 (rs6000_legitimize_address): Likewise.
25481 (rs6000_legitimize_reload_address): Likewise.
25482 (rs6000_legitimate_address_p): Likewise.
25483 (rs6000_eliminate_indexed_memrefs): Likewise.
25484 (rs6000_emit_move): Likewise.
25485 (rs6000_secondary_reload): Likewise.
25486 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25487 reloads to fpr registers to continue to use reg+offset addressing,
25488 but 64-bit reloads to altivec registers need reg+reg addressing.
25489 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25490 it. Treat LO_SUM like a PLUS operation.
25491 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25492 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25493 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25494 registers to share a register with a smaller sized type, since VSX
25495 puts scalars in the upper 64-bits.
25496 (print_operand): Add support for PTImode.
25497 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25498 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25499 registers, but don't have arithmetic support.
25500 (rs6000_memory_move_cost): Add test for VSX.
25501 (rs6000_opt_masks): Add -mvsx-timode.
25503 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25506 (VSr): Use wt constraint for TImode.
25507 (VSv): Drop TImode support.
25508 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25509 (vsx_movti_64bit): Likewise.
25510 (vsx_movti_32bit): Likewise.
25511 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25512 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25513 one '?' on the appropriate output constraint. Do not allow TImode
25514 logical operations on 32-bit systems.
25515 (vsx_ior<mode>3): Likewise.
25516 (vsx_xor<mode>3): Likewise.
25517 (vsx_one_cmpl<mode>2): Likewise.
25518 (vsx_nor<mode>3): Likewise.
25519 (vsx_andc<mode>3): Likewise.
25520 (vsx_concat_<mode>): Likewise.
25521 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25523 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25524 OPTION_MASK_VSX_TIMODE.
25525 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25526 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25528 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25529 (TI2 iterator): New iterator for TImode, PTImode.
25530 (wd mode attribute): Add values for vector types.
25531 (movti_string): Replace TI move operations with operations for TImode
25532 and PTImode. Add support for TImode being allowed in VSX registers.
25533 (mov<mode>_string, TImode/PTImode): Likewise.
25534 (movti_ppc64): Likewise.
25535 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25536 (TI mode splitters): Likewise.
25538 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25541 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25543 PR tree-optimization/56355
25544 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25545 Also handle integers with undefined overflow.
25547 2013-03-20 Catherine Moore <clm@codesourcery.com>
25548 Maciej W. Rozycki <macro@codesourcery.com>
25549 Tom de Vries <tom@codesourcery.com>
25550 Nathan Sidwell <nathan@codesourcery.com>
25551 Iain Sandoe <iain@codesourcery.com>
25552 Nathan Froyd <froydnj@codesourcery.com>
25553 Chao-ying Fu <fu@mips.com>
25555 * doc/extend.texi: (micromips, nomicromips, nocompression):
25556 Document new function attributes.
25557 * doc/invoke.texi (minterlink-compressed, mmicromips,
25558 m14k, m14ke, m14kec): Document new options.
25559 (minterlink-mips16): Update documentation.
25560 * doc/md.texi (ZC, ZD): Document new constraints.
25561 * configure.ac (gcc_cv_as_micromips): Check if linker
25562 supports the .set micromips directive.
25563 * configure: Regenerate.
25564 * config.in: Regenerate.
25565 * config/mips/mips-tables.opt: Regenerate.
25566 * config/mips/micromips.md: New file.
25567 * constraints.md (ZC, ZD): New constraints.
25568 * config/mips/predicates.md (movep_src_register): New predicate.
25569 (movep_src_operand): New predicate.
25570 (non_volatile_mem_operand): New predicate.
25571 * config/mips/mips.md (multimem): New type.
25572 (length): Differentiate between 17-bit and 18-bit branch offsets.
25573 (MOVEP1, MOVEP2): New mode iterator.
25574 (mov_<load>l): Use ZC constraint.
25575 (mov_<load>r): Likewise.
25576 (mov_<store>l): Likewise.
25577 (mov_<store>r): Likewise.
25578 (*branch_equality<mode>_inverted): Add microMIPS support.
25579 (*branch_equality<mode>): Likewise.
25580 (*jump_absolute): Likewise.
25581 (indirect_jump_<mode>): Likewise.
25582 (tablejump_<mode>): Likewise.
25583 (<optab>_internal): Likewise.
25584 (sibcall_internal): Likewise.
25585 (sibcall_value_internal): Likewise.
25586 (prefetch): Use constraint ZD.
25587 * config/mips/mips.opt (minterlink-compressed): New option.
25588 (minterlink-mips16): Now an alias for minterlink-compressed.
25589 (mmicromips): New option.
25590 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25591 (compare_and_swap_12): Likewise.
25592 (sync_add<mode>): Likewise.
25593 (sync_<optab>_12): Likewise.
25594 (sync_old_<optab>_12): Likewise.
25595 (sync_new_<optab>_12): Likewise.
25596 (sync_nand_12): Likewise.
25597 (sync_old_nand_12): Likewise.
25598 (sync_new_nand_12): Likewise.
25599 (sync_sub<mode>): Likewise.
25600 (sync_old_add<mode>): Likewise.
25601 (sync_old_sub<mode>): Likewise.
25602 (sync_new_add<mode>): Likewise.
25603 (sync_new_sub<mode>): Likewise.
25604 (sync_<optab><mode>): Likewise.
25605 (sync_old_<optab><mode>): Likewise.
25606 (sync_new_<optab><mode>): Likewise.
25607 (sync_nand<mode>): Likewise.
25608 (sync_old_nand<mode>): Likewise.
25609 (sync_new_nand<mode>): Likewise.
25610 (sync_lock_test_and_set<mode>): Likewise.
25611 (test_and_set_12): Likewise.
25612 (atomic_compare_and_swap<mode>): Likewise.
25613 (atomic_exchange<mode>_llsc): Likewise.
25614 (atomic_fetch_add<mode>_llsc): Likewise.
25615 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25616 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25617 (umips_save_restore_pattern_p): Likewise.
25618 (umips_load_store_pair_p): Likewise.
25619 (umips_output_load_store_pair): Likewise.
25620 (umips_movep_target_p): Likewise.
25621 (umips_12bit_offset_address_p): Likewise.
25622 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25623 (mips_base_mips16): Rename this...
25624 (mips_base_compression_flags): ...to this. Update all uses.
25625 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25626 (mips_mips16_decl_p): Delete.
25627 (mips_nomips16_decl_p): Delete.
25628 (mips_get_compress_on_flags): New function.
25629 (mips_get_compress_off_flags): New function.
25630 (mips_get_compress_mode): New function.
25631 (mips_get_compress_on_name): New function.
25632 (mips_get_compress_off_name): New function.
25633 (mips_insert_attributes): Support multiple compression types.
25634 (mips_merge_decl_attributes): Likewise.
25635 (umips_12bit_offset_address_p): New function.
25636 (mips_start_function_definition): Emit .set micromips directive.
25637 (mips_call_may_need_jalx_p): New function.
25638 (mips_function_ok_for_sibcall): Add microMIPS support.
25639 (mips_print_operand_punctuation): Support short delay slots and
25641 (umips_swm_mask, umips_swm_encoding): New.
25642 (umips_build_save_restore): New function.
25643 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25644 (was_mips16_p): Remove.
25645 (old_compression_mode): New.
25646 (mips_set_compression_mode): New function.
25647 (mips_set_current_function): Add microMIPS support.
25648 (mips_option_override): Likewise.
25649 (umips_save_restore_pattern_p): New function.
25650 (umips_output_save_restore): New function.
25651 (umips_load_store_pair_p_1): New function.
25652 (umips_load_store_pair_p): New function.
25653 (umips_output_load_store_pair_1): New function.
25654 (umips_output_load_store_pair): New function.
25655 (umips_movep_target_p) New function.
25656 (mips_prepare_pch_save): Add microMIPS support.
25657 * config/mips/mips.h (TARGET_COMPRESSION): New.
25658 (TARGET_CPU_CPP_BUILTINS): Update macro
25659 to use new compression flags and to support microMIPS.
25660 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25661 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25662 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25663 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25664 (ASM_SPEC): Support mmicromips and mno-micromips.
25665 (M16STORE_REG_P): New macro.
25666 (MIPS_CALL): Support TARGET_MICROMIPS.
25667 (MICROMIPS_J): New macro.
25668 (mips_base_mips16): Rename this...
25669 (mips_base_compression_flags): ...to this.
25670 (UMIPS_12BIT_OFFSET_P): New macro.
25671 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25672 (MULTILIB_DIRNAMES): Likewise.
25673 2013-03-20 Richard Biener <rguenther@suse.de>
25675 PR tree-optimization/56661
25676 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
25677 the result does not have to be distinct.
25679 2013-03-20 Richard Biener <rguenther@suse.de>
25681 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
25684 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25685 Steven Bosscher <steven@gcc.gnu.org>
25687 PR rtl-optimization/56605
25688 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
25690 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25693 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
25694 that require movd instead of movq.
25696 2013-03-20 Richard Biener <rguenther@suse.de>
25698 * tree-ssa-structalias.c (struct variable_info): Add pointer
25699 to the first field of an aggregate with sub-vars. Make
25700 this and the pointer to the next subfield its ID.
25701 (vi_next): New function.
25702 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
25703 storedanything_id, integer_id): Increment by one.
25704 (new_var_info, get_call_vi, lookup_call_clobber_vi,
25705 get_call_clobber_vi): Adjust.
25706 (solution_set_expand): Simplify and speedup.
25707 (solution_set_add): Inline into ...
25708 (set_union_with_increment): ... this. Adjust accordingly.
25709 (do_sd_constraint): Likewise.
25710 (do_ds_constraint): Likewise.
25711 (do_complex_constraint): Simplify.
25712 (build_pred_graph): Adjust.
25713 (solve_graph): Likewise. Simplify and speedup.
25714 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
25715 get_constraint_for_component_ref, get_constraint_for_1,
25716 first_vi_for_offset, first_or_preceding_vi_for_offset,
25717 create_function_info_for, create_variable_info_for_1,
25718 create_variable_info_for, intra_create_variable_infos): Adjust.
25719 (init_base_vars): Push NULL for ID zero.
25720 (compute_points_to_sets): Adjust.
25722 2013-03-20 Richard Biener <rguenther@suse.de>
25724 * cfgloop.c (verify_loop_structure): Streamline and avoid
25725 ICEing on corrupt loop tree.
25726 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
25729 2013-03-20 Richard Biener <rguenther@suse.de>
25731 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
25732 check whether an SSA update is needed.
25734 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
25736 * config/mips/constraints.md (T): Rename to...
25740 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
25741 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
25743 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25745 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
25746 (*subsi3_carryin_uxtw): Likewise.
25748 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25750 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
25751 (*rorsi3_insn_uxtw): Likewise.
25753 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25755 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
25756 (*extrsi5_insn_uxtw): Likewise.
25758 2013-03-19 Richard Biener <rguenther@suse.de>
25760 PR tree-optimization/56273
25761 * passes.c (init_optimization_passes): Move second VRP after DOM.
25763 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
25765 * config/i386/i386.md (*movti_internal): Merge from
25766 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
25767 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
25768 nox64 isa attributes.
25770 2013-03-18 Richard Biener <rguenther@suse.de>
25772 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
25774 (merge_node_constraints): Likewise.
25775 (build_succ_graph): Likewise.
25776 (valid_graph_edge): Inline into single caller.
25777 (unify_nodes): Likewise. Use bitmap_set_bit return value
25779 (scc_visit): Fix formatting and variable use.
25780 (do_sd_constraint): Use gcc_checking_assert.
25781 (do_ds_constraint): Likewise.
25782 (do_complex_constraint): Likewise.
25783 (condense_visit): Likewise. Cleanup.
25784 (dump_pred_graph): New function.
25785 (perform_var_substitution): Dump the pred-graph before
25786 variable substitution.
25787 (find_equivalent_node): Use gcc_checking_assert.
25788 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
25790 2013-03-18 Richard Biener <rguenther@suse.de>
25792 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
25793 Remove cond_expr_stmt_list argument and do not gimplify the
25795 (vect_loop_versioning): Adjust.
25796 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
25797 Cleanup to use less temporaries.
25798 (vect_create_data_ref_ptr): Cleanup.
25800 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25802 PR tree-optimization/56635
25803 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
25804 require types_compatible_p types.
25806 2013-03-18 Nick Clifton <nickc@redhat.com>
25808 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
25809 spurious backslash.
25811 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
25812 Add missing line to comment describing function.
25814 2013-03-18 Richard Biener <rguenther@suse.de>
25816 PR tree-optimization/56210
25817 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25818 Handle string / character search functions.
25819 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25821 2013-03-18 Richard Biener <rguenther@suse.de>
25823 PR middle-end/56483
25824 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25825 and implement properly.
25826 * gimple.h (gimple_cond_single_var_p): Remove.
25828 2013-03-18 Richard Biener <rguenther@suse.de>
25830 * tree-data-ref.h (find_data_references_in_loop): Declare.
25831 * tree-data-ref.c (get_references_in_stmt): Use a stack
25832 vector pre-allocated in the callers.
25833 (find_data_references_in_stmt): Adjust.
25834 (graphite_find_data_references_in_stmt): Likewise.
25835 (create_rdg_vertices): Likewise.
25836 (find_data_references_in_loop): Export.
25837 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25838 Compute dependences here...
25839 (vect_analyze_data_refs): ...not here. When we encounter
25840 a non-vectorizable data reference in basic-block vectorization
25841 truncate the data reference vector. Do not bother to
25842 fixup data-dependence information for gather loads.
25843 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25844 of data references, as reported.
25846 2013-03-18 Richard Biener <rguenther@suse.de>
25848 PR tree-optimization/3713
25849 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25850 has_constants and expr.
25851 (stmt_has_constants): Properly valueize SSA names when deciding
25852 whether the stmt has constants.
25854 2013-03-18 Richard Biener <rguenther@suse.de>
25856 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
25857 whole function when there is nothing to do.
25858 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
25859 * tree-vectorizer.c (vectorize_loops): Update virtual and
25860 loop-closed SSA once.
25861 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
25863 2013-03-18 Richard Biener <rguenther@suse.de>
25865 PR middle-end/56113
25866 * domwalk.c (bb_postorder): New global static.
25867 (cmp_bb_postorder): New function.
25868 (walk_dominator_tree): Replace scheme imposing an order for
25869 visiting dominator sons by one sorting them at the time they
25870 are pushed on the stack.
25872 2013-03-18 Richard Biener <rguenther@suse.de>
25874 PR tree-optimization/39326
25875 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
25876 (struct mem_ref): Replace mem member with ao_ref typed member.
25877 (MEM_ANALYZABLE): Adjust.
25878 (memref_eq): Likewise.
25879 (mem_ref_alloc): Likewise.
25880 (gather_mem_refs_stmt): Likewise.
25881 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
25882 (execute_sm_if_changed_flag_set): Adjust.
25883 (execute_sm): Likewise.
25884 (ref_always_accessed_p): Likewise.
25885 (refs_independent_p): Likewise.
25886 (can_sm_ref_p): Likewise.
25888 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25891 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
25892 return 1 even for !unsignedp.
25894 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25896 * config/i386/i386.md (isa): Add x64 and nox64.
25897 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
25898 (*pushtf): Enable *roF alternative for x64 isa only.
25899 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
25900 mode attribute of integer alternatives to DImode for TARGET_64BIT.
25901 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
25902 (*movtf_internal): Merge from *movtf_internal_rex64 and
25903 *movtf_internal_sse. Use x64 and nox64 isa attributes.
25904 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
25905 nox64 isa attributes.
25906 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
25907 nox64 isa attributes.
25908 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
25910 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25912 * config/alpha/alpha.c (TARGET_LRA_P): New define.
25914 2013-03-17 Jakub Jelinek <jakub@redhat.com>
25917 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
25918 class names. Remove trailing comma after "ALL_REGS".
25920 2013-03-16 Jan Hubicka <jh@suse.cz>
25922 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
25923 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
25924 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
25925 of cgraph_get_create_node.
25926 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
25928 2013-03-16 Jason Merrill <jason@redhat.com>
25931 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
25932 with DW_AT_default_value.
25934 2013-03-16 Jakub Jelinek <jakub@redhat.com>
25936 * BASE-VER: Set to 4.9.0.
25938 2013-03-14 Andi Kleen <ak@linux.intel.com>
25941 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
25942 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
25943 Document _x* TSX intrinsics.
25945 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
25946 David Holsgrove <david.holsgrove@xilinx.com>
25948 * configure.ac: Add MicroBlaze TLS support detection.
25949 * configure: Regenerate.
25950 * config/microblaze/microblaze-protos.h
25951 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
25952 symbol_mentioned_p, label_mentioned_p): Add prototypes.
25953 * config/microblaze/microblaze.c (microblaze_address_type): Add
25954 ADDRESS_TLS and tls_reloc address types.
25955 (microblaze_address_info): Add tls_reloc.
25956 (TARGET_HAVE_TLS): Define.
25957 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
25958 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
25959 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
25960 load_tls_operand, microblaze_call_tls_get_addr,
25961 microblaze_legitimize_tls_address): New functions.
25962 (microblaze_classify_unspec): Handle UNSPEC_TLS.
25963 (get_base_reg): Use microblaze_tls_symbol_p.
25964 (microblaze_classify_address): Handle TLS.
25965 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
25966 label_mentioned_p and microblaze_tls_referenced_p.
25967 (microblaze_legitimize_address): Handle TLS.
25968 (microblaze_address_insns): Handle ADDRESS_TLS.
25969 (pic_address_needs_scratch): Handle TLS.
25970 (print_operand_address): Handle TLS.
25971 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
25972 (microblaze_expand_move): Handle TLS.
25973 (microblaze_legitimate_constant_p): Check
25974 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
25975 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
25976 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
25977 (PIC_OFFSET_TABLE_REGNUM): Set.
25978 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
25979 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
25980 (addsi3, movsi_internal2, movdf_internal): Update constraints
25981 * config/microblaze/predicates.md (arith_plus_operand): Define
25982 (move_operand): Redefine as move_src_operand,
25983 check microblaze_tls_referenced_p.
25985 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25987 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
25988 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
25990 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25992 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
25995 2013-03-14 Jakub Jelinek <jakub@redhat.com>
25997 PR tree-optimization/53265
25998 * common.opt (Waggressive-loop-optimizations): New option.
25999 * tree-ssa-loop-niter.c: Include tree-pass.h.
26000 (do_warn_aggressive_loop_optimizations): New function.
26001 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
26002 if number_of_latch_executions returned constant.
26003 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
26004 early. If number_of_latch_executions returned constant, set
26005 nb_iterations_upper_bound back to it.
26006 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
26008 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
26009 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
26011 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
26012 (MULTILIB_OSDIRNAMES): Set.
26013 * genmultilib: If defaultosdirname doesn't start with :: , set
26014 defaultosdirname2 instead, clear it and emit two . multilib_raw
26015 entries instead of just one.
26017 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26019 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26020 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26021 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26022 (SUBTARGET_OVERRIDE_OPTIONS): New.
26024 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26027 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26028 (musermode): Convert to Var(TARGET_USERMODE).
26029 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26030 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26031 * config/sh/sh.c (sh_option_override): Use
26032 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26033 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26035 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26037 (udivsi3_i4_single, divsi3_i4_single): Use
26038 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26040 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26042 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26045 2013-03-13 Richard Biener <rguenther@suse.de>
26047 PR tree-optimization/56608
26048 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26049 calls when vectorizing basic-blocks.
26051 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26054 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26057 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26059 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26061 2013-03-11 Jan Hubicka <jh@suse.cz>
26064 * lto-streamer-out.c (output_symbol_p): Skip references from
26065 constructors of external variables.
26067 2013-03-11 Jan Hubicka <jh@suse.cz>
26069 PR middle-end/56571
26070 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26072 * emit-rtl.c (verify_rtx_sharing): Likewise.
26073 (copy_insn_1): Likewise.
26074 * rtl.c (copy_rtx): Likewise.
26076 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26079 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26080 output_operand_lossage message.
26082 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26085 * arm.c (shift_op): Validate RTL pattern on the fly.
26086 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26089 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26092 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26093 new call patterns that clobber %fr12.
26094 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26095 split and postreload patterns.
26096 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26097 registers %fr12 and %fr12R as call used.
26099 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26101 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26102 (canon_address, record_store, replace_read, check_mem_read_rtx,
26103 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26104 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26105 rest_of_handle_dse): Likewise.
26107 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26109 PR middle-end/56524
26110 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26112 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26113 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26114 (save_optabs_if_changed): Replace with...
26115 (init_tree_optimization_optabs): ...this.
26116 * optabs.c (save_optabs_if_changed): Rename to...
26117 (init_tree_optimization_optabs): ...this. Take the optimization node
26118 as argument. Do nothing if the base optabs are already correct.
26119 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26120 to recompute optabs.
26121 * function.h (function): Remove optabs field.
26122 * function.c (invoke_set_current_function_hook): Call
26123 init_tree_optimization_optabs. Use the result to initialize
26126 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26128 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26129 if GTMA_HAS_NO_INSTRUMENTATION.
26130 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26131 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26132 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26133 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26134 GTMA_HAS_NO_INSTRUMENTATION.
26136 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26138 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26141 2013-03-08 Marek Polacek <polacek@redhat.com>
26142 Jakub Jelinek <jakub@redhat.com>
26144 PR tree-optimization/56478
26145 * predict.c (is_comparison_with_loop_invariant_p): Change the
26146 type of loop_step to tree.
26147 (predict_loops): Adjust.
26148 (predict_iv_comparison): Perform the computations on double_ints.
26150 2013-03-08 Richard Biener <rguenther@suse.de>
26152 PR tree-optimization/56570
26153 * tree-cfg.c (verify_expr_location_1): Verify locations for
26155 * tree-sra.c (create_access_replacement): Strip locations
26156 from DECL_DEBUG_EXPRs.
26158 2013-03-08 Richard Biener <rguenther@suse.de>
26160 * tree-inline.c (expand_call_inline): Do not associate
26161 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26162 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
26164 2013-03-08 Richard Biener <rguenther@suse.de>
26166 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
26167 or block changes with -Og. Fix for location / block encoding
26168 changes and PHI arguments with locations.
26170 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
26172 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
26174 (struct output_info): Likewise.
26175 (register_overhead): Remove bad gcc_assert.
26176 (bitmap_find_bit): If there is only a single bitmap element, do not
26177 count a miss as a search.
26178 (print_statistics): Update for counter type changes.
26179 (dump_bitmap_statistics): Likewise. Print headers such that they
26180 are properly lined up with the printed counters.
26182 2013-03-07 Jakub Jelinek <jakub@redhat.com>
26184 PR tree-optimization/56559
26185 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
26186 check that it has only a single use.
26188 2013-03-07 Richard Biener <rguenther@suse.de>
26190 * doc/invoke.texi (fwhole-program): Discourage use in combination
26193 2013-03-06 Jakub Jelinek <jakub@redhat.com>
26195 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
26197 PR tree-optimization/56539
26198 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
26199 instead of GSI_CONTINUE_LINKING as last argument to
26200 force_gimple_operand_gsi. Adjust function comment.
26202 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
26205 PR middle-end/56548
26206 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
26207 promoted mode, convert the result back to the original mode.
26209 2013-03-06 Richard Biener <rguenther@suse.de>
26211 PR middle-end/56294
26212 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26213 (insert_updated_phi_nodes_compare_uids): New function.
26214 (update_ssa): Sort symbols_to_rename after UID before
26215 traversing it to insert PHI nodes.
26217 2013-03-06 Richard Biener <rguenther@suse.de>
26219 PR middle-end/50494
26220 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26221 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26224 2013-02-13 Richard Biener <rguenther@suse.de>
26227 * varasm.c (output_constant_def_1): Get the decl representing
26228 the constant as argument.
26229 (output_constant_def): Wrap output_constant_def_1.
26230 (make_decl_rtl): Use output_constant_def_1 with the decl
26231 representing the constant.
26232 (build_constant_desc): Optionally re-use a decl already
26233 representing the constant.
26234 (tree_output_constant_def): Adjust.
26236 2013-03-06 Joey Ye <joey.ye@arm.com>
26239 * gcc.c (convert_white_space): New function.
26240 (main): Handles white space in function name.
26242 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26245 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26246 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26247 to SH_DIV_CALL_TABLE for TARGET_SH2.
26248 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26250 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26251 call-table options.
26253 2013-03-05 Sterling Augustine <saugustine@google.com>
26254 Cary Coutant <ccoutant@google.com>
26257 * dwarf2out.c (resolve_addr): Don't call
26258 remove_loc_list_addr_table_entries a second time for the same
26261 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26264 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26265 (avoid_complex_debug_insns): New function.
26266 (expand_debug_locations): Call it.
26268 PR rtl-optimization/56484
26269 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26270 lifetimes of hard registers on small register class machines.
26272 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26274 * config/microblaze/microblaze-protos.h: Rename
26275 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26276 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26278 (microblaze_fast_interrupt_function_p): New function.
26279 (microblaze_is_interrupt_handler): Rename to
26280 microblaze_is_interrupt_variant and add fast_interrupt check.
26281 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26282 (save_restore_insns): Likewise.
26283 (compute_frame_size): Likewise.
26284 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26285 (microblaze_globalize_label): Likewise.
26286 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26287 * config/microblaze/microblaze.md: Use wrapper
26288 microblaze_is_interrupt_variant.
26290 2013-03-05 Kai Tietz <ktietz@redhat.com>
26292 * sdbout.c (sdbout_one_type): Switch to current function's section
26293 supporting cold/hot.
26295 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26297 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26300 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26302 PR middle-end/56461
26303 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26304 if VALGRIND_GET_VBITS is defined, temporarily make object
26305 memory all defined, and restore previous valgrind addressability
26306 and definability afterwards. Free this_object at the end.
26308 PR middle-end/56461
26309 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26310 right before calling lra_create_live_ranges, also call it
26311 when clearing live_p. Only call lra_clear_live_ranges
26312 at the end if live_p.
26314 PR middle-end/56461
26315 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26317 2013-03-05 Richard Biener <rguenther@suse.de>
26319 PR tree-optimization/56521
26320 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26323 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26326 * except.h (remove_unreachable_eh_regions): New prototype.
26327 * except.c (remove_eh_handler_splicer): New function, split out
26328 of remove_eh_handler.
26329 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26330 warning about running it on many EH regions one at a time.
26331 (remove_unreachable_eh_regions_worker): New function, walk the
26332 EH tree in depth-first order and remove non-marked regions.
26333 (remove_unreachable_eh_regions): New function.
26334 * tree-eh.c (mark_reachable_handlers): New function, split out
26335 from remove_unreachable_handlers.
26336 (remove_unreachable_handlers): Use mark_reachable_handlers and
26337 remove_unreachable_eh_regions.
26338 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26339 and remove_unreachable_eh_regions.
26341 2013-03-05 Richard Biener <rguenther@suse.de>
26343 PR middle-end/56525
26344 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26345 not freeing them until the end.
26347 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26349 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26351 2013-03-05 Richard Biener <rguenther@suse.de>
26353 PR tree-optimization/56270
26354 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26355 of loads after scheduling an SLP instance.
26357 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26359 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26361 (check_gcc_parallelize): Run guality.exp as a separate job from
26362 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26363 struct-layout-1.exp with stackalign.exp.
26365 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26367 PR middle-end/56461
26368 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26369 load_index sbitmap even if some bit in it isn't set.
26371 PR middle-end/56461
26372 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26373 (discover_iteration_bound_by_body_walk): Change queues to
26374 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26375 spelling in comment. Call safe_push on queues[bound_index] directly.
26376 Release queues[queue_index] in every iteration unconditionally.
26377 Release bounds vector.
26379 PR middle-end/56461
26380 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26381 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26382 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26385 2013-03-05 Richard Biener <rguenther@suse.de>
26388 * tree-inline.c (remap_blocks_to_null): New function.
26389 (expand_call_inline): When expanding a call stmt without
26390 an associated block inline remap all callee blocks to NULL.
26392 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26394 PR rtl-optimization/56494
26395 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26396 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26397 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26399 PR middle-end/56461
26400 * sel-sched-ir.c (free_sched_pools): Release
26401 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26402 if succs_info_pool.max_top isn't -1.
26405 * opts.c (opts_obstack, opts_concat): Moved to...
26406 * opts-common.c (opts_obstack, opts_concat): ... here.
26408 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26410 PR middle-end/56461
26411 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26413 2013-03-04 Martin Jambor <mjambor@suse.cz>
26415 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26416 all appropriate places.
26418 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26420 PR tree-optimization/56424
26421 * ipa-split.c (split_function): Do not set the RSO flag if result is
26422 not by reference and its type is a register type.
26424 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26426 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26427 (microblaze_legitimate_pic_operand): Likewise
26428 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26429 new function microblaze_legitimate_pic_operand
26430 * config/microblaze/microblaze-protos.h
26431 (microblaze_legitimate_pic_operand): Declare.
26433 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26435 * config/microblaze/predicates.md (call_insn_simple_operand):
26436 New predicate for supported rtx code types.
26437 * config/microblaze/microblaze.md (call_internal1): Use
26438 call_insn_simple_operand predicate.
26440 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26442 PR middle-end/56461
26443 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26444 partitions.ordered_remove.
26446 PR middle-end/56461
26447 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26448 vec_oprnds0.create (1) for modifier == NONE.
26450 PR middle-end/56461
26451 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26452 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26453 vec_oprnds1 right before pushing anything to it for
26456 PR middle-end/56461
26457 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26458 set nbbs to 0 instead of having separate code path.
26459 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26460 instead of false as last argument if returning NULL.
26462 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26464 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26465 the attribute is now called "target" instead of "option".
26466 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26467 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26468 attribute/pragma name for TARGET_OPTION_VALID_P and
26469 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26470 * doc/tm.texi: Regenerated.
26472 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26474 * config/microblaze/microblaze.c:
26475 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26476 * config/microblaze/microblaze.h: Add -mxl-reorder to
26478 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26479 instructions emitted if TARGET_REORDER.
26480 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26481 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26484 2013-03-01 Xinliang David Li <davidxl@google.com>
26486 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26489 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26491 PR middle-end/56461
26492 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26493 vector even when returning true. Fix up function comment formatting.
26495 PR middle-end/56461
26496 * ira-build.c (ira_loop_nodes_count): New variable.
26497 (create_loop_tree_nodes): Initialize it.
26498 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26500 PR middle-end/56461
26501 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26502 method on dr_chain and result_chain.
26503 * tree-vect-stmts.c (vectorizable_store): Only call
26504 result_chain.create if j == 0.
26506 PR middle-end/56461
26507 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26508 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26509 before overwriting it.
26511 2013-03-01 Tobias Burnus <burnus@net-b.de>
26513 * doc/extended.texi (C Extensions): Change order in @menu
26515 (Other MIPS Built-in Functions): Move last MIPS entry before
26516 "picoChip Built-in Functions".
26517 (SH Built-in Functions): Move after RX Built-in Functions.
26518 * doc/gcc.texi (Introduction): Change order in @menu
26520 * doc/md.texi (Constraints): Ditto.
26521 * gty.texi (Type Information): Ditto.
26522 (User-provided marking routines for template types): Make
26524 * doc/invoke.texi (AArch64 Options): Move before
26525 "Adapteva Epiphany Options".
26527 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26528 Jakub Jelinek <jakub@redhat.com>
26531 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26532 no_address_safety_analysis attribute.
26533 * doc/extend.texi (no_address_safety_attribute): Rename to
26534 no_sanitize_address attribute, mention no_address_safety_analysis
26535 attribute as deprecated alias.
26537 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26539 PR middle-end/56461
26540 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26541 type to vec<vec<tree> > *.
26542 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26543 to be vec<tree> instead of vec<tree> *, set vec_defs
26544 to vNULL and call vec_defs.create (number_of_vects), adjust other
26546 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26547 vectorizable_condition): Adjust vect_get_slp_defs callers.
26549 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26551 * config/aarch64/aarch64.c
26552 (aarch64_float_const_representable): Remove unused variable.
26554 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26556 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26558 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26560 * config/aarch64/aarch64-builtins.c
26561 (aarch64_init_simd_builtins): Make static.
26563 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26565 * config/aarch64/aarch64.c
26566 (aarch64_simd_make_constant): Make static.
26568 2013-02-28 Martin Jambor <mjambor@suse.cz>
26570 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26571 with no initialization to the RHS of debug statements.
26573 2013-02-28 Martin Jambor <mjambor@suse.cz>
26575 PR tree-optimization/56294
26576 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26578 (get_access_replacement): Do not call create_access_replacement.
26579 Assert a replacement exists.
26580 (get_repl_default_def_ssa_name): Create the replacement declaration
26583 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26585 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26586 final_end_function.
26588 2013-02-28 Marek Polacek <polacek@redhat.com>
26590 PR rtl-optimization/56466
26591 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26592 if we're changing a loop.
26593 (peel_loops_completely): Likewise.
26595 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26598 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26600 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26603 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26604 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26605 INTX_FTYPE_FX, FX_FTYPE_INTX.
26606 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26608 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26610 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26611 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26612 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26613 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26614 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26615 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26616 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26617 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26618 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26619 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26620 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26621 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26622 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26623 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26624 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26625 (avrxmega6): Increase max flash segments from 5 to 6.
26626 * config/avr/t-multilib: Regenerate.
26627 * config/avr/avr-tables.opt: Regenerate.
26628 * doc/avr-mmcu.texi: Regenerate.
26630 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26632 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
26633 (avr_device_to_arch): Rename to avr_device_to_ld.
26634 (avr_device_to_as): New prototype.
26635 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
26636 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
26637 * config/avr/driver-avr.c (avr_device_to_as): New.
26638 (avr_device_to_arch): Rename to avr_device_to_ld.
26640 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26642 PR middle-end/56461
26643 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
26644 method on dr_chain and result_chain.
26646 PR middle-end/56461
26647 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
26648 pointer_set_destroy on not_executed_last_iteration.
26650 PR middle-end/56461
26651 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
26653 PR middle-end/56461
26654 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
26655 FOR_EACH_DEFINED_FUNCTION when freeing state.
26657 PR middle-end/56461
26658 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
26660 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
26663 PR middle-end/56461
26664 * ipa-cp.c (decide_whether_version_node): Call vec_free on
26665 known_aggs[i].items and release known_aggs vector.
26667 PR middle-end/56461
26668 * ipa-reference.c (propagate): Free node_info even for alias nodes.
26670 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26672 * config/microblaze/microblaze.c (microblaze_emit_compare):
26673 Use xor for EQ/NE comparisions.
26674 * config/microblaze/microblaze.md (cstoresf4): Add constraints
26675 (cbranchsf4): Adjust operator to comparison_operator.
26677 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26679 PR middle-end/56461
26680 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
26682 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
26683 vec_safe_push, always update *slot.
26684 (redirect_edge_var_map_clear): Use vec_free.
26685 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
26686 (free_var_map_entry): Use vec_free.
26687 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
26688 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
26690 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
26692 PR middle-end/45472
26693 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
26694 when the may_trap_p bit of the exprs being merged differs.
26695 Reorder tests for speculativeness in the logical and operator.
26697 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26699 * incpath.c (add_standard_paths): Use reconcat instead of concat
26700 where appropriate and avoid leaking memory.
26702 * opts.h: Include obstack.h.
26703 (opts_concat): New prototype.
26704 (opts_obstack): New declaration.
26705 * opts.c (opts_concat): New function.
26706 (opts_obstack): New variable.
26707 (init_options_struct): Call gcc_init_obstack on opts_obstack.
26708 (finish_options): Use opts_concat instead of concat
26709 and XOBNEWVEC instead of XNEWVEC.
26710 * opts-common.c (generate_canonical_option, decode_cmdline_option,
26711 generate_option): Likewise.
26712 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
26713 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
26716 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
26717 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
26719 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26721 PR middle-end/56461
26722 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
26724 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26726 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
26727 (arm_block_move_unaligned_straight): Likewise.
26728 (arm_adjust_block_mem): Likewise.
26730 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26733 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
26734 temp, cond and label.
26735 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
26738 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
26739 * config/c6x/c6x.h (dbx_register_map): Update declaration.
26742 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
26743 of prologue/epilogue functions.
26746 * config/tilegx/tilegx.c (tilegx_expand_prologue):
26747 Remove unused variable cfa_offset.
26748 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
26751 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
26752 type promotion to unsigned.
26755 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
26756 for HOST_WIDE_INT of 32 bit / same size as int.
26757 (arm_block_move_unaligned_straight): Likewise.
26758 (arm_adjust_block_mem): Likewise.
26761 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
26764 2013-02-26 Marek Polacek <polacek@redhat.com>
26766 PR tree-optimization/56426
26767 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
26769 2013-02-26 Richard Biener <rguenther@suse.de>
26772 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
26773 unused variable loops.
26775 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26777 PR tree-optimization/56448
26778 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
26779 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
26780 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
26781 later operands of the references, or even first operand for
26782 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
26784 PR tree-optimization/56443
26785 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
26786 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
26787 to type_for_mode langhook.
26789 2013-02-25 Matt Turner <mattst88@gmail.com>
26791 * doc/invoke.texi: Document r4700.
26793 2013-02-25 Richard Biener <rguenther@suse.de>
26795 PR tree-optimization/56175
26796 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
26798 (simplify_bitwise_binary): ... here. Also guard the conversion
26799 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
26801 2013-02-25 Catherine Moore <clm@codesourcery.com>
26804 2013-02-24 Catherine Moore <clm@codesourcery.com>
26805 Maciej W. Rozycki <macro@codesourcery.com>
26806 Tom de Vries <tom@codesourcery.com>
26807 Nathan Sidwell <nathan@codesourcery.com>
26808 Iain Sandoe <iain@codesourcery.com>
26809 Nathan Froyd <froydnj@codesourcery.com>
26810 Chao-ying Fu <fu@mips.com>
26812 * doc/extend.texi: (micromips, nomicromips, nocompression):
26813 Document new function attributes.
26814 * doc/invoke.texi (minterlink-compressed, mmicromips,
26815 m14k, m14ke, m14kec): Document new options.
26816 (minterlink-mips16): Update documentation.
26817 * doc/md.texi (ZC, ZD): Document new constraints.
26818 * configure.ac (gcc_cv_as_micromips): Check if linker
26819 supports the .set micromips directive.
26820 * configure: Regenerate.
26821 * config.in: Regenerate.
26822 * config/mips/mips-tables.opt: Regenerate.
26823 * config/mips/micromips.md: New file.
26824 * constraints.md (ZC, AD): New constraints.
26825 * config/mips/predicates.md (movep_src_register): New predicate.
26826 (movep_src_operand): New predicate.
26827 (non_volatile_mem_operand): New predicate.
26828 * config/mips/mips.md (multimem): New type.
26829 (length): Differentiate between 17-bit and 18-bit branch offsets.
26830 (MOVEP1, MOVEP2): New mode iterator.
26831 (mov_<load>l): Use ZC constraint.
26832 (mov_<load>r): Likewise.
26833 (mov_<store>l): Likewise.
26834 (mov_<store>r): Likewise.
26835 (*branch_equality<mode>_inverted): Add microMIPS support.
26836 (*branch_equality<mode>): Likewise.
26837 (*jump_absolute): Likewise.
26838 (indirect_jump_<mode>): Likewise.
26839 (tablejump_<mode>): Likewise.
26840 (<optab>_internal): Likewise.
26841 (sibcall_internal): Likewise.
26842 (sibcall_value_internal): Likewise.
26843 (prefetch): Use constraint ZD.
26844 * config/mips/mips.opt (minterlink-compressed): New option.
26845 (minterlink-mips16): Now an alias for minterlink-compressed.
26846 (mmicromips): New option.
26847 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26848 (compare_and_swap_12): Likewise.
26849 (sync_add<mode>): Likewise.
26850 (sync_<optab>_12): Likewise.
26851 (sync_old_<optab>_12): Likewise.
26852 (sync_new_<optab>_12): Likewise.
26853 (sync_nand_12): Likewise.
26854 (sync_old_nand_12): Likewise.
26855 (sync_new_nand_12): Likewise.
26856 (sync_sub<mode>): Likewise.
26857 (sync_old_add<mode>): Likewise.
26858 (sync_old_sub<mode>): Likewise.
26859 (sync_new_add<mode>): Likewise.
26860 (sync_new_sub<mode>): Likewise.
26861 (sync_<optab><mode>): Likewise.
26862 (sync_old_<optab><mode>): Likewise.
26863 (sync_new_<optab><mode>): Likewise.
26864 (sync_nand<mode>): Likewise.
26865 (sync_old_nand<mode>): Likewise.
26866 (sync_new_nand<mode>): Likewise.
26867 (sync_lock_test_and_set<mode>): Likewise.
26868 (test_and_set_12): Likewise.
26869 (atomic_compare_and_swap<mode>): Likewise.
26870 (atomic_exchange<mode>_llsc): Likewise.
26871 (atomic_fetch_add<mode>_llsc): Likewise.
26872 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26873 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26874 (umips_save_restore_pattern_p): Likewise.
26875 (umips_load_store_pair_p): Likewise.
26876 (umips_output_load_store_pair): Likewise.
26877 (umips_movep_target_p): Likewise.
26878 (umips_12bit_offset_address_p): Likewise.
26879 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26880 (mips_base_mips16): Rename this...
26881 (mips_base_compression_flags): ...to this. Update all uses.
26882 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26883 (mips_mips16_decl_p): Delete.
26884 (mips_nomips16_decl_p): Delete.
26885 (mips_get_compress_on_flags): New function.
26886 (mips_get_compress_off_flags): New function.
26887 (mips_get_compress_mode): New function.
26888 (mips_get_compress_on_name): New function.
26889 (mips_get_compress_off_name): New function.
26890 (mips_insert_attributes): Support multiple compression types.
26891 (mips_merge_decl_attributes): Likewise.
26892 (umips_12bit_offset_address_p): New function.
26893 (mips_start_function_definition): Emit .set micromips directive.
26894 (mips_call_may_need_jalx_p): New function.
26895 (mips_function_ok_for_sibcall): Add microMIPS support.
26896 (mips_print_operand_punctuation): Support short delay slots and
26898 (umips_swm_mask, umips_swm_encoding): New.
26899 (umips_build_save_restore): New function.
26900 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26901 (was_mips16_p): Remove.
26902 (old_compression_mode): New.
26903 (mips_set_compression_mode): New function.
26904 (mips_set_current_function): Add microMIPS support.
26905 (mips_option_override): Likewise.
26906 (umips_save_restore_pattern_p): New function.
26907 (umips_output_save_restore): New function.
26908 (umips_load_store_pair_p_1): New function.
26909 (umips_load_store_pair_p): New function.
26910 (umips_output_load_store_pair_1): New function.
26911 (umips_output_load_store_pair): New function.
26912 (umips_movep_target_p) New function.
26913 (mips_prepare_pch_save): Add microMIPS support.
26914 * config/mips/mips.h (TARGET_COMPRESSION): New.
26915 (TARGET_CPU_CPP_BUILTINS): Update macro
26916 to use new compression flags and to support microMIPS.
26917 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26918 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26919 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26920 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26921 (ASM_SPEC): Support mmicromips and mno-micromips.
26922 (M16STORE_REG_P): New macro.
26923 (MIPS_CALL): Support TARGET_MICROMIPS.
26924 (MICROMIPS_J): New macro.
26925 (mips_base_mips16): Rename this...
26926 (mips_base_compression_flags): ...to this.
26927 (UMIPS_12BIT_OFFSET_P): New macro.
26928 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26929 (MULTILIB_DIRNAMES): Likewise.
26931 2013-02-25 Tom de Vries <tom@codesourcery.com>
26933 PR rtl-optimization/56131
26934 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
26935 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
26936 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
26938 2013-02-25 Tobias Burnus <burnus@net-b.de>
26940 * doc/invoke.texi (-fsanitize=): Move from optimization
26941 to debugging options.
26943 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26945 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
26947 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26948 Alexander Monakov <amonakov@ispras.ru>
26950 PR middle-end/56077
26951 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
26952 flush pending lists also on non-jumps. Adjust comment.
26954 2013-02-24 Catherine Moore <clm@codesourcery.com>
26955 Maciej W. Rozycki <macro@codesourcery.com>
26956 Tom de Vries <tom@codesourcery.com>
26957 Nathan Sidwell <nathan@codesourcery.com>
26958 Iain Sandoe <iain@codesourcery.com>
26959 Nathan Froyd <froydnj@codesourcery.com>
26960 Chao-ying Fu <fu@mips.com>
26962 * doc/extend.texi: (micromips, nomicromips, nocompression):
26963 Document new function attributes.
26964 * doc/invoke.texi (minterlink-compressed, mmicromips,
26965 m14k, m14ke, m14kec): Document new options.
26966 (minterlink-mips16): Update documentation.
26967 * doc/md.texi (ZC, ZD): Document new constraints.
26968 * configure.ac (gcc_cv_as_micromips): Check if linker
26969 supports the .set micromips directive.
26970 * configure: Regenerate.
26971 * config.in: Regenerate.
26972 * config/mips/mips-tables.opt: Regenerate.
26973 * config/mips/micromips.md: New file.
26974 * constraints.md (ZC, AD): New constraints.
26975 * config/mips/predicates.md (movep_src_register): New predicate.
26976 (movep_src_operand): New predicate.
26977 (non_volatile_mem_operand): New predicate.
26978 * config/mips/mips.md (multimem): New type.
26979 (length): Differentiate between 17-bit and 18-bit branch offsets.
26980 (MOVEP1, MOVEP2): New mode iterator.
26981 (mov_<load>l): Use ZC constraint.
26982 (mov_<load>r): Likewise.
26983 (mov_<store>l): Likewise.
26984 (mov_<store>r): Likewise.
26985 (*branch_equality<mode>_inverted): Add microMIPS support.
26986 (*branch_equality<mode>): Likewise.
26987 (*jump_absolute): Likewise.
26988 (indirect_jump_<mode>): Likewise.
26989 (tablejump_<mode>): Likewise.
26990 (<optab>_internal): Likewise.
26991 (sibcall_internal): Likewise.
26992 (sibcall_value_internal): Likewise.
26993 (prefetch): Use constraint ZD.
26994 * config/mips/mips.opt (minterlink-compressed): New option.
26995 (minterlink-mips16): Now an alias for minterlink-compressed.
26996 (mmicromips): New option.
26997 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26998 (compare_and_swap_12): Likewise.
26999 (sync_add<mode>): Likewise.
27000 (sync_<optab>_12): Likewise.
27001 (sync_old_<optab>_12): Likewise.
27002 (sync_new_<optab>_12): Likewise.
27003 (sync_nand_12): Likewise.
27004 (sync_old_nand_12): Likewise.
27005 (sync_new_nand_12): Likewise.
27006 (sync_sub<mode>): Likewise.
27007 (sync_old_add<mode>): Likewise.
27008 (sync_old_sub<mode>): Likewise.
27009 (sync_new_add<mode>): Likewise.
27010 (sync_new_sub<mode>): Likewise.
27011 (sync_<optab><mode>): Likewise.
27012 (sync_old_<optab><mode>): Likewise.
27013 (sync_new_<optab><mode>): Likewise.
27014 (sync_nand<mode>): Likewise.
27015 (sync_old_nand<mode>): Likewise.
27016 (sync_new_nand<mode>): Likewise.
27017 (sync_lock_test_and_set<mode>): Likewise.
27018 (test_and_set_12): Likewise.
27019 (atomic_compare_and_swap<mode>): Likewise.
27020 (atomic_exchange<mode>_llsc): Likewise.
27021 (atomic_fetch_add<mode>_llsc): Likewise.
27022 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27023 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27024 (umips_save_restore_pattern_p): Likewise.
27025 (umips_load_store_pair_p): Likewise.
27026 (umips_output_load_store_pair): Likewise.
27027 (umips_movep_target_p): Likewise.
27028 (umips_12bit_offset_address_p): Likewise.
27029 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27030 (mips_base_mips16): Rename this...
27031 (mips_base_compression_flags): ...to this. Update all uses.
27032 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27033 (mips_mips16_decl_p): Delete.
27034 (mips_nomips16_decl_p): Delete.
27035 (mips_get_compress_on_flags): New function.
27036 (mips_get_compress_off_flags): New function.
27037 (mips_get_compress_mode): New function.
27038 (mips_get_compress_on_name): New function.
27039 (mips_get_compress_off_name): New function.
27040 (mips_insert_attributes): Support multiple compression types.
27041 (mips_merge_decl_attributes): Likewise.
27042 (umips_12bit_offset_address_p): New function.
27043 (mips_start_function_definition): Emit .set micromips directive.
27044 (mips_call_may_need_jalx_p): New function.
27045 (mips_function_ok_for_sibcall): Add microMIPS support.
27046 (mips_print_operand_punctuation): Support short delay slots and
27048 (umips_swm_mask, umips_swm_encoding): New.
27049 (umips_build_save_restore): New function.
27050 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27051 (was_mips16_p): Remove.
27052 (old_compression_mode): New.
27053 (mips_set_compression_mode): New function.
27054 (mips_set_current_function): Add microMIPS support.
27055 (mips_option_override): Likewise.
27056 (umips_save_restore_pattern_p): New function.
27057 (umips_output_save_restore): New function.
27058 (umips_load_store_pair_p_1): New function.
27059 (umips_load_store_pair_p): New function.
27060 (umips_output_load_store_pair_1): New function.
27061 (umips_output_load_store_pair): New function.
27062 (umips_movep_target_p) New function.
27063 (mips_prepare_pch_save): Add microMIPS support.
27064 * config/mips/mips.h (TARGET_COMPRESSION): New.
27065 (TARGET_CPU_CPP_BUILTINS): Update macro
27066 to use new compression flags and to support microMIPS.
27067 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27068 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27069 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27070 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27071 (ASM_SPEC): Support mmicromips and mno-micromips.
27072 (M16STORE_REG_P): New macro.
27073 (MIPS_CALL): Support TARGET_MICROMIPS.
27074 (MICROMIPS_J): New macro.
27075 (mips_base_mips16): Rename this...
27076 (mips_base_compression_flags): ...to this.
27077 (UMIPS_12BIT_OFFSET_P): New macro.
27078 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27079 (MULTILIB_DIRNAMES): Likewise.
27081 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27084 * target-globals.c (save_target_globals): For init_reg_sets and
27085 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27087 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27089 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27090 * config/aarch64/t-aarch64
27091 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27093 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27095 PR inline-asm/56148
27096 * lra-constraints.c (process_alt_operands): Reload operand
27097 conflicting with earlier clobber only if no more other conflicting
27100 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27103 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27104 if not linking a shared library.
27106 2013-02-22 Seth LaForge <sethml@google.com>
27108 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27110 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27112 * config/arm/arm.md (split for extendsidi): Update condition.
27113 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27114 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27115 (qhs_zextenddi_cstr): Likewise.
27117 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27119 PR middle-end/56420
27120 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27121 avoid signed wrapping.
27122 (expand_mult): Handle properly multiplication by
27123 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27124 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27125 in the compiler if coeff is HOST_WIDE_INT_MIN.
27126 (expand_divmod): Don't make ext_op1 static, change it's type to
27127 uhwi. Avoid undefined behavior in -INTVAL (op1).
27129 PR rtl-optimization/50339
27130 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27132 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27133 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27134 into splitting_ashiftrt field.
27135 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27137 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27140 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27142 PR middle-end/56108
27143 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27144 are sure to go irrevocable.
27146 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27148 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27149 scalars are valid operands.
27151 2013-02-21 Martin Jambor <mjambor@suse.cz>
27153 PR tree-optimization/56310
27154 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27155 only matching indices and non-negative final offsets.
27156 (intersect_aggregates_with_edge): Pass src_idx to
27157 agg_replacements_to_vector. Pass src_idx insstead of index to
27158 intersect_with_agg_replacements.
27160 2013-02-21 Martin Jambor <mjambor@suse.cz>
27162 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
27163 instead of hard-wired defaults.
27165 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
27167 * doc/invoke.texi (MIPS Options): Update documentation of the
27168 floating-point multiply-accumulate instruction restrictions.
27170 2013-02-21 Kostya Serebryany <kcc@google.com>
27172 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
27173 asan_shadow_offset on x86_64 linux.
27175 2013-02-21 Richard Biener <rguenther@suse.de>
27177 PR tree-optimization/56415
27179 2013-02-11 Richard Biener <rguenther@suse.de>
27181 PR tree-optimization/56273
27182 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27185 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27188 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
27191 PR inline-asm/56405
27192 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
27193 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
27195 2013-02-20 Jan Hubicka <jh@suse.cz>
27197 PR tree-optimization/56265
27198 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
27199 when target is referenced for first time.
27201 2013-02-20 Richard Biener <rguenther@suse.de>
27203 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
27204 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
27205 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
27206 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
27207 not return anything.
27208 (rename_ssa_copies): Do not remove unused locals.
27209 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27210 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27211 * passes.c (execute_function_todo): Do not schedule unused locals
27212 removal if cleanup_tree_cfg did something.
27213 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27214 about the number of removed locals.
27216 2013-02-20 Richard Biener <rguenther@suse.de>
27218 PR tree-optimization/56398
27219 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27221 2013-02-20 Martin Jambor <mjambor@suse.cz>
27223 PR tree-optimization/55334
27224 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27225 restricted pointers to arrays.
27227 2013-02-20 Richard Biener <rguenther@suse.de>
27228 Jakub Jelinek <jakub@redhat.com>
27230 PR tree-optimization/56396
27231 * tree-ssa-ccp.c (n_const_val): New static variable.
27232 (get_value): Return NULL for SSA names we don't have a lattice
27234 (ccp_initialize): Initialize n_const_val.
27235 * tree-ssa-copy.c (n_copy_of): New static variable.
27236 (init_copy_prop): Initialize n_copy_of.
27237 (get_value): Return NULL_TREE for SSA names we don't have a
27240 2013-02-20 Martin Jambor <mjambor@suse.cz>
27242 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27244 2013-02-20 Richard Biener <rguenther@suse.de>
27246 * genpreds.c (write_lookup_constraint): Do not compare first
27247 letter of the constraint again.
27249 2013-02-20 Richard Biener <rguenther@suse.de>
27251 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27253 (get_use_iv_cost): Terminate hashtable walk when coming across
27256 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27258 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27259 reassociation for avx2 targets.
27261 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27263 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27264 Add version check for v8.10.a to enable microblaze_has_clz
27265 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27266 version and TARGET_PATTERN_COMPARE check
27267 * config/microblaze/microblaze.md: New clzsi2 instruction
27269 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27271 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27272 function before branching.
27274 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27276 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27278 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27280 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27282 PR middle-end/55889
27283 * sel-sched.c: Include ira.h.
27284 (implicit_clobber_conflict_p): New function.
27285 (moveup_expr): Use it.
27286 * Makefile.in (sel-sched.o): Depend on ira.h.
27288 2013-02-19 Richard Biener <rguenther@suse.de>
27290 PR tree-optimization/56384
27291 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27292 (vn_hash_type): Split out from ...
27293 (vn_hash_constant_with_type): ... here.
27294 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27295 (vn_phi_eq): Compare types from vn_phi_s structure.
27296 (vn_phi_lookup): Populate vn_phi_s type.
27297 (vn_phi_insert): Likewise.
27299 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27301 PR tree-optimization/56350
27302 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27303 if haven't found reduction or nested cycle operand, rather than
27304 asserting we must find it.
27306 PR tree-optimization/56381
27307 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27310 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27311 Jakub Jelinek <jakub@redhat.com>
27314 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27315 (swap_optab_enable): Same.
27316 (init_all_optabs): Use argument instead of global.
27317 * tree.h (struct tree_optimization_option): New field target_optabs.
27318 * expr.h (init_all_optabs): Add argument to prototype.
27319 (TREE_OPTIMIZATION_OPTABS): New.
27320 (save_optabs_if_changed): Protoize.
27321 * optabs.h: Declare this_fn_optabs.
27322 * optabs.c (save_optabs_if_changed): New.
27323 Declare this_fn_optabs.
27324 (init_optabs): Add argument to init_all_optabs() call.
27325 * function.c (invoke_set_current_function_hook): Handle per
27327 * function.h (struct function): New field optabs.
27328 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27329 optimization_current_node has changed.
27330 * target-globals.h (save_target_globals_default_opts): Protoize.
27331 * target-globals.c (save_target_globals_default_opts): New.
27333 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27336 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27337 registers %fr12 and %fr12R as call used.
27340 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27341 and HImode, require all displacements to be an integer multiple of
27343 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27344 only allow QImode and HImode when reload is in progress and strict is
27345 true. Likewise for symbolic addresses. Use base14_operand to check
27346 displacements in REG+BASE addresses.
27348 2013-02-18 Richard Biener <rguenther@suse.de>
27350 PR tree-optimization/56366
27351 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27352 handle sign-conversion of outer-loop initial induction value.
27354 2013-02-18 Richard Biener <rguenther@suse.de>
27356 PR middle-end/56349
27357 * cfghooks.c (merge_blocks): If we merge a latch into another
27358 block adjust references to it.
27359 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27360 (verify_loop_structure): Verify that a recorded latch is in fact
27363 2013-02-18 Richard Biener <rguenther@suse.de>
27365 PR tree-optimization/56321
27366 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27367 order SSA name release and virtual operand unlinking.
27369 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27371 * config/microblaze/microblaze.md (save_stack_block): Define.
27372 (restore_stack_block): Likewise.
27374 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27376 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27377 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27378 * config/microblaze/microblaze.c (microblaze_option_override):
27379 Bail out early for PIC modes when target does not support PIC.
27381 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27383 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27384 Replace with a microblaze version.
27385 (microblaze_trampoline_init): Adapt for microblaze.
27386 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27389 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27390 Dodji Seketeli <dodji@redhat.com>
27393 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27394 (instrument_mem_region_access): Do not forget to always put
27395 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27396 0) statement, even for cases where either 'base' or 'base + len'
27397 are not instrumented -- because they have been previously
27398 instrumented. Simplify the logic by putting all the statements
27399 instrument 'base + len' inside a sequence, and then insert that
27400 sequence right before the current insertion point. Then, to
27401 instrument 'base + len', just get an iterator on that statement.
27402 And do not forget to update the pointer to iterator the function
27403 received as argument.
27405 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27407 PR rtl-optimization/56348
27408 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27410 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27412 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27413 (clean_graph_dump_file): Pass base to start_graph_dump.
27415 2013-02-14 Richard Henderson <rth@redhat.com>
27418 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27420 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27422 * collect2-aix.h: Define F_LOADONLY.
27424 2013-02-14 Richard Biener <rguenther@suse.de>
27427 * varasm.c (output_constant_def_1): Get the decl representing
27428 the constant as argument.
27429 (output_constant_def): Wrap output_constant_def_1.
27430 (make_decl_rtl): Use output_constant_def_1 with the decl
27431 representing the constant.
27432 (build_constant_desc): Optionally re-use a decl already
27433 representing the constant.
27434 (tree_output_constant_def): Adjust.
27436 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27439 * asan.c (instrument_builtin_call): Really put the length of the
27440 second source argument into src1_len.
27442 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27444 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27445 argument. If it is false, don't create edge from then_bb to
27447 (insert_if_then_before_iter): Pass true to it.
27448 (build_check_stmt): Pass false to it.
27449 (transform_statements): Flush hash table only on extended basic
27450 block boundaries, rather than at the beginning of every bb.
27451 Don't flush hash table on nonfreeing_call_p calls.
27452 * tree-flow.h (nonfreeing_call_p): New prototype.
27453 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27455 2013-02-13 David S. Miller <davem@davemloft.net>
27457 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27459 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27462 * ira.c (max_regno_before_ira): Move from ...
27464 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27467 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27469 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27471 2013-02-13 Richard Biener <rguenther@suse.de>
27474 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27475 globals in MEM_REFs.
27477 2013-02-13 Richard Biener <rguenther@suse.de>
27479 * loop-init.c (loop_optimizer_init): Clear loop state when
27480 re-initializing preserved loops.
27481 * loop-unswitch.c (unswitch_single_loop): Return whether
27482 we unswitched the loop. Do not verify loop state here.
27483 (unswitch_loops): When we unswitched a loop discover new loops.
27485 2013-02-13 Kostya Serebryany <kcc@google.com>
27487 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27489 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27491 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27493 Avoid instrumenting duplicated memory access in the same basic block
27494 * Makefile.in (asan.o): Add new dependency on hash-table.h
27495 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27496 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27497 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27498 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27499 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27500 (get_mem_ref_of_assignment): New functions.
27501 (get_mem_refs_of_builtin_call): Extract from
27502 instrument_builtin_call and tweak a little bit to make it fit with
27504 (instrument_builtin_call): Use the new
27505 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27506 of is_gimple_builtin_call.
27507 (instrument_derefs, instrument_mem_region_access): Insert the
27508 instrumented memory reference into the hash table.
27509 (maybe_instrument_assignment): Renamed instrument_assignment into
27510 this, and change it to advance the iterator when instrumentation
27511 actually happened and return true in that case. This makes it
27512 homogeneous with maybe_instrument_assignment, and thus give a
27513 chance to callers to be more 'regular'.
27514 (transform_statements): Clear the memory reference hash table
27515 whenever we enter a new BB, when we cross a function call, or when
27516 we are done transforming statements. Use
27517 maybe_instrument_assignment instead of instrumentation. No more
27518 need to special case maybe_instrument_assignment and advance the
27519 iterator after calling it; it's now handled just like
27520 maybe_instrument_call. Update comment.
27522 2013-02-13 Richard Biener <rguenther@suse.de>
27524 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27525 Fix loop discovery code.
27527 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27529 PR inline-asm/56148
27530 * lra-constraints.c (process_alt_operands): Match early clobber
27531 operand with itself. Check conflicts with earlyclobber only if
27532 the operand is not reloaded. Prefer to reload conflicting operand
27533 if earlyclobber and matching operands are the same.
27535 2013-02-12 Richard Biener <rguenther@suse.de>
27538 * lto-streamer-out.c (write_symbol): Do not output symbols
27539 for hard register variables.
27541 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27544 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27545 (umulsidi3_insn, mulsidi3_insn): New insns.
27547 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27549 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27550 (struct tune_params): Add vec_costs field.
27551 * config/arm/arm.c (arm_builtin_vectorization_cost)
27552 (arm_add_stmt_cost): New functions.
27553 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27554 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27555 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27556 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27557 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27558 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27559 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27561 2013-02-12 Richard Biener <rguenther@suse.de>
27564 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27565 decls again if possible.
27567 2013-02-12 Richard Biener <rguenther@suse.de>
27569 PR middle-end/56288
27570 * tree-ssa.c (verify_ssa_name): Fix check, move
27571 SSA_NAME_IN_FREE_LIST check up.
27573 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27574 Steven Bosscher <steven@gcc.gnu.org>
27576 PR rtl-optimization/56151
27577 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27578 equal to op0 or op1, and last_insn pattern is CODE operation
27579 with MEM dest and one of the operands matches that MEM.
27581 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27583 * doc/extend.texi: Document Function Multiversioning and "default"
27584 parameter string to target attribute.
27585 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27586 target attribute parameter is "default".
27587 (ix86_compare_version_priority): Remove checks for target attribute.
27588 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27589 Remove check for target attribute equal to NULL. Add assert.
27590 (ix86_generate_version_dispatcher_body): Change error to sorry.
27592 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27593 Jack Howarth <howarth@bromo.med.uc.edu>
27594 Patrick Marlier <patrick.marlier@gmail.com>
27597 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27598 define ENDFILE_SPEC as TM_DESTRUCTOR.
27599 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27601 2013-02-11 Alexander Potapenko <glider@google.com>
27602 Jack Howarth <howarth@bromo.med.uc.edu>
27603 Jakub Jelinek <jakub@redhat.com>
27606 * config/darwin.c (cdtor_record): Rename ctor_record.
27607 (sort_cdtor_records): Rename sort_ctor_records.
27608 (finalize_dtors): New routine to sort destructors by
27609 priority before use in assemble_integer.
27610 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27612 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27614 PR rtl-optimization/56275
27615 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27616 offset is non-negative and less than cmode size before
27617 calling simplify_subreg.
27619 2013-02-11 Richard Biener <rguenther@suse.de>
27621 PR tree-optimization/56264
27622 * cfgloop.h (fix_loop_structure): Adjust prototype.
27623 * loop-init.c (fix_loop_structure): Return the number of
27624 newly discovered loops.
27625 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27626 are discovered, do a full loop-closed SSA rewrite.
27628 2013-02-11 Richard Biener <rguenther@suse.de>
27630 PR tree-optimization/56273
27631 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27633 (check_array_ref): Fix missing newline in dumps.
27634 (search_for_addr_array): Likewise.
27636 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
27638 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
27640 2013-02-09 Jakub Jelinek <jakub@redhat.com>
27643 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
27645 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
27647 PR rtl-optimization/56246
27648 * lra-constraints.c (simplify_operand_subreg): Try to reuse
27650 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
27651 constraints are satisfied.
27653 2013-02-08 Jeff Law <law@redhat.com>
27656 * emit-rtl.c (reg_is_parm_p): New function.
27657 * regs.h (reg_is_parm_p): New prototype.
27658 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
27659 callee-clobbered registers.
27661 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
27664 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
27665 If there is no implicit builtin declaration, just return NULL.
27667 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27669 * config/i386/sse.md (FMAMODEM): New mode iterator.
27670 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
27671 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
27673 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27675 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
27676 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
27677 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
27679 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27681 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
27682 (microblaze*-*-elf): Likewise.
27683 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
27685 * config/microblaze/microblaze-c.c: Add builtin defines for
27686 _LITTLE_ENDIAN and _BIG_ENDIAN.
27687 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
27688 add to TARGET_DEFAULT flags.
27689 Expand ASM_SPEC and LINK_SPEC.
27690 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
27691 * config/microblaze/microblaze.md: Update extendsidi2 and
27692 movdi_internal instructions to use low-order / high-order reg
27694 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
27695 options and inversemask / mask of LITTLE_ENDIAN.
27696 * config/microblaze/t-microblaze: Expand multilib options to
27697 include mlittle-endian (le) and update exceptions patterns.
27699 2013-02-08 Jakub Jelinek <jakub@redhat.com>
27701 PR rtl-optimization/56195
27702 * lra-constraints.c (get_reload_reg): Don't reuse regs
27703 if they have smaller mode than requested, if they have
27704 wider mode than requested, try to return a SUBREG.
27706 PR tree-optimization/56250
27707 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
27708 if type is unsigned and code isn't MULT_EXPR.
27710 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27712 PR tree-optimization/56064
27713 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
27714 bits according to mode.
27715 * fixed-value.h (fixed_from_double_int)
27716 (const_fixed_from_double_int): Adjust comments.
27718 2013-02-08 Richard Biener <rguenther@suse.de>
27721 * lto-streamer.h (struct data_in): Remove current_file, current_line
27722 and current_col members.
27723 * lto-streamer-out.c (lto_output_location): Stream changed bits
27724 en-block for efficiency.
27725 * lto-streamer-in.c (clear_line_info): Remove.
27726 (lto_input_location): Cache current file, line and column
27727 globally via local statics. Read changed bits en-block.
27728 (input_function): Do not call clear_line_info.
27729 (lto_read_body): Likewise.
27730 (lto_input_toplevel_asms): Likewise.
27732 2013-02-08 Michael Matz <matz@suse.de>
27734 PR tree-optimization/52448
27735 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
27736 (nt_call_phase): New static.
27737 (add_or_mark_expr): Only mark accesses with newer phase than any
27739 (nonfreeing_call_p): New.
27740 (nt_init_block): Update nt_call_phase, mark blocks as visited.
27741 (nt_fini_block): Keep blocks marked as visited.
27742 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
27744 2013-02-08 Richard Biener <rguenther@suse.de>
27746 * ira.c (ira): Free broken dominator information.
27748 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27750 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
27752 2013-02-08 Marek Polacek <polacek@redhat.com>
27754 * cfgloop.c (verify_loop_structure): Add more checking of headers.
27756 2013-02-08 Richard Biener <rguenther@suse.de>
27758 PR middle-end/56181
27759 * cfgloop.h (flow_loops_find): Adjust.
27760 (bb_loop_header_p): Declare.
27761 * cfgloop.c (bb_loop_header_p): New function split out from ...
27762 (flow_loops_find): ... here. Adjust function signature,
27763 support incremental loop structure update.
27764 (verify_loop_structure): Cleanup. Verify a loop is a loop.
27765 * cfgloopmanip.c (fix_loop_structure): Move ...
27766 * loop-init.c (fix_loop_structure): ... here.
27767 (apply_loop_flags): Split out from ...
27768 (loop_optimizer_init): ... here.
27769 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
27770 in incremental mode, only remove dead loops here.
27772 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27775 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
27776 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
27777 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
27778 (*round<mode>3.libgcc): New insns for fixed-modes.
27779 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
27780 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
27781 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
27782 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
27783 implementations. Define to __builtin_avr_absFX,
27784 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
27785 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
27786 __builtin_avr_countlsFX, respectively.
27787 * config/avr/avr-c.c (target.h): Include it.
27788 (enum avr_builtin_id): New enum.
27789 (avr_resolve_overloaded_builtin): New static function.
27790 (avr_register_target_pragmas): Use it to set
27791 targetm.resolve_overloaded_builtin.
27792 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
27793 tree nodes used by DEF_BUILTIN.
27794 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
27795 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
27796 <AVR_BUILTIN_xxBITS>: Same.
27798 2013-02-08 Richard Biener <rguenther@suse.de>
27800 * cfgloop.c (verify_loop_structure): Properly handle
27801 a loop exiting to another loop header.
27802 * ira-int.h (ira_loops): Remove.
27803 * ira.c (ira_loops): Remove.
27804 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
27805 (do_reload): Use loop_optimizer_finalize.
27806 * ira-build.c (create_loop_tree_nodes): Use get_loops and
27807 number_of_loops to access the loop tree.
27808 (more_one_region_p): Likewise.
27809 (finish_loop_tree_nodes): Likewise.
27810 (rebuild_regno_allocno_maps): Likewise.
27811 (mark_loops_for_removal): Likewise.
27812 (mark_all_loops_for_removal): Likewise.
27813 (remove_unnecessary_regions): Likewise.
27814 (ira_build): Likewise.
27815 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
27817 2013-02-08 Richard Biener <rguenther@suse.de>
27819 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27820 * ipa-pure-const.c (analyze_function): Avoid calling
27821 mark_irreducible_loops twice.
27822 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27824 2013-02-07 David S. Miller <davem@davemloft.net>
27826 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27828 * var-tracking.c (vt_add_function_parameter): Test the presence of
27829 HAVE_window_save properly and do not remap argument registers when
27830 we have a leaf function.
27832 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27835 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27837 * config/i386/i386.c (ix86_print_operand): Ditto.
27839 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27841 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27843 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27845 PR rtl-optimization/56225
27846 * lra-constraints.c (process_alt_operands): Check that reload hard
27847 reg can hold value for strict_low_part.
27849 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27852 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
27853 dwarf2out_end_function.
27854 (in_first_function_p, maybe_at_text_label_p,
27855 first_loclabel_num_not_at_text_label): New variables.
27856 (dwarf2out_var_location): In the first function find out
27857 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
27858 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
27861 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
27863 PR rtl-optimization/56178
27864 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
27865 SUBREG of a register. Tidy up related block of code.
27866 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
27867 note if the source is a register or a SUBREG of a register.
27869 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27872 * config/rs6000/rs6000.md (ptrm): New mode attr.
27873 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
27874 call_value_indirect_aix<pttrsize>,
27875 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
27878 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27880 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
27881 if -bnortl. Convert to strcmp and strncmp.
27883 2013-02-07 Alan Modra <amodra@gmail.com>
27886 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
27887 addresses won't wrap when offsetting.
27888 (rs6000_secondary_reload): Provide secondary reloads needed for
27889 wrapping LO_SUM addresses.
27891 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
27893 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
27894 MACH, just __MACH__.
27896 2013-02-06 Richard Biener <rguenther@suse.de>
27898 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
27899 instead of calling fix_loop_structure.
27901 2013-02-06 Jakub Jelinek <jakub@redhat.com>
27903 PR middle-end/56217
27904 * omp-low.c (use_pointer_for_field): Return false if
27905 lower_send_shared_vars doesn't generate any copy-out code.
27907 2013-02-06 Tom de Vries <tom@codesourcery.com>
27909 PR rtl-optimization/56131
27910 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
27911 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
27912 of the label is NULL. Add comment.
27914 2013-02-05 Jakub Jelinek <jakub@redhat.com>
27916 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
27919 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
27920 (STATIC_LIBTSAN_LIBS): Likewise.
27921 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
27922 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
27923 is defined, don't add anything else beyond that.
27924 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
27925 (LINK_COMMAND_SPEC): Use them.
27927 PR tree-optimization/56205
27928 * tree-stdarg.c (check_all_va_list_escapes): Return true if
27929 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
27930 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
27932 2013-02-05 Richard Biener <rguenther@suse.de>
27934 PR tree-optimization/53342
27935 PR tree-optimization/53185
27936 * tree-vectorizer.h (vect_check_strided_load): Remove.
27937 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
27938 not disallow peeling for vectorized strided loads.
27939 (vect_check_strided_load): Make static and simplify.
27940 (vect_analyze_data_refs): Adjust.
27941 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
27942 correctly when vectorizing strided loads.
27944 2013-02-05 Richard Biener <rguenther@suse.de>
27946 * doc/install.texi: Refer to ISL, not PPL.
27948 2013-02-05 Jan Hubicka <jh@suse.cz>
27950 PR tree-optimization/55789
27951 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
27953 2013-02-05 Jan Hubicka <jh@suse.cz>
27955 PR tree-optimization/55789
27956 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
27957 the dead call anyway.
27959 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
27962 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
27964 2013-02-04 Alexander Potapenko <glider@google.com>
27965 Jack Howarth <howarth@bromo.med.uc.edu>
27966 Jakub Jelinek <jakub@redhat.com>
27969 * config/darwin.c (sort_ctor_records): Stabilized qsort
27970 on constructor priority by using original position.
27971 (finalize_ctors): New routine to sort constructors by
27972 priority before use in assemble_integer.
27973 (machopic_asm_out_constructor): Use finalize_ctors if needed.
27975 2013-02-04 Jakub Jelinek <jakub@redhat.com>
27978 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
27979 about visibility on artificial decls.
27980 * config/sol2.c (solaris_assemble_visibility): Likewise.
27982 2013-02-04 Kai Tietz <ktietz@redhat.com>
27985 * config/i386/i386.c (function_value_ms_64): Add additional valtype
27986 argument and improve checking of return-argument types for 16-byte
27988 (ix86_function_value_1): Add additional valtype argument on call
27989 of function_value_64.
27990 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
27991 handling infunction_value_64 function.
27993 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
27995 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
27997 2013-02-04 Richard Biener <rguenther@suse.de>
27999 PR tree-optimization/56188
28000 * tree-ssa-structalias.c (label_visit): Consider case with
28001 initially non-empty points-to set.
28002 (perform_var_substitution): Dump node mapping and clean up.
28004 2013-02-04 Richard Guenther <rguenther@suse.de>
28007 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
28008 node prevail as last resort.
28009 (lto_symtab_merge_decls): Remove guard on LTRANS here.
28010 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
28012 2013-02-04 Richard Biener <rguenther@suse.de>
28014 PR tree-optimization/56113
28015 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
28017 (equiv_class_lookup_or_add): ... this.
28018 (label_visit): Adjust and fix error in previous patch.
28019 (perform_var_substitution): Adjust.
28021 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28023 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28024 * config/sh/sh4-300.md: Likewise.
28025 * config/sh/sh4a.md: Likewise.
28026 * config/sh/constraints.md: Likewise.
28027 * config/sh/sh.md: Likewise.
28028 * config/sh/netbsd-elf.h: Likewise.
28029 * config/sh/predicates.md: Likewise.
28030 * config/sh/sh-protos.h: Likewise.
28031 * config/sh/ushmedia.h: Likewise.
28032 * config/sh/linux.h: Likewise.
28033 * config/sh/sh.c: Likewise.
28034 * config/sh/superh.h: Likewise.
28035 * config/sh/elf.h: Likewise.
28036 * config/sh/sh4.md: Likewise.
28037 * config/sh/sh.h: Likewise.
28039 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28041 * config/pa/constraints.md: Adjust unused letters. Change "T"
28042 constraint to match_test floating_point_store_memory_operand().
28043 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28044 (base14_operand): New.
28045 (floating_point_store_memory_operand): New.
28046 (integer_store_memory_operand): Revise to use base14_operand and
28047 reg_plus_base_memory_operand.
28048 (move_dest_operand): Allow symbolic_memory_operands.
28049 (symbolic_memory_operand): Check for LO_SOM.
28050 (symbolic_operand): Change default case to break.
28051 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28052 CONST_DOUBLE values to be reloaded by putting them into memory when
28053 the destination is a floating point register.
28054 (movdf): Remove code to handle CONST_DOUBLE.
28056 (reload_indf_r1): New.
28057 (reload_insf_r1): New.
28058 Consistently use "Q" and "T" constraints with integer and floating
28059 point move instructions, respectively.
28060 (movdi): Remove FAIL.
28061 Change predicate for source operand unamed DImode move from
28062 general_operand to move_src_operand.
28063 (umulsidi3): Change predicate for destination operand to
28065 Likewise for similar unamed patterns.
28066 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28067 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28068 (hppa_legitimize_address): Simplify mask calculation.
28069 (pa_emit_move_sequence): Revised handling of secondary reloads from
28070 REG+D addresses for floating point loads and stores. Directly handle
28071 loading CONST0_RTX (mode) to a floating point register.
28072 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28073 to floating point registers. Don't restrict secondary reloads to
28074 floating point registers to integer modes. Revise some comments and
28076 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28077 (pa_legitimate_address_p): New.
28078 (pa_legitimize_reload_address): New.
28079 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28080 (STRICT_REG_OK_FOR_BASE_P): New.
28081 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28082 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28084 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28085 Andrew Dixie <andrewd@gentrack.com>
28087 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28090 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28092 * expmed.c (extract_bit_field_1): Pass the full width of the
28093 structure to get_best_reg_extraction_insn.
28095 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28098 * configure.ac (use_cxa_atexit): Add AIX.
28099 * configure: Regenerate.
28101 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28103 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28106 * final.c (need_profile_function): New variable.
28107 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28108 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28109 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28110 notes, targetm.asm_out.function_prologue doesn't emit anything,
28111 HAVE_prologue and profiler should be emitted before prologue,
28112 set need_profile_function instead of emitting it.
28113 (final_scan_insn): If need_profile_function, emit
28114 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28115 NOTE_INSN_FUNCTION_BEG note.
28117 2013-02-01 Richard Henderson <rth@redhat.com>
28119 * config/rs6000/rs6000.md (smulditi3): New.
28122 * config/alpha/alpha.md (umulditi3): New.
28124 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28126 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28127 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28129 2013-02-01 Richard Biener <rguenther@suse.de>
28131 PR tree-optimization/56113
28132 * tree-ssa-structalias.c (label_visit): Reduce work for
28133 single-predecessor nodes.
28135 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28137 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28138 range isn't testing for zero.
28140 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28142 PR middle-end/56113
28143 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28145 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28146 Nick Clifton <nickc@redhat.com>
28148 * config/v850/constraints.md (Q): Define as a memory constraint.
28149 * config/v850/predicates.md (label_ref_operand): New predicate.
28150 (e3v5_shift_operand): New predicate.
28151 (ior_operator): New predicate.
28152 * config/v850/t-v850: Add e3v5 multilib.
28153 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28154 (v850_gen_movdi): Prototype.
28155 * config/v850/v850.c: Add support for e3v5 architecture.
28156 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28158 (construct_save_jarl): Add e3v5 long JARL support.
28159 (v850_adjust_insn_length): New function. Adjust length of call
28160 insns when using e3v5 instructions.
28161 (v850_gen_movdi): New function: Generate instructions to move a
28163 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
28164 (CPP_SPEC): Define __v850e3v5__ as appropriate.
28165 (TARGET_USE_FPU): Enable for e3v5.
28166 (CONST_OK_FOR_W): New macro.
28167 (ADJUST_INSN_LENGTH): Define.
28168 * config/v850/v850.md (UNSPEC_LOOP): Define.
28169 (attr cpu): Add v850e3v5.
28170 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
28171 (movdi): New pattern.
28172 (movdi_internal): New pattern.
28173 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
28174 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
28175 (cstoresf4): Likewise.
28176 (cstoredf4): Likewise.
28177 (insv): New pattern.
28178 (rotlso3_a): New pattern.
28179 (rotlsi3_b): New pattern
28180 (rotlsi3_v850e3v5): New pattern.
28181 (doloop_begin): New pattern.
28182 (fix_loop_counter): New pattern.
28183 (doloop_end): New pattern.
28184 (branch_normal): Add e3v5 long branch support.
28185 (branch_invert): Likewise.
28186 (branch_z_normal): Likewise.
28187 (branch_z_invert): Likewise.
28188 (branch_nz_normal): Likewise.
28189 (branch_nz_invert): Likewise.
28190 (call_internal_short): Add e3v5 register-indirect JARL support.
28191 (call_internal_long): Likewise.
28192 (call_value_internal_short): Likewise.
28193 (call_value_internal_long): Likewise.
28194 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
28195 (mloop): New option.
28196 * config.gcc: Add support for configuring v840e3v5 target.
28197 * doc/invoke.texi: Document new v850 specific command line options.
28199 2013-01-31 Paul Koning <ni1d@arrl.net>
28203 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
28204 children if parent is a class.
28205 (prune_unused_types_prune): Don't add DW_AT_declaration.
28207 2013-01-31 Richard Biener <rguenther@suse.de>
28209 PR tree-optimization/56157
28210 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28211 match up operand with SLP child.
28213 2013-01-31 Jason Merrill <jason@redhat.com>
28216 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28217 parameters the first time.
28218 (gen_scheduled_generic_parms_dies): Check completeness here.
28220 2013-01-31 Richard Biener <rguenther@suse.de>
28222 PR middle-end/53073
28223 * common.opt (faggressive-loop-optimizations): New flag,
28224 enabled by default.
28225 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28226 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28227 infer_loop_bounds_from_undefined by it.
28229 2013-01-31 Richard Biener <rguenther@suse.de>
28231 PR tree-optimization/56150
28232 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28233 visit virtual operands.
28234 (find_uses_to_rename_bb): Likewise.
28236 2013-01-31 Richard Biener <rguenther@suse.de>
28238 PR tree-optimization/56150
28239 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28240 mixed store non-store stmts.
28242 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28245 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28246 LIBASAN_EARLY_SPEC is defined.
28247 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28248 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28250 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28253 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28254 invalid args instead of ICEing on it.
28255 (ix86_valid_target_attribute_tree): Return error_mark_node if
28256 ix86_valid_target_attribute_inner_p failed.
28257 (ix86_valid_target_attribute_p): Return false only if
28258 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28259 target("default") attribute.
28260 (sorted_attr_string): Change argument from const char * to tree,
28261 merge in all target attribute arguments rather than just one.
28262 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28263 instead of free. Avoid using strcat.
28264 (ix86_mangle_function_version_assembler_name): Mangle
28265 target("default") as if no target attribute is present. Adjust
28266 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28267 instead of xmalloc and XDELETEVEC instead of free.
28268 (ix86_function_versions): Don't return true if one of the decls
28269 doesn't have target attribute. If they don't and one of the decls
28270 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28271 sorted_attr_string caller. Use XDELETEVEC instead of free.
28272 (ix86_supports_function_versions): Remove.
28273 (make_name): Fix up formatting.
28274 (make_dispatcher_decl): Remove resolver_name and its initialization.
28275 Avoid leaking memory.
28276 (is_function_default_version): Return true if there is
28277 target("default") attribute rather than no target attribute at all.
28278 (make_resolver_func): Avoid leaking memory.
28279 (ix86_generate_version_dispatcher_body): Likewise.
28280 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28281 * target.def (supports_function_versions): Remove.
28282 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28283 * doc/tm.texi: Regenerated.
28285 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28287 PR rtl-optimization/56144
28288 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28289 for values with side effects.
28291 2013-01-30 Richard Biener <rguenther@suse.de>
28293 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28294 (sparseset_pop): Likewise.
28295 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28296 to be able to use quick_push in the worker loop.
28298 2013-01-30 Marek Polacek <polacek@redhat.com>
28300 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28302 2013-01-30 Richard Biener <rguenther@suse.de>
28305 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28307 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28309 PR tree-optimization/56064
28310 * fixed-value.c (fixed_from_double_int): New function.
28311 * fixed-value.h (fixed_from_double_int): New prototype.
28312 (const_fixed_from_double_int): New static inline function.
28313 * fold-const.c (native_interpret_fixed): New static function.
28314 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28315 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28316 (native_encode_fixed): New static function.
28317 (native_encode_expr) <FIXED_CST>: Use it.
28318 (native_interpret_int): Move double_int worker code to...
28319 * double-int.c (double_int::from_buffer): ...this new static method.
28320 * double-int.h (double_int::from_buffer): Prototype it.
28322 2013-01-30 Richard Biener <rguenther@suse.de>
28324 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28325 New pointer-map and obstack.
28326 (init_alias_vars): Allocate pointer-map and obstack.
28327 (delete_points_to_sets): Free them.
28328 (find_what_var_points_to): Cache result.
28329 (find_what_p_points_to): Adjust for changed interface of
28330 find_what_var_points_to.
28331 (compute_points_to_sets): Likewise.
28332 (ipa_pta_execute): Likewise.
28334 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28336 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28337 * configure: Regenerate.
28338 * config.in: Regenerate.
28339 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28340 #nobits/#progbits if supported.
28342 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28345 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28346 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28347 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28349 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28351 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28352 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28354 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28356 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28357 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28359 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28361 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28363 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28364 * config/arm/cortex-a7.md: New bypasses using
28365 arm_mac_accumulator_is_result.
28367 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28369 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28370 (cortex_a7_neon_mla): Likewise.
28371 (cortex_a7_fpfmad): New reservation.
28372 (cortex_a7_fpmacs): Use ffmas and update required units.
28373 (cortex_a7_fpmuld): Update required units and latency.
28374 (cortex_a7_fpmacd): Likewise.
28375 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28376 (cortex_a7_neon). Likewise.
28377 (bypass) Update participating units.
28379 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28381 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28382 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28384 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28385 (vfp_fmul): Use ffmad.
28386 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28387 (cortex_r4_fmacd): Use ffmad.
28388 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28389 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28390 (cortex_a9_fmacd): Use ffmad.
28391 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28392 (cortex_a8_vfp_macd): Use ffmad.
28393 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28394 (cortex_a5_fpmacd): Use ffmad.
28395 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28396 (cortex_a15_vfp_macd): Use ffmad.
28397 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28399 2013-01-29 Jason Merrill <jason@redhat.com>
28402 * varasm.c (default_assemble_visibility): Don't warn about
28403 visibility on artificial decls.
28405 2013-01-29 Richard Biener <rguenther@suse.de>
28407 PR tree-optimization/56113
28408 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28410 (label_visit): Free duplicate bitmaps and record the leader instead.
28411 (perform_var_substitution): Adjust.
28413 2013-01-29 Richard Biener <rguenther@suse.de>
28415 PR tree-optimization/55270
28416 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28417 the CFG, schedule loops for fixup.
28419 2013-01-29 Nick Clifton <nickc@redhat.com>
28421 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28424 2013-01-28 Leif Ekblad <leif@rdos.net>
28426 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28427 * config/i386/i386.h (TARGET_RDOS): New macro.
28428 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28429 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28430 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28431 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28432 DEFAULT_LARGE_SECTION_THRESHOLD.
28433 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28434 * config/i386/rdos.h: New file.
28435 * config/i386/rdos64.h: New file.
28437 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28440 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28443 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28445 PR rtl-optimization/56117
28446 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28447 call cselib_lookup_from_insn on the MEM before calling
28448 add_insn_mem_dependence.
28450 2013-01-28 Richard Biener <rguenther@suse.de>
28452 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28453 to a stmt that didn't have one.
28454 (copy_phis_for_bb): Likewise for PHI arguments.
28455 (copy_debug_stmt): Likewise for debug stmts.
28457 2013-01-28 Richard Biener <rguenther@suse.de>
28459 PR tree-optimization/56034
28460 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28461 (partition_builtin_p): Adjust.
28462 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28463 it is the last partition.
28464 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28465 up the vertex for the definition.
28466 (classify_partition): Classify whether a partition is a
28467 PKIND_REDUCTION, thus has uses outside of the loop.
28468 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28469 Merge all PKIND_REDUCTION partitions into the last partition.
28470 (tree_loop_distribution): Seed partitions from reductions as well.
28472 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28474 PR tree-optimization/56125
28475 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28476 pow(x,c) into sqrt(x) * powi(x, n/2) or
28477 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28478 optimizing for size.
28479 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28480 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28483 PR tree-optimization/56094
28484 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28485 to UNKNOWN_LOCATION while gimplifying expr.
28487 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28490 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28491 operand 0 in movabs insn template for -masm=intel asm alternative.
28492 (*movabs<mode>_2): Ditto for operand 1.
28494 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28497 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28500 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28502 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28505 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28507 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28508 Undef to avoid warning.
28510 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28512 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28513 * configure: Regenerate.
28515 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28517 PR tree-optimization/56098
28518 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28519 for stmts with volatile ops.
28520 (cond_store_replacement): Don't optimize if assign has volatile ops.
28521 (cond_if_else_store_replacement_1): Don't optimize if either
28522 then_assign or else_assign have volatile ops.
28523 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28526 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28528 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28530 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28532 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28533 missing ':' in asm example.
28535 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28537 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28538 entries into lane and laneq entries.
28539 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28540 Remove AdvSIMD scalar modes.
28541 (aarch64_sq<r>dmulh_laneq<mode>): New.
28542 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28544 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28545 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28546 * config/aarch64/iterators.md (VCOND): New.
28549 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28552 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28553 Add NULL LIBNAME argument to existing definitions.
28554 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28555 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28556 * config/avr/avr.c (DEF_BUILTIN): Same.
28557 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28558 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28559 implementation is available (DECL_ASSEMBLER_NAME is set).
28560 (avr_fold_absfx): New static function.
28561 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28562 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28563 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28564 AVR_BUILTIN_ABSLLK.
28565 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28566 (abshk, absk, abslk, absllk): Provide as static inline functions.
28568 2013-01-25 Marek Polacek <polacek@redhat.com>
28570 PR tree-optimization/56035
28571 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28573 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28575 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28576 (*movtf_internal_rex64): Add (!o,C) alternative
28577 (*movxf_internal_rex64): Ditto.
28578 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28580 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28582 * doc/invoke.texi: fix typo.
28583 * doc/objc.texi: fix typo.
28585 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28587 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28588 for the first two alternatives.
28590 2013-01-24 Diego Novillo <dnovillo@google.com>
28592 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28593 (ggc-zone.o): Remove.
28594 * configure.ac: Remove option --with-gc.
28595 * configure: Re-generate.
28596 * doc/install.texi: Remove documentation for --with-gc.
28597 * gengtype.c (write_enum_defn): Remove. Update all users.
28598 (write_Types_process_field): Remove generation of gt_e_* argument.
28599 (output_type_enum): Remove. Update all users.
28600 (write_enum_defn): Remove. Update all users.
28601 (enum alloc_zone): Remove. Update all users.
28602 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28603 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28604 Update all callers.
28605 (struct ptr_data): Remove field TYPE. Update all users.
28606 (gt_pch_note_object): Remove argument TYPE. Update all users.
28607 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28609 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28610 (struct alloc_zone): Remove.
28611 (ggc_internal_alloc_zone_stat): Remove.
28612 (ggc_internal_cleared_alloc_zone_stat): Remove.
28613 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28614 (ggc_pch_count_object): Remove last argument. Update all users.
28615 (ggc_pch_alloc_object): Remove last argument. Update all users.
28616 (struct alloc_zone): Remove.
28617 * ggc-zone.c: Remove.
28618 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28619 (struct alloc_zone): Remove.
28620 (ggc_alloc_typed_stat): Remove.
28621 (ggc_alloc_typed): Remove.
28622 (ggc_splay_alloc): Remove first argument.
28623 (rtl_zone): Remove. Update all users.
28624 (tree_zone): Remove. Update all users.
28625 (tree_id_zone): Remove. Update all users.
28626 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28627 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28628 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28629 * tree-ssanames.c: Remove references to zone allocator in comments.
28631 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
28633 * config/avr/avr.c (avr_out_fract): Make register numbers that
28634 might be outside of source operand signed.
28636 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
28638 * config/i386/constraints.md (Yf): New constraint.
28639 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
28640 of f constraint to conditionaly disable x87 register preferences.
28641 (*movdf_internal): Ditto.
28642 (*movsf_internal): Ditto.
28644 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
28646 PR inline-asm/55934
28647 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
28648 that have operands with impossible constraints.
28649 Add a FIXME for a speed-up opportunity.
28650 * lra-constraints.c (process_alt_operands): Verify that a class
28651 selected from constraints on asms is valid for the operand mode.
28652 (curr_insn_transform): Remove incorrect comment.
28654 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
28656 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
28657 TOC operand is a valid symbol ref in the constant pool.
28659 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28661 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
28663 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
28666 * config/avr/stdfix.h: New file.
28667 * t-avr (stdfix-gcc.h): New rule to build it.
28668 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
28670 2013-01-23 Kostya Serebryany <kcc@google.com>
28672 * config/darwin.h: remove dependency on
28673 CoreFoundation (asan on Mac OS).
28675 2013-01-23 Jakub Jelinek <jakub@redhat.com>
28678 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
28679 instead of cmpdi_operand for first comparison operand.
28680 Don't assert that comparison operands aren't both constants.
28682 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28684 * doc/install.texi (Downloading the Source): Update references to
28685 downloading separate components.
28687 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28689 * doc/extend.texi (__int128): Improve grammar.
28691 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
28694 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
28695 alternative to (o,r).
28696 (*movdi_internal_rex64): Remove (!o,n) alternative.
28697 (DImode immediate->memory splitter): Remove.
28698 (DImode immediate->memory peephole2): Remove.
28699 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
28700 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
28701 alternative to (!o,*r).
28702 (*movtf_internal_sse): New pattern.
28703 (*movxf_internal_rex64): New pattern.
28704 (*movxf_internal): Disable for TARGET_64BIT.
28705 (*movdf_internal_rex64): Remove (!o,F) alternative.
28707 2013-01-22 Jakub Jelinek <jakub@redhat.com>
28709 PR middle-end/56074
28710 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
28711 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
28712 * tree-vect-loop-manip.c (find_loop_location): Also ignore
28713 stmt locations where LOCATION_LOCUS of the stmt location is
28714 UNKNOWN_LOCATION or BUILTINS_LOCATION.
28717 * config/i386/i386.md (UNSPEC_STOS): New.
28718 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
28719 *strsetqi_1): Add UNSPEC_STOS.
28721 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
28724 * doc/invoke.texi: Remove left over -Wsynth example.
28726 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28728 PR tree-optimization/56051
28729 * fold-const.c (fold_binary_loc): Don't fold
28730 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
28731 a narrowing conversion, or widening conversion from signed
28734 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28736 PR rtl-optimization/56023
28737 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
28738 dependent on debug instruction.
28740 2013-01-21 Martin Jambor <mjambor@suse.cz>
28742 PR middle-end/56022
28743 * function.c (allocate_struct_function): Call
28744 invoke_set_current_function_hook earlier.
28746 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28748 * reload1.c (init_reload): Only initialize reload_obstack
28749 during the first call.
28751 2013-01-21 Marek Polacek <polacek@redhat.com>
28753 * cfgloop.c (verify_loop_structure): Fix up grammar.
28755 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
28757 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
28758 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
28760 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28763 * config/arm/marvell-pj4.md: Update copyright year.
28764 Fix up use of alu to alu_reg and simple_alu_imm.
28766 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28768 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
28770 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
28773 * lra-constraints.c (curr_insn_transform): Don't reuse original
28774 insn for secondary memory move when memory mode should be different.
28776 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28778 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
28779 atomic_storedi_1): New patterns.
28781 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28783 btver2 pipeline descriptions.
28784 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
28786 * config/i386/i386.md (btver2_decode): New type attributes.
28787 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
28789 * config/i386/btver2.md: New file describing btver2 pipelines.
28791 2013-01-19 Andrew Pinski <apinski@cavium.com>
28793 PR tree-optimization/52631
28794 * tree-ssa-sccvn (visit_use): Before looking up the original
28795 statement, try looking up the simplified expression.
28797 2013-01-19 Anthony Green <green@moxielogic.com>
28799 * config/moxie/moxie.c (moxie_expand_prologue): Set
28800 current_function_static_stack_size.
28802 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28804 PR tree-optimization/56029
28805 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
28806 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
28808 2013-01-18 Sharad Singhai <singhai@google.com>
28810 PR tree-optimization/55995
28811 * dumpfile.c (dump_loc): Print location only if available.
28812 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
28814 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
28817 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28818 secondary memory move.
28819 (inherit_reload_reg): Use rclass instead of cl for
28820 check_secondary_memory_needed_p.
28822 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28824 PR middle-end/56015
28825 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28826 the case where writing real complex part of target modifies op1.
28828 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28830 * config/aarch64/aarch64-simd.md
28831 (aarch64_vcond_internal<mode>): Handle unordered cases.
28832 * config/aarch64/iterators.md (v_cmp_result): New.
28834 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28835 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28837 * config/arm/marvell-pj4.md: New file.
28838 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28839 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28840 (generic_vfp): Likewise.
28841 * config/arm/arm-cores.def: Add marvell-pj4.
28842 * config/arm/arm-tune.md: Regenerate.
28843 * config/arm/arm-tables.opt: Regenerate.
28844 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28845 * doc/invoke.texi: Document marvell-pj4.
28847 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28849 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28851 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28856 * var-tracking.c (negative_power_of_two_p): New.
28857 (global_get_addr_cache, local_get_addr_cache): New.
28858 (get_addr_from_global_cache, get_addr_from_local_cache): New.
28859 (vt_canonicalize_addr): Rewrite using the above. Adjust the
28861 (vt_stack_offset_p): Remove.
28862 (vt_canon_true_dep): Always canonicalize loc's address.
28863 (clobber_overlapping_mems): Make sure we have a MEM.
28864 (local_get_addr_clear_given_value): New.
28865 (val_reset): Clear local cached entries.
28866 (compute_bb_dataflow): Create and release the local cache.
28867 Disable duplicate MEMs clobbering.
28868 (emit_notes_in_bb): Clobber MEMs likewise.
28869 (vt_emit_notes): Create and release the local cache.
28870 (vt_initialize, vt_finalize): Create and release the global
28871 cache, respectively.
28872 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
28874 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28876 PR libmudflap/53359
28877 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
28878 not found in the symtab.
28880 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28883 PR rtl-optimization/55547
28884 PR rtl-optimization/53827
28887 * alias.c (offset_overlap_p): New, factored out of...
28888 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
28889 the conservative special case for symbolic constants. Don't
28890 adjust zero sizes on alignment.
28892 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
28894 PR rtl-optimization/52573
28895 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
28896 REG_UNUSED for the same register.
28898 2013-01-17 Richard Biener <rguenther@suse.de>
28899 Marek Polacek <polacek@redhat.com>
28901 PR rtl-optimization/55833
28902 * loop-unswitch.c (unswitch_loops): Move loop verification...
28903 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
28904 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
28905 Set it to true when we're removing a loop from hierarchy tree in
28906 an irreducible region.
28907 (fix_bb_placements): Adjust caller.
28908 (fix_loop_placements): Likewise.
28910 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
28912 * config/avr/builtins.def (DEF_BUILTIN): Factor out
28913 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
28914 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
28915 Remove ID. Adjust comments.
28916 * config/avr/avr-c.c (avr_builtin_name): Remove.
28917 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
28918 * config/avr/avr.c (avr_tolower): New static function.
28919 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
28920 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
28921 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
28924 2013-01-17 Jan Hubicka <jh@suse.cz>
28926 PR tree-optimization/55273
28927 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
28929 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
28932 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
28933 store through atomic_store<mode>_1.
28934 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
28936 2013-01-17 Martin Jambor <mjambor@suse.cz>
28938 PR tree-optimizations/55264
28939 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
28940 for virtual methods.
28941 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
28942 virtual methods before inlining is over.
28943 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
28945 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
28948 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
28950 PR rtl-optimization/56005
28951 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
28952 pending reads for prefetch.
28954 2013-01-16 Ian Bolton <ian.bolton@arm.com>
28956 * config/aarch64/aarch64.md
28957 (*cstoresi_neg_uxtw): New pattern.
28958 (*cmovsi_insn_uxtw): New pattern.
28959 (*<optab>si3_uxtw): New pattern.
28960 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
28961 (*<optab>si3_insn_uxtw): New pattern.
28962 (*bswapsi2_uxtw): New pattern.
28964 2013-01-16 Richard Biener <rguenther@suse.de>
28966 * tree-inline.c (tree_function_versioning): Remove set but
28967 never used variable.
28969 2013-01-16 Richard Biener <rguenther@suse.de>
28971 PR tree-optimization/55964
28972 * tree-flow.h (rename_variables_in_loop): Remove.
28973 (rename_variables_in_bb): Likewise.
28974 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
28975 (copy_loop_before): Adjust and delete update-ssa status.
28976 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
28977 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
28978 (rename_variables_in_loop): Remove.
28979 (slpeel_update_phis_for_duplicate_loop): Likewise.
28980 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
28981 use available cfg machinery instead of duplicating it.
28982 Update PHI nodes and perform poor-mans SSA update here.
28983 (slpeel_tree_peel_loop_to_edge): Adjust.
28985 2013-01-16 Richard Biener <rguenther@suse.de>
28987 PR tree-optimization/54767
28988 PR tree-optimization/53465
28989 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
28990 (vrp_visit_phi_node): For PHI arguments coming via backedges
28991 drop all symbolical range information.
28992 (execute_vrp): Compute backedges.
28994 2013-01-16 Richard Biener <rguenther@suse.de>
28996 * doc/install.texi: Update CLooG and ISL requirements to
28999 2013-01-16 Christian Bruel <christian.bruel@st.com>
29002 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
29003 (broken_move): Handle UNSPECV_SP_SWITCH_B.
29004 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
29006 2013-01-16 DJ Delorie <dj@redhat.com>
29008 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
29009 (UNSPECV_SP_SWITCH_E): New.
29010 (sp_switch_1): Change to an unspec.
29011 (sp_switch_2): Change to an unspec. Don't use post-inc when we
29014 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
29016 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29017 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29018 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29019 (expand_mem_thread_fence): Ditto.
29020 (expand_mem_signal_fence): Ditto.
29021 (expand_atomic_load): Ditto.
29022 (expand_atomic_store): Ditto.
29024 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29026 PR rtl-optimization/55547
29027 PR rtl-optimization/53827
29030 * alias.c (memrefs_conflict_p): Set sizes to negative after
29033 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29036 * function.c (thread_prologue_and_epilogue_insns): Always
29037 add crtl->drap_reg to set_up_by_prologue.set, even if
29038 stack_realign_drap is false.
29040 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29042 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29043 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29044 *call): Fix indention.
29046 2013-01-15 Tom de Vries <tom@codesourcery.com>
29049 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29052 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29054 PR rtl-optimization/55153
29055 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29057 2013-01-15 Martin Jambor <mjambor@suse.cz>
29059 PR tree-optimization/55920
29060 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29061 accesses as grp_to_be_debug_replaced.
29063 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29065 PR tree-optimization/55920
29066 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29067 there is non-useless type conversion needed from debug rhs to lhs,
29068 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29070 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29071 Mikael Pettersson <mikpe@it.uu.se>
29074 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29076 (ASM_OUTPUT_CASE_LABEL): Remove.
29077 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29078 * final.c (shorten_branches): Update alignment of labels before
29079 jump tables if CASE_VECTOR_SHORTEN_MODE.
29081 2013-01-15 Richard Biener <rguenther@suse.de>
29084 * system.h: Do not include gmp.h for building host tools.
29086 2013-01-15 Richard Biener <rguenther@suse.de>
29088 PR middle-end/55882
29089 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29090 account for bitpos when computing alignment.
29092 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29094 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29095 (ix86_target_macros_internal): Likewise.
29097 * config/i386/i386.c (m_CORE2I7): Removed.
29098 (m_CORE_HASWELL): New macro.
29099 (m_CORE_ALL): Likewise.
29100 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29101 (initial_ix86_arch_features): Likewise.
29102 (processor_target_table): Initializations for Core avx2.
29103 (cpu_names): New names "core-avx2".
29104 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29105 PROCESSOR_CORE_HASWELL.
29106 (ix86_issue_rate): New case.
29107 (ia32_multipass_dfa_lookahead): Likewise.
29108 (ix86_sched_init_global): Likewise.
29110 * config/i386/i386.h (TARGET_HASWELL): New macro.
29111 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29112 (processor_type): New PROCESSOR_HASWELL.
29114 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29116 PR tree-optimization/55955
29117 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29118 *SHIFT_EXPR and *ROTATE_EXPR codes.
29120 PR tree-optimization/48766
29121 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29122 -ftrapv disable -fwrapv.
29124 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29127 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29128 etc. to 1 and not to __flash.
29129 Use LL suffix for __INT24_MAX__ with -mint8.
29130 Use ULL suffix for __UINT24_MAX__ with -mint8.
29132 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29134 * config/avr/avr-arch.h
29135 (struct base_arch_s): Use typedef avr_arch_t instead.
29136 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29137 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29138 * config/avr/avr.c: Same.
29139 * config/avr/avr-devices.c: Same.
29140 * config/avr/driver-avr.c: Same.
29141 * config/avr/gen-avr-mmcu-texi.c: Same.
29142 * config/avr/avr-mcus.def: Adjust comment.
29144 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29146 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29147 * config/aarch64/iterators.md (VALLDI): New.
29149 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29150 Andi Kleen <ak@linux.intel.com>
29153 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29154 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29157 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29159 * config/avr/avr-stdint.h: Remove trailing blanks.
29160 * config/avr/avr-log.h: Same.
29161 * config/avr/avr-arch.h: Same.
29162 * config/avr/avr-devices.c: Same.
29163 * config/avr/avr-dimode.md: Same.
29164 * config/avr/predicates.md: Same.
29165 * config/avr/avr-c.c: Same. And fix typo.
29167 * config/avr/avr-protos.h: Same. And:
29168 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
29169 (init_cumulative_args): Rename to avr_init_cumulative_args.
29170 (expand_prologue): Rename to avr_expand_prologue.
29171 (expand_epilogue): Rename to avr_expand_epilogue.
29172 (adjust_insn_length): Rename to avr_adjust_insn_length.
29173 (notice_update_cc): Rename to avr_notice_update_cc.
29174 (final_prescan_insn): Rename to avr_final_prescan_insn.
29175 * config/avr/avr.c: Same.
29176 * config/avr/avr.h: Same.
29177 * config/avr/avr.md: Remove trailing blanks.
29178 (prologue): Use avr_expand_prologue.
29179 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
29181 2013-01-14 Richard Biener <rguenther@suse.de>
29183 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
29184 verify_location, collect_subblocks): New functions.
29185 (verify_gimple_in_cfg): Verify that locations only reference
29186 BLOCKs in the functions BLOCK tree.
29188 2013-01-14 Richard Biener <rguenther@suse.de>
29190 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
29192 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
29194 (insert_out_of_ssa_copy_on_edge): Likewise.
29195 (rewrite_close_phi_out_of_ssa): Likewise.
29196 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
29198 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
29199 propagated constants.
29200 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
29203 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29205 * config/avr/avr-modes.def: Add GPL copyright notice.
29207 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
29209 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29210 MEMMODEL_MASK to determine memory model.
29211 (atomic_store<mode>): Ditto from operands[2].
29212 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29214 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29217 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29218 (fold_gimple_assign): Don't call unshare_expr here.
29219 (fold_ctor_reference): Call unshare_expr.
29221 2013-01-13 Terry Guo <terry.guo@arm.com>
29223 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29224 * doc/fragments.texi: Document MULTILIB_REUSE.
29225 * gcc.c (multilib_reuse): New internal spec.
29226 (set_multilib_dir): Also search multilib from multilib_reuse.
29227 * genmultilib (tmpmultilib3): Refactor code.
29228 (tmpmultilib4): Ditto.
29229 (multilib_reuse): New multilib argument.
29231 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29233 * Makefile.in: Update copyright.
29235 2013-01-12 Tom de Vries <tom@codesourcery.com>
29237 PR middle-end/55890
29238 * calls.c (expand_call): Check if arg_nr is valid.
29240 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29242 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29243 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29244 documentation. Add missing '__' in front of
29245 __builtin_ia32_packssdw256.
29247 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29250 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29251 NO_REGS for larl operands.
29252 (s390_reload_larl_operand): Use s390_load_address instead of
29255 2013-01-11 Richard Biener <rguenther@suse.de>
29257 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29258 (verify_node_sharing): ... here.
29259 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29261 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29263 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29264 Substitute TREECHECKING.
29265 * configure: Regenerate.
29266 * Makefile.in (TREECHECKING): New.
29268 2013-01-11 Richard Guenther <rguenther@suse.de>
29270 PR tree-optimization/44061
29271 * tree-vrp.c (extract_range_basic): Compute zero as
29272 value-range for __builtin_constant_p of function parameters.
29274 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29276 Update copyright years.
29278 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29280 PR rtl-optimization/55672
29281 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29282 const to be eliminable.
29284 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29286 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29287 * configure: Regenerate.
29289 2013-01-10 Richard Biener <rguenther@suse.de>
29291 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29293 2013-01-10 Richard Biener <rguenther@suse.de>
29296 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29297 locations for virtual PHI arguments.
29298 (rewrite_update_phi_arguments): Likewise.
29300 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29302 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29305 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29307 PR tree-optimization/55921
29308 * tree-complex.c (expand_complex_asm): New function.
29309 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29311 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29314 * config/s390/s390.c (s390_symref_operand_p)
29315 (s390_loadrelative_operand_p): Merge the two functions.
29316 (s390_check_qrst_address, print_operand_address): Add parameters
29317 to s390_loadrelative_operand_p invokation.
29318 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29319 (s390_reload_larl_operand, s390_secondary_reload): Use
29320 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29321 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29323 2013-01-09 Mike Stump <mikestump@comcast.net>
29325 * dse.c (record_store): Remove unnecessary assert.
29327 2013-01-09 Jan Hubicka <jh@suse.cz>
29329 PR tree-optimization/55569
29330 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29331 * cfgloop.h (scale_loop_profile): Likewise.
29333 2013-01-09 Jan Hubicka <jh@suse.cz>
29336 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29338 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29340 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29342 PR middle-end/55114
29343 * expr.h (maybe_emit_group_store): Declare.
29344 * expr.c (maybe_emit_group_store): New function.
29345 * builtins.c (expand_builtin_int_roundingfn): Call it.
29346 (expand_builtin_int_roundingfn_2): Likewise.
29348 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29350 PR rtl-optimization/55829
29351 * lra-constraints.c (match_reload): Add code for absent output.
29352 (curr_insn_transform): Add code for reloads of matched inputs
29355 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29357 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29358 attribute of movddup insn to DF.
29359 (*vec_interleave_lowv2df): Ditto.
29360 (vec_dupv2df): Ditto.
29362 2013-01-09 Jan Hubicka <jh@suse.cz>
29364 PR tree-optimiation/55875
29365 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29366 EVERY_ITERATION parameter.
29367 (number_of_iterations_exit): Check if exit is executed every iteration.
29368 (idx_infer_loop_bounds): Similarly here.
29369 (n_of_executions_at_most): Simplify
29370 to only test for cases where statement is dominated by the
29371 particular bound; handle correctly the "postdominance" test.
29372 (scev_probably_wraps_p): Use max loop iterations info
29373 as a global bound first.
29375 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29376 Nick Clifton <nickc@redhat.com>
29378 * config/v850/v850.md (cbranchsf4): New pattern.
29379 (cstoresf4): New pattern.
29380 (cbranchdf4): New pattern.
29381 (cstoredf4): New pattern.
29382 (movsicc): Disallow floating point comparisons.
29383 (cmpsf_le_insn): Fix order of operators.
29384 (cmpsf_lt_insn): Likewise.
29385 (cmpsf_eq_insn): Likewise.
29386 (cmpdf_le_insn): Likewise.
29387 (cmpdf_lt_insn): Likewise.
29388 (cmpdf_eq_insn): Likewise.
29389 (cmpsf_ge_insn): Use LE comparison.
29390 (cmpdf_ge_insn): Likewise.
29391 (cmpsf_gt_insn): Use LT comparison.
29392 (cmpdf_gt_insn): Likewise.
29393 (cmpsf_ne_insn): Delete pattern.
29394 (cmpdf_ne_insn): Delete pattern.
29395 * config/v850/v850.c (v850_gen_float_compare): Use
29396 gen_cmpdf_eq_insn for NE comparison.
29397 (v850_float_z_comparison_operator)
29398 (v850_float_nz_comparison_operator): Move from here ...
29399 * config/v850/predicates.md: ... to here. Move GT and GE
29400 comparisons into v850_float_z_comparison_operator.
29401 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29403 (v850_float_nz_comparison_operator): Likewise.
29405 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29407 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29408 with calls to gen_insvsi/gen_insvdi.
29410 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29412 * config/i386/i386.c (initial_ix86_tune_features): Set up
29413 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29415 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29416 Jakub Jelinek <jakub@redhat.com>
29418 PR tree-optimization/48189
29419 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29420 If nitercst is 0, don't predict the exit edge.
29422 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29424 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29425 in asm_fprintf with reg_names.
29426 (aarch64_print_operand_address): Likewise.
29427 (aarch64_return_addr): Likewise.
29428 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29430 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29432 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29433 (INT_U6_BITS): Likewise.
29434 * config/pa/predicates.md (uint6_operand): New predicate.
29435 (shift5_operand, shift6_operand): Likewise.
29436 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29438 (lshrdi3): Use shift6_operand.
29439 (shrpsi4, shrpdi4): New insn patterns.
29440 (extzv): Delete expander.
29441 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29442 predicates in unamed zero extract patterns. Tighten common constraint.
29443 (extv): Delete expander.
29444 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29445 predicates in unamed sign extract patterns. Tighten common constraint.
29446 (insv): Delete expander.
29447 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29448 predicates in unamed insert patterns. Tighten common constraint.
29449 Change uint32_operand predicate to uint6_operand predicate in unamed
29450 DImode pattern to insert constant values of type 1...1xxxx.
29452 2013-01-04 Jan Hubicka <jh@suse.cz>
29454 PR tree-optimization/55823
29455 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29458 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29459 Uros Bizjak <ubizjak@gmail.com>
29461 PR rtl-optimization/55845
29462 * df-problems.c (can_move_insns_across): Stop scanning at
29463 volatile_insn_p source instruction or give up if
29464 across_from .. across_to range contains any volatile_insn_p
29467 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29469 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29470 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29472 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29473 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29475 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29478 * asan.c (asan_clear_shadow): New function.
29479 (asan_emit_stack_protection): Use it.
29481 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29483 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29484 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29485 with tab instead of space.
29487 2013-01-08 Nick Clifton <nickc@redhat.com>
29489 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29490 register bank 0 at the start of an interrupt handler.
29491 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29494 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29496 * config/aarch64/aarch64-simd.md
29497 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29498 (aarch64_simd_bsl): Likewise.
29499 (aarch64_vcond_internal<mode>): Likewise.
29500 (vcond<mode><mode>): Likewise.
29501 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29502 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29504 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29506 * config/aarch64/aarch64-builtins.c
29507 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29509 2013-01-08 Martin Jambor <mjambor@suse.cz>
29512 * tree-sra.c (analyze_access_subtree): Return true also after
29513 potentially creating a debug-only replacement.
29515 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29517 PR middle-end/55890
29518 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29520 PR tree-optimization/54120
29521 * tree-vrp.c (range_fits_type_p): Don't allow
29522 src_precision < precision from signed vr to unsigned_p
29523 if vr->min or vr->max is negative.
29524 (simplify_float_conversion_using_ranges): Test can_float_p
29525 against CODE_FOR_nothing.
29527 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29528 Richard Biener <rguenther@suse.de>
29530 PR middle-end/55851
29531 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29532 types instead of just INTEGER_TYPE types.
29534 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29536 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29539 2013-01-07 Steve Ellcey <sellcey@mips.com>
29542 * config/mips/mips.opt: Change mad to mmad to match documentation.
29544 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29547 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29548 .progmemx.data now.
29550 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29553 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29554 (avr_addrspace_t): Add .section_name field.
29555 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29557 (avr_addrspace): Same. Initialize .section_name. Remove last
29558 NULL entry. Put __memx into .progmemx.data.
29559 (progmem_section_prefix): Remove.
29560 (avr_asm_init_sections): No need to initialize progmem_section.
29561 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29562 section name prefix.
29563 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29564 retrieve the progmem section.
29565 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29566 boundary to run over avr_addrspace[].
29567 (avr_register_target_pragmas): Ditto.
29569 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29571 * varasm.c (output_constant_def_contents): For asan_protect_global
29572 protected strings, adjust DECL_ALIGN if needed, before testing for
29574 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29575 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29577 (output_object_block): For asan protected decls, emit asan padding
29578 after their contents.
29579 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29580 (asan_finish_file): Test it here instead.
29582 2013-01-07 Nick Clifton <nickc@redhat.com>
29583 Matthias Klose <doko@debian.org>
29584 Doug Kwan <dougkwan@google.com>
29585 H.J. Lu <hongjiu.lu@intel.com>
29588 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29590 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29592 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29594 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29596 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29598 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29601 * doc/install.texi (Cross-Compiler-Specific Options): Document
29604 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29606 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29607 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29608 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29609 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29610 vqmovun_high_s64): Fix source operand number and update copyright.
29612 2013-01-07 Richard Biener <rguenther@suse.de>
29614 PR middle-end/55890
29615 * gimple.h (gimple_call_builtin_p): New overload.
29616 * gimple.c (validate_call): New function.
29617 (gimple_call_builtin_p): Likewise.
29618 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29619 Use gimple_call_builtin_p.
29620 (find_func_clobbers): Likewise.
29621 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29622 (strlen_optimize_stmt): Likewise.
29624 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29626 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29627 (vld1q_dup_*): Likewise.
29628 (vld1_*): Likewise.
29629 (vld1q_*): Likewise.
29630 (vld1_lane_*): Likewise.
29631 (vld1q_lane_*): Likewise.
29633 2013-01-07 Richard Biener <rguenther@suse.de>
29635 * lto-streamer.h (LTO_minor_version): Bump to 2.
29637 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29639 * config/aarch64/aarch64-protos.h
29640 (aarch64_const_double_zero_rtx_p): Rename to...
29641 (aarch64_float_const_zero_rtx_p): ...this.
29642 (aarch64_float_const_representable_p): New.
29643 (aarch64_output_simd_mov_immediate): Likewise.
29644 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
29645 move immediate case.
29646 * config/aarch64/aarch64.c
29647 (aarch64_const_double_zero_rtx_p): Rename to...
29648 (aarch64_float_const_zero_rtx_p): ...this.
29649 (aarch64_print_operand): Allow printing of new constants.
29650 (aarch64_valid_floating_const): New.
29651 (aarch64_legitimate_constant_p): Check for valid floating-point
29653 (aarch64_simd_valid_immediate): Likewise.
29654 (aarch64_vect_float_const_representable_p): New.
29655 (aarch64_float_const_representable_p): Likewise.
29656 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
29657 (aarch64_output_simd_mov_immediate): New.
29658 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
29659 (*movdf_aarch64): Likewise.
29660 * config/aarch64/constraints.md (Ufc): New.
29661 (Y): call aarch64_float_const_zero_rtx.
29662 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
29664 2013-01-07 Richard Biener <rguenther@suse.de>
29666 PR tree-optimization/55888
29667 PR tree-optimization/55862
29668 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
29669 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
29670 not if it is contained therein.
29672 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29674 * config/avr/t-avr: Typo.
29676 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29679 * config/avr/t-avr: Don't automatically rebuild
29680 $(srcdir)/config/avr/t-multilib
29681 $(srcdir)/config/avr/avr-tables.opt
29682 $(srcdir)/doc/avr-mmcu.texi
29683 (avr-mcus): New phony target to build them on request.
29684 (s-avr-mlib, s-avr-mmcu-texi): Remove.
29685 * avr/avr-mcus.def: Adjust comments.
29687 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
29689 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
29691 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29693 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
29695 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29697 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
29699 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
29701 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
29702 to generate profiling.
29703 * config/rs6000/aix64.h (LIB_SPEC): Same.
29705 2013-01-04 Andrew Pinski <apinski@cavium.com>
29707 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
29709 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
29711 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
29713 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
29715 (ix86_expand_move): Ditto.
29716 (ix86_zero_extend_to_Pmode): Ditto.
29717 (ix86_expand_call): Ditto.
29718 (ix86_expand_special_args_builtin): Ditto.
29719 (ix86_expand_builtin): Ditto.
29721 2013-01-04 Richard Biener <rguenther@suse.de>
29723 PR tree-optimization/55862
29724 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
29725 translating them through PHI nodes.
29727 2013-01-04 Martin Jambor <mjambor@suse.cz>
29729 PR tree-optimization/55755
29730 * tree-sra.c (sra_modify_assign): Do not check that an access has no
29731 children when trying to avoid producing a VIEW_CONVERT_EXPR.
29733 2013-01-04 Marek Polacek <polacek@redhat.com>
29735 PR middle-end/55859
29736 * opts.c (default_options_optimization): Clarify error message.
29738 2013-01-04 Richard Biener <rguenther@suse.de>
29740 PR middle-end/55863
29741 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
29744 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29747 * config/pa/pa.md (movsi): Revert previous change.
29748 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
29751 2013-01-03 Richard Henderson <rth@redhat.com>
29753 * config/i386/i386.c (ix86_expand_move): Always assign to op1
29754 after eliminating TLS symbols.
29756 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
29759 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
29760 * graphite-poly.c (debug_gmp_value): Likewise.
29762 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
29765 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
29766 selected code model, define __code_mode_small__, __code_model_medium__,
29767 __code_model_large__, __code_model_32__ or __code_model_kernel__.
29768 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
29769 xchg temporary register with %k. Declare temporary register as
29771 [__x86_64__]: For medium and large code models, preserve %rbx register.
29773 2013-01-03 Richard Biener <rguenther@suse.de>
29775 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
29776 (dump_subscript): Adjust.
29777 (finalize_ddr_dependent): Do not dump redundant info.
29778 (analyze_siv_subscript): Adjust.
29779 (subscript_dependence_tester): Likewise.
29780 (compute_affine_dependence): Likewise.
29782 2013-01-03 Richard Biener <rguenther@suse.de>
29785 2013-01-03 Richard Biener <rguenther@suse.de>
29787 PR tree-optimization/55857
29788 * tree-vect-stmts.c (vectorizable_load): Do not setup
29789 re-alignment for invariant loads.
29791 2013-01-02 Richard Biener <rguenther@suse.de>
29793 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29794 invariant load do not generate a vector load from the scalar location.
29796 2013-01-03 Richard Biener <rguenther@suse.de>
29798 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
29799 for not vectorizing.
29800 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
29801 not build INDIRECT_REFs, call get_name once only.
29802 (vect_create_data_ref_ptr): Likewise. Dump base object kind
29803 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
29805 2013-01-03 Richard Biener <rguenther@suse.de>
29807 PR tree-optimization/55857
29808 * tree-vect-stmts.c (vectorizable_load): Do not setup
29809 re-alignment for invariant loads.
29811 2013-01-03 Richard Biener <rguenther@suse.de>
29814 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
29815 prefer a built-in decl.
29817 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29819 * gcc.c (process_command): Update copyright notice dates.
29820 * gcov.c (print_version): Likewise.
29821 * gcov-dump.c (print_version): Likewise.
29823 PR rtl-optimization/55838
29824 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29825 iv0.step, iv1.step and step.
29827 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29828 Marc Glisse <marc.glisse@inria.fr>
29830 PR tree-optimization/55832
29831 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29832 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29833 integer_{one,zero}_node.
29835 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29838 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29839 * var-tracking.c (reverse_op): Don't add reverse ops to
29840 VALUEs that have already
29841 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29843 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29845 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29847 2013-01-02 Teresa Johnson <tejohnson@google.com>
29849 * dumpfile.c (dump_loc): Print filename with location.
29850 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29851 new location_t parameter to emit complete unroll message with
29852 new dump framework.
29853 (canonicalize_loop_induction_variables): Compute loops location
29854 and pass to try_unroll_loop_completely.
29855 * loop-unroll.c (report_unroll_peel): New function.
29856 (peel_loops_completely): Use new dump format with location
29857 for main dumpfile message, and invoke report_unroll_peel on success.
29858 (decide_unrolling_and_peeling): Ditto.
29859 (decide_peel_once_rolling): Remove old dumpfile message subsumed
29860 by report_unroll_peel.
29861 (decide_peel_completely): Ditto.
29862 (decide_unroll_constant_iterations): Ditto.
29863 (decide_unroll_runtime_iterations): Ditto.
29864 (decide_peel_simple): Ditto.
29865 (decide_unroll_stupid): Ditto.
29866 * cfgloop.c (get_loop_location): New function.
29867 * cfgloop.h (get_loop_location): Declare.
29869 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29871 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
29874 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29876 PR middle-end/55198
29877 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
29878 BLKmode objects when EXPAND_MEMORY is specified.
29880 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29882 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
29884 (fold_builtin_cpu): Do not share cpu model decls across statements.
29886 2013-01-02 Jason Merrill <jason@redhat.com>
29889 * tree.c (build_array_type_1): Revert earlier change.
29891 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
29893 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
29895 * config/aarch64/aarch64-tune.md: Re-generate.
29897 2013-01-02 Richard Biener <rguenther@suse.de>
29899 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29900 invariant load do not generate a vector load from the scalar location.
29902 2013-01-02 Richard Biener <rguenther@suse.de>
29905 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
29906 * configure: Regenerate.
29908 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29910 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
29911 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
29912 (expand_builtin_int_roundingfn_2): Keep the original target around
29913 for the fallback case.
29915 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29917 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
29918 to be clear for sign changes.
29920 2013-01-01 Jan Hubicka <jh@suse.cz>
29922 * ipa-inline-analysis.c: Fix formatting.
29924 2013-01-01 Jakub Jelinek <jakub@redhat.com>
29926 PR tree-optimization/55831
29927 * tree-vect-loop.c (get_initial_def_for_induction): Use
29928 gsi_after_labels instead of gsi_start_bb.
29930 Copyright (C) 2013 Free Software Foundation, Inc.
29932 Copying and distribution of this file, with or without modification,
29933 are permitted in any medium without royalty provided the copyright
29934 notice and this notice are preserved.